michal/mcpctl
1
0
Fork 0
Code Issues Pull Requests 7 Actions Packages Projects Releases Wiki Activity
Files
fix/injector-image-server-argv
mcpctl/pnpm-lock.yaml

10346 lines
334 KiB
YAML
Raw Permalink Normal View History

first commit
2026-02-21 03:10:39 +00:00
lockfileVersion: '9.0'
settings:
autoInstallPeers: true
excludeLinksFromLockfile: false
importers:
.:
devDependencies:
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
'@earendil-works/pi-ai':
specifier: ^0.84.1
version: 0.84.1(@modelcontextprotocol/sdk@1.26.0(zod@3.25.76))(ws@8.19.0)(zod@3.25.76)
'@earendil-works/pi-coding-agent':
specifier: 0.84.1
version: 0.84.1(@modelcontextprotocol/sdk@1.26.0(zod@3.25.76))(ws@8.19.0)(zod@3.25.76)
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
'@opencode-ai/plugin':
specifier: 1.18.15
version: 1.18.15(@opentui/core@0.4.5(typescript@5.9.3)(web-tree-sitter@0.25.10))(@opentui/keymap@0.4.5(@opentui/solid@0.4.5(solid-js@1.9.12)(typescript@5.9.3)(web-tree-sitter@0.25.10))(react@19.2.5)(solid-js@1.9.12)(typescript@5.9.3)(web-tree-sitter@0.25.10))(@opentui/solid@0.4.5(solid-js@1.9.12)(typescript@5.9.3)(web-tree-sitter@0.25.10))
'@opentui/core':
specifier: 0.4.5
version: 0.4.5(typescript@5.9.3)(web-tree-sitter@0.25.10)
'@opentui/keymap':
specifier: 0.4.5
version: 0.4.5(@opentui/solid@0.4.5(solid-js@1.9.12)(typescript@5.9.3)(web-tree-sitter@0.25.10))(react@19.2.5)(solid-js@1.9.12)(typescript@5.9.3)(web-tree-sitter@0.25.10)
'@opentui/solid':
specifier: 0.4.5
version: 0.4.5(solid-js@1.9.12)(typescript@5.9.3)(web-tree-sitter@0.25.10)
feat: implement MCP registry client with multi-source search Add registry client that queries Official, Glama, and Smithery MCP registries with caching, request deduplication, retry logic, and result ranking/dedup. Includes 53 tests covering all components. Also fix null priority values in cancelled tasks (19-21) that broke Task Master, and add new tasks 25-27 for registry completion and CLI discover/install commands. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 03:46:14 +00:00
'@types/node':
specifier: ^25.3.0
version: 25.3.0
first commit
2026-02-21 03:10:39 +00:00
'@typescript-eslint/eslint-plugin':
specifier: ^8.56.0
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
version: 8.56.0(@typescript-eslint/parser@8.56.0(eslint@10.0.1(jiti@2.7.0))(typescript@5.9.3))(eslint@10.0.1(jiti@2.7.0))(typescript@5.9.3)
first commit
2026-02-21 03:10:39 +00:00
'@typescript-eslint/parser':
specifier: ^8.56.0
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
version: 8.56.0(eslint@10.0.1(jiti@2.7.0))(typescript@5.9.3)
first commit
2026-02-21 03:10:39 +00:00
'@vitest/coverage-v8':
specifier: ^4.0.18
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
version: 4.0.18(vitest@4.0.18(@opentelemetry/api@1.9.0)(@types/node@25.3.0)(jiti@2.7.0)(jsdom@28.1.0)(lightningcss@1.32.0)(tsx@4.21.0)(yaml@2.9.0))
first commit
2026-02-21 03:10:39 +00:00
eslint:
specifier: ^10.0.1
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
version: 10.0.1(jiti@2.7.0)
first commit
2026-02-21 03:10:39 +00:00
eslint-config-prettier:
specifier: ^10.1.8
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
version: 10.1.8(eslint@10.0.1(jiti@2.7.0))
first commit
2026-02-21 03:10:39 +00:00
rimraf:
specifier: ^6.1.3
version: 6.1.3
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
solid-js:
specifier: 1.9.12
version: 1.9.12
first commit
2026-02-21 03:10:39 +00:00
tsx:
specifier: ^4.21.0
version: 4.21.0
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
typebox:
specifier: 1.3.11
version: 1.3.11
first commit
2026-02-21 03:10:39 +00:00
typescript:
specifier: ^5.9.3
version: 5.9.3
vitest:
specifier: ^4.0.18
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
version: 4.0.18(@opentelemetry/api@1.9.0)(@types/node@25.3.0)(jiti@2.7.0)(jsdom@28.1.0)(lightningcss@1.32.0)(tsx@4.21.0)(yaml@2.9.0)
first commit
2026-02-21 03:10:39 +00:00
src/cli:
dependencies:
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
'@inkjs/ui':
specifier: ^2.0.0
version: 2.0.0(ink@6.8.0(@types/react@19.2.14)(react@19.2.4))
first commit
2026-02-21 03:10:39 +00:00
'@mcpctl/db':
specifier: workspace:*
version: link:../db
'@mcpctl/shared':
specifier: workspace:*
version: link:../shared
chalk:
specifier: ^5.4.0
version: 5.6.2
commander:
specifier: ^13.0.0
version: 13.1.0
feat: eager vLLM warmup and smart page titles in paginate stage - Add warmup() to LlmProvider interface for eager subprocess startup - ManagedVllmProvider.warmup() starts vLLM in background on project load - ProviderRegistry.warmupAll() triggers all managed providers - NamedProvider proxies warmup() to inner provider - paginate stage generates LLM-powered descriptive page titles when available, cached by content hash, falls back to generic "Page N" - project-mcp-endpoint calls warmupAll() on router creation so vLLM is loading while the session initializes Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-03-03 19:07:39 +00:00
diff:
specifier: ^8.0.3
version: 8.0.3
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
ink:
specifier: ^6.8.0
version: 6.8.0(@types/react@19.2.14)(react@19.2.4)
first commit
2026-02-21 03:10:39 +00:00
inquirer:
specifier: ^12.0.0
feat: implement MCP registry client with multi-source search Add registry client that queries Official, Glama, and Smithery MCP registries with caching, request deduplication, retry logic, and result ranking/dedup. Includes 53 tests covering all components. Also fix null priority values in cancelled tasks (19-21) that broke Task Master, and add new tasks 25-27 for registry completion and CLI discover/install commands. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 03:46:14 +00:00
version: 12.11.1(@types/node@25.3.0)
first commit
2026-02-21 03:10:39 +00:00
js-yaml:
specifier: ^4.1.0
version: 4.1.1
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
react:
specifier: ^19.2.4
version: 19.2.4
feat: implement MCP registry client with multi-source search Add registry client that queries Official, Glama, and Smithery MCP registries with caching, request deduplication, retry logic, and result ranking/dedup. Includes 53 tests covering all components. Also fix null priority values in cancelled tasks (19-21) that broke Task Master, and add new tasks 25-27 for registry completion and CLI discover/install commands. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 03:46:14 +00:00
zod:
specifier: ^3.24.0
version: 3.25.76
feat: build CLI core framework with Commander.js Add CLI entry point with Commander.js, config management (~/.mcpctl/config.json with Zod validation), output formatters (table/json/yaml), config and status commands with dependency injection for testing. Fix sanitizeString regex ordering. 67 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:17:31 +00:00
devDependencies:
feat: eager vLLM warmup and smart page titles in paginate stage - Add warmup() to LlmProvider interface for eager subprocess startup - ManagedVllmProvider.warmup() starts vLLM in background on project load - ProviderRegistry.warmupAll() triggers all managed providers - NamedProvider proxies warmup() to inner provider - paginate stage generates LLM-powered descriptive page titles when available, cached by content hash, falls back to generic "Page N" - project-mcp-endpoint calls warmupAll() on router creation so vLLM is loading while the session initializes Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-03-03 19:07:39 +00:00
'@types/diff':
specifier: ^8.0.0
version: 8.0.0
feat: build CLI core framework with Commander.js Add CLI entry point with Commander.js, config management (~/.mcpctl/config.json with Zod validation), output formatters (table/json/yaml), config and status commands with dependency injection for testing. Fix sanitizeString regex ordering. 67 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:17:31 +00:00
'@types/js-yaml':
specifier: ^4.0.9
version: 4.0.9
'@types/node':
specifier: ^25.3.0
version: 25.3.0
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
'@types/react':
specifier: ^19.2.14
version: 19.2.14
first commit
2026-02-21 03:10:39 +00:00
src/db:
dependencies:
'@mcpctl/shared':
specifier: workspace:*
version: link:../shared
'@prisma/client':
specifier: ^6.0.0
version: 6.19.2(prisma@6.19.2(typescript@5.9.3))(typescript@5.9.3)
devDependencies:
prisma:
specifier: ^6.0.0
version: 6.19.2(typescript@5.9.3)
src/mcpd:
dependencies:
'@fastify/cors':
specifier: ^10.0.0
version: 10.1.0
'@fastify/helmet':
specifier: ^12.0.0
version: 12.0.1
'@fastify/rate-limit':
specifier: ^10.0.0
version: 10.3.0
feat(mcpd+deploy): serve web UI at /ui + smoke tests + docs (Stage 6) The closing stage. mcpd now hosts the Stage 5 SPA, the Docker image bundles the build artifact, a smoke test exercises the personality HTTP surface end-to-end, and the user-facing docs spell out the mental model. mcpd: - Add @fastify/static dep. - New routes/web-ui.ts: registers /ui/* against a static bundle. Looks for the bundle at $MCPD_WEB_ROOT, then /usr/share/mcpd/web (the Docker image path), then a dev-tree fallback. Logs and skips cleanly if missing — API-only deploys keep working. - SPA fallback: any /ui/<path> that doesn't match a file falls through to index.html so direct hits to react-router URLs work. - /ui/* falls through to `kind: skip` in mapUrlToPermission, so the static assets are served unauthenticated. Each API call from the SPA still carries the bearer token. Deploy: - Dockerfile.mcpd builds the @mcpctl/web bundle in the same builder stage and copies dist/ to /usr/share/mcpd/web in the runtime image. Smoke (personality.smoke.test.ts): - Live mcpd flow: create secret/llm/agent/personality, attach an agent-direct prompt, verify the binding listing, reject double- attach (409) + foreign-agent prompt (400), set defaultPersonality by name, detach + delete cleanup. Docs: - New docs/personalities.md: VLAN-on-ethernet model, system-block ordering table, three prompt scopes, CLI walkthrough, web UI walkthrough, full API surface, RBAC notes. - agents.md and chat.md cross-link. - README's Agents section gains a Personalities subsection. Test count after Stage 6: mcpd: 801/801 cli: 430/430 web: 7/7 db: 58/62 (4 pre-existing) Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:48:43 +01:00
'@fastify/static':
specifier: ^8.0.0
version: 8.3.0
feat: add Kubernetes orchestrator for MCP server pod management mcpd can now deploy MCP server instances as Kubernetes pods instead of Docker containers. Set MCPD_ORCHESTRATOR=kubernetes to enable. - Add @kubernetes/client-node with thin wrapper (context enforcement via MCPD_K8S_CONTEXT to prevent multi-cluster mishaps) - Rewrite KubernetesOrchestrator: pod CRUD, pod IP extraction, exec via SPDY (one-shot + interactive), log streaming - Manifest generator: stdin:true for STDIO servers, args (not command) to preserve runner image entrypoint, security hardening - Orchestrator selection in main.ts via MCPD_ORCHESTRATOR env var - 25 unit tests for k8s orchestrator, all 624 tests pass Tested end-to-end on local k3s: - mcpd deployed via Pulumi, creates pods in mcpctl-servers namespace - NetworkPolicy verified: only mcpd can reach MCP server pods - Python runner (uvx) successfully runs aws-documentation-mcp-server Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-08 01:55:13 +01:00
'@kubernetes/client-node':
specifier: ^1.4.0
version: 1.4.0
first commit
2026-02-21 03:10:39 +00:00
'@mcpctl/db':
specifier: workspace:*
version: link:../db
'@mcpctl/shared':
specifier: workspace:*
version: link:../shared
feat: add MCP server and profile management API Add validation schemas (Zod), repository pattern with Prisma, service layer with business logic (NotFoundError, ConflictError), and REST routes for MCP server and profile CRUD. 86 mcpd tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:26:18 +00:00
'@prisma/client':
specifier: ^6.0.0
version: 6.19.2(prisma@6.19.2(typescript@5.9.3))(typescript@5.9.3)
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
bcrypt:
specifier: ^5.1.1
version: 5.1.1
feat(mcpd): ResourceRevision + ResourceProposal services + Prompt revision integration Phase 2 of the Skills + Revisions + Proposals work. Stands up the generic revision/proposal layer and wires Prompt into it. Skills will plug into the same infrastructure in PR-3 with no further service changes required. This PR is intentionally additive: PromptRequest table and routes are unchanged. The /api/v1/proposals API runs side-by-side with the legacy /api/v1/promptrequests API. The PromptRequest cutover (rename + backfill + mcplocal rewire) is deferred to a later PR so this one stays reviewable. ## What's added ### Repositories (src/mcpd/src/repositories/) - resource-revision.repository.ts — append-only revision log keyed by (resourceType, resourceId). Soft FK; no relations declared. Supports history listing, semver lookup, and contentHash cross-resource search. - resource-proposal.repository.ts — generic propose queue. Status lifecycle pending → approved | rejected. Mirrors Prompt's `?? ''` workaround for nullable-FK compound lookups. ### Services (src/mcpd/src/services/) - resource-revision.service.ts — record() inserts a revision with a stable sha256 contentHash computed from canonicalised JSON (key-sorted at every level so reordered objects produce the same hash). Caller passes a pre-computed semver; service does NOT decide bump policy. - resource-proposal.service.ts — propose / approve / reject / list, with a per-resourceType handler registry. PromptService registers the 'prompt' handler at construction; the SkillService will register 'skill' in PR-3. approve() runs in a Prisma $transaction so the resource update + revision insert + proposal status flip are atomic. ### Pure utility (src/mcpd/src/utils/semver.ts) - bumpSemver(current, kind) for major / minor / patch - compareSemver(a, b) — numeric, not lex (10 > 9) - isValidSemver(s) - Invalid input falls back to '0.1.0' rather than throwing — keeps the audit-write path from blowing up the prompt update if a row's semver ever drifts out of MAJOR.MINOR.PATCH shape. ### Routes (src/mcpd/src/routes/) - revisions.ts — GET /api/v1/revisions?resourceType=&resourceId=, GET /api/v1/revisions/:id, GET /api/v1/revisions/:id/diff?against=<id|live> (unified-format diff via the `diff` package), and POST /api/v1/prompts/:id/restore-revision { revisionId, note? }. - proposals.ts — GET / POST /api/v1/proposals, GET /api/v1/proposals/:id, PUT for body updates, POST .../approve and POST .../reject, plus DELETE. ## What's changed - PromptService.create / update now record a ResourceRevision when the revision service is wired. Update auto-bumps patch on content change; authors can override via `--bump major|minor|patch` or `--semver X.Y.Z` on the CLI (forwarded into the PUT body). Best-effort: revision write failures are swallowed so the prompt save still succeeds (revision is audit, not source of truth). - PromptService.setProposalService registers a 'prompt' approval handler with the proposal service. Approval runs in a Prisma transaction: upsert prompt → record revision → update currentRevisionId → flip proposal status. semver bumps to 0.1.0 on first approval, patch thereafter. - New CLI flags on `mcpctl edit prompt`: --bump, --semver, --note. They're prompt-only (validated client-side); other resources reject them. - Aliases in shared.ts: `proposal`/`prop` → proposals, `revision`/`rev` → revisions. - diff dependency added to mcpd. ## Tests - src/mcpd/tests/utils/semver.test.ts — covers bump/compare/validate including numeric (not lex) semver compare and invalid-input fallback. - prompt-service.test.ts updated: makePrompt fixture now sets semver + agentId + currentRevisionId; updatePrompt assertion expects the auto-bumped patch in the same update call. - prompt-routes.test.ts updated symmetrically. ## RBAC `proposals` and `revisions` URL segments map to the existing `prompts` permission for now. PR-7 may split if a "reviewer" role becomes useful. ## Verification Full suite: 158 test files / 2127 tests green. `pnpm build` clean across all 6 workspace packages. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-07 00:38:35 +01:00
diff:
specifier: ^5.2.0
version: 5.2.2
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
dockerode:
specifier: ^4.0.9
version: 4.0.9
first commit
2026-02-21 03:10:39 +00:00
fastify:
specifier: ^5.0.0
version: 5.7.4
feat: add MCP server templates and deployment infrastructure Introduce a Helm-chart-like template system for MCP servers. Templates are YAML files in templates/ that get seeded into the DB on startup. Users can browse them with `mcpctl get templates`, inspect with `mcpctl describe template`, and instantiate with `mcpctl create server --from-template=`. Also adds Portainer deployment scripts, mcplocal systemd service, Streamable HTTP MCP endpoint, and RPM packaging for mcpctl-local. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 22:24:35 +00:00
js-yaml:
specifier: ^4.1.0
version: 4.1.1
feat(errors): mcpd error-log ring buffer + 'mcpctl errors' Operators can now see recent mcpd error/fatal logs without kubectl: - mcpd tees level>=error pino records into an in-memory ring buffer (src/mcpd/src/services/error-log-buffer.ts; wired via pino.multistream in server.ts so stdout logging is unchanged). Captures structured errors incl. fatal kinds like BACKEND_TOKEN_DEAD. - GET /api/v1/logs/errors?limit=N (RBAC: 'logs' operation). - CLI: 'mcpctl errors [-n N]' renders TIME/LEVEL/DETAIL, most-recent-first. Buffer unit tests (6); full suite 2223 passing. Needs a deploy to go live. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-16 23:25:55 +01:00
pino:
specifier: ^10.3.1
version: 10.3.1
first commit
2026-02-21 03:10:39 +00:00
zod:
specifier: ^3.24.0
version: 3.25.76
feat: implement mcpd core server framework with Fastify Add Fastify server with config validation (Zod), health/healthz endpoints, auth middleware (Bearer token + session lookup), security plugins (CORS, Helmet, rate limiting), error handler, audit logging, and graceful shutdown. 36 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:22:01 +00:00
devDependencies:
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
'@types/bcrypt':
specifier: ^5.0.2
version: 5.0.2
feat(mcpd): ResourceRevision + ResourceProposal services + Prompt revision integration Phase 2 of the Skills + Revisions + Proposals work. Stands up the generic revision/proposal layer and wires Prompt into it. Skills will plug into the same infrastructure in PR-3 with no further service changes required. This PR is intentionally additive: PromptRequest table and routes are unchanged. The /api/v1/proposals API runs side-by-side with the legacy /api/v1/promptrequests API. The PromptRequest cutover (rename + backfill + mcplocal rewire) is deferred to a later PR so this one stays reviewable. ## What's added ### Repositories (src/mcpd/src/repositories/) - resource-revision.repository.ts — append-only revision log keyed by (resourceType, resourceId). Soft FK; no relations declared. Supports history listing, semver lookup, and contentHash cross-resource search. - resource-proposal.repository.ts — generic propose queue. Status lifecycle pending → approved | rejected. Mirrors Prompt's `?? ''` workaround for nullable-FK compound lookups. ### Services (src/mcpd/src/services/) - resource-revision.service.ts — record() inserts a revision with a stable sha256 contentHash computed from canonicalised JSON (key-sorted at every level so reordered objects produce the same hash). Caller passes a pre-computed semver; service does NOT decide bump policy. - resource-proposal.service.ts — propose / approve / reject / list, with a per-resourceType handler registry. PromptService registers the 'prompt' handler at construction; the SkillService will register 'skill' in PR-3. approve() runs in a Prisma $transaction so the resource update + revision insert + proposal status flip are atomic. ### Pure utility (src/mcpd/src/utils/semver.ts) - bumpSemver(current, kind) for major / minor / patch - compareSemver(a, b) — numeric, not lex (10 > 9) - isValidSemver(s) - Invalid input falls back to '0.1.0' rather than throwing — keeps the audit-write path from blowing up the prompt update if a row's semver ever drifts out of MAJOR.MINOR.PATCH shape. ### Routes (src/mcpd/src/routes/) - revisions.ts — GET /api/v1/revisions?resourceType=&resourceId=, GET /api/v1/revisions/:id, GET /api/v1/revisions/:id/diff?against=<id|live> (unified-format diff via the `diff` package), and POST /api/v1/prompts/:id/restore-revision { revisionId, note? }. - proposals.ts — GET / POST /api/v1/proposals, GET /api/v1/proposals/:id, PUT for body updates, POST .../approve and POST .../reject, plus DELETE. ## What's changed - PromptService.create / update now record a ResourceRevision when the revision service is wired. Update auto-bumps patch on content change; authors can override via `--bump major|minor|patch` or `--semver X.Y.Z` on the CLI (forwarded into the PUT body). Best-effort: revision write failures are swallowed so the prompt save still succeeds (revision is audit, not source of truth). - PromptService.setProposalService registers a 'prompt' approval handler with the proposal service. Approval runs in a Prisma transaction: upsert prompt → record revision → update currentRevisionId → flip proposal status. semver bumps to 0.1.0 on first approval, patch thereafter. - New CLI flags on `mcpctl edit prompt`: --bump, --semver, --note. They're prompt-only (validated client-side); other resources reject them. - Aliases in shared.ts: `proposal`/`prop` → proposals, `revision`/`rev` → revisions. - diff dependency added to mcpd. ## Tests - src/mcpd/tests/utils/semver.test.ts — covers bump/compare/validate including numeric (not lex) semver compare and invalid-input fallback. - prompt-service.test.ts updated: makePrompt fixture now sets semver + agentId + currentRevisionId; updatePrompt assertion expects the auto-bumped patch in the same update call. - prompt-routes.test.ts updated symmetrically. ## RBAC `proposals` and `revisions` URL segments map to the existing `prompts` permission for now. PR-7 may split if a "reviewer" role becomes useful. ## Verification Full suite: 158 test files / 2127 tests green. `pnpm build` clean across all 6 workspace packages. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-07 00:38:35 +01:00
'@types/diff':
specifier: ^5.2.3
version: 5.2.3
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
'@types/dockerode':
specifier: ^4.0.1
version: 4.0.1
feat: add MCP server templates and deployment infrastructure Introduce a Helm-chart-like template system for MCP servers. Templates are YAML files in templates/ that get seeded into the DB on startup. Users can browse them with `mcpctl get templates`, inspect with `mcpctl describe template`, and instantiate with `mcpctl create server --from-template=`. Also adds Portainer deployment scripts, mcplocal systemd service, Streamable HTTP MCP endpoint, and RPM packaging for mcpctl-local. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 22:24:35 +00:00
'@types/js-yaml':
specifier: ^4.0.9
version: 4.0.9
feat: implement mcpd core server framework with Fastify Add Fastify server with config validation (Zod), health/healthz endpoints, auth middleware (Bearer token + session lookup), security plugins (CORS, Helmet, rate limiting), error handler, audit logging, and graceful shutdown. 36 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:22:01 +00:00
'@types/node':
specifier: ^25.3.0
version: 25.3.0
first commit
2026-02-21 03:10:39 +00:00
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
src/mcplocal:
dependencies:
'@fastify/cors':
specifier: ^10.0.0
version: 10.1.0
'@mcpctl/shared':
specifier: workspace:*
version: link:../shared
'@modelcontextprotocol/sdk':
specifier: ^1.0.0
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
version: 1.26.0(zod@4.1.8)
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
fastify:
specifier: ^5.0.0
version: 5.7.4
feat: eager vLLM warmup and smart page titles in paginate stage - Add warmup() to LlmProvider interface for eager subprocess startup - ManagedVllmProvider.warmup() starts vLLM in background on project load - ProviderRegistry.warmupAll() triggers all managed providers - NamedProvider proxies warmup() to inner provider - paginate stage generates LLM-powered descriptive page titles when available, cached by content hash, falls back to generic "Page N" - project-mcp-endpoint calls warmupAll() on router creation so vLLM is loading while the session initializes Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-03-03 19:07:39 +00:00
yaml:
specifier: ^2.8.2
version: 2.8.2
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
devDependencies:
'@types/node':
specifier: ^25.3.0
version: 25.3.0
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
src/pulumi:
devDependencies:
'@pulumi/pulumi':
specifier: ^3.0.0
version: 3.253.0(typescript@5.9.3)
first commit
2026-02-21 03:10:39 +00:00
src/shared:
dependencies:
zod:
specifier: ^3.24.0
version: 3.25.76
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
src/web:
dependencies:
'@monaco-editor/react':
specifier: ^4.7.0
version: 4.7.0(monaco-editor@0.55.1)(react-dom@19.2.5(react@19.2.5))(react@19.2.5)
feat(web): bold redesign — Tailwind v4 + shadcn-style primitives + Skills/Proposals/Revisions UI Phase 6 of the Skills + Revisions + Proposals work. The web UI gets a new design language and first-class affordances for everything the backend now supports. ## Visual direction - Tailwind v4 with custom @theme block (oklch tokens). Dark-mode-only (internal tool — light mode doubles QA surface). - Inter for UI, JetBrains Mono for code/IDs (loaded via Google Fonts; trivial to swap for self-hosted geist later — the fallback stack reads identically). - Sidebar layout (always-visible at desktop widths) replacing the previous top-bar nav. Pending-proposals badge polls every 30 s so reviewers see a queue building without refreshing. - Lucide icons throughout. - Spacing and radii on Tailwind defaults. Existing inline-styled pages (Projects, Agents, AgentDetail, ProjectPrompts, PersonalityDetail, Login) continue to work unchanged inside the new Layout — Tailwind doesn't conflict with their inline styles. A follow-up can migrate them incrementally. ## What's added ### Build infra (src/web/) - package.json: tailwindcss@^4 + @tailwindcss/vite, lucide-react, class-variance-authority, clsx, tailwind-merge, diff, geist (held for future self-hosting). - vite.config.ts: registers the @tailwindcss/vite plugin. - src/index.css: Tailwind import + @theme tokens + @layer base. - src/main.tsx: imports index.css. - src/lib/utils.ts: shadcn-style cn() helper. ### shadcn-style primitives (src/components/ui/) Hand-written rather than generated via `npx shadcn` so the repo doesn't depend on a CLI tool that needs interactive runtime: - button.tsx — variants: primary / secondary / ghost / danger / link; sizes: sm / md / lg / icon. - card.tsx — Card + Header/Title/Description/Content/Footer subparts. - badge.tsx — variants: default / info / success / warning / danger / outline. - input.tsx — Input + Textarea + Label. - tabs.tsx — no-dep accessible Tabs (no Radix needed for our use). - separator.tsx — h/v separator with role=separator. ### Diff component (src/components/Diff.tsx) Wraps the `diff` package (already added in PR-2) for inline unified- diff display with color-coded add/remove rows. Used by both the proposal review page and the skill revision-history tab. ### New pages (src/pages/) - Dashboard.tsx — at-a-glance home. Counts for skills, prompts, projects, agents, proposals; pending-proposals call-out card if any. - Skills.tsx — list view, separated into Global vs Project/Agent- scoped sections. - SkillDetail.tsx — name + semver + description; tabs for SKILL.md / Files / Metadata / History. History tab shows revisions with click-to-diff against the live body. - Proposals.tsx — queue with Pending/Approved/Rejected tabs. Pending count is highlighted in amber. - ProposalDetail.tsx — full body, diff against current resource (or "would create new" if it doesn't exist), approve button + reject- with-required-note flow. ### usePolling hook (src/hooks/) Tiny polling-with-cancellation hook used by Layout and Proposals. ### Layout rewrite (src/components/Layout.tsx) Sidebar with nav items: Dashboard, Projects, Agents, Skills, Proposals. Lucide icons. Active-route highlighting via NavLink. Pending-proposals warning badge on the Proposals item. ### Routes (src/App.tsx) New routes: /dashboard, /skills, /skills/:name, /proposals, /proposals/:id. Default redirects to /dashboard. ### API types (src/api.ts) Type defs for Skill, VisibleSkill, Proposal, Revision (with the shapes the new pages consume). ## Tests Existing 7 web tests still pass (Login + api). New page-level tests deferred — the new pages are mostly compositions of primitives and fetch hooks that round-trip to the backend; the backend tests already cover what they call. PR-7 polish can add render-and-click tests if coverage drift surfaces. ## Verification - `pnpm --filter @mcpctl/web build` clean, no warnings. - `pnpm test:run` whole monorepo: 162 test files / 2157 tests green. - Visual smoke deferred — needs a running mcpd to populate the fixtures. Manual smoke tested locally is the next step. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-07 17:54:55 +01:00
class-variance-authority:
specifier: ^0.7.1
version: 0.7.1
clsx:
specifier: ^2.1.1
version: 2.1.1
diff:
specifier: ^5.2.0
version: 5.2.2
geist:
specifier: ^1.5.1
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
version: 1.7.0(next@16.2.5(@babel/core@7.29.0)(@opentelemetry/api@1.9.1)(react-dom@19.2.5(react@19.2.5))(react@19.2.5))
feat(web): bold redesign — Tailwind v4 + shadcn-style primitives + Skills/Proposals/Revisions UI Phase 6 of the Skills + Revisions + Proposals work. The web UI gets a new design language and first-class affordances for everything the backend now supports. ## Visual direction - Tailwind v4 with custom @theme block (oklch tokens). Dark-mode-only (internal tool — light mode doubles QA surface). - Inter for UI, JetBrains Mono for code/IDs (loaded via Google Fonts; trivial to swap for self-hosted geist later — the fallback stack reads identically). - Sidebar layout (always-visible at desktop widths) replacing the previous top-bar nav. Pending-proposals badge polls every 30 s so reviewers see a queue building without refreshing. - Lucide icons throughout. - Spacing and radii on Tailwind defaults. Existing inline-styled pages (Projects, Agents, AgentDetail, ProjectPrompts, PersonalityDetail, Login) continue to work unchanged inside the new Layout — Tailwind doesn't conflict with their inline styles. A follow-up can migrate them incrementally. ## What's added ### Build infra (src/web/) - package.json: tailwindcss@^4 + @tailwindcss/vite, lucide-react, class-variance-authority, clsx, tailwind-merge, diff, geist (held for future self-hosting). - vite.config.ts: registers the @tailwindcss/vite plugin. - src/index.css: Tailwind import + @theme tokens + @layer base. - src/main.tsx: imports index.css. - src/lib/utils.ts: shadcn-style cn() helper. ### shadcn-style primitives (src/components/ui/) Hand-written rather than generated via `npx shadcn` so the repo doesn't depend on a CLI tool that needs interactive runtime: - button.tsx — variants: primary / secondary / ghost / danger / link; sizes: sm / md / lg / icon. - card.tsx — Card + Header/Title/Description/Content/Footer subparts. - badge.tsx — variants: default / info / success / warning / danger / outline. - input.tsx — Input + Textarea + Label. - tabs.tsx — no-dep accessible Tabs (no Radix needed for our use). - separator.tsx — h/v separator with role=separator. ### Diff component (src/components/Diff.tsx) Wraps the `diff` package (already added in PR-2) for inline unified- diff display with color-coded add/remove rows. Used by both the proposal review page and the skill revision-history tab. ### New pages (src/pages/) - Dashboard.tsx — at-a-glance home. Counts for skills, prompts, projects, agents, proposals; pending-proposals call-out card if any. - Skills.tsx — list view, separated into Global vs Project/Agent- scoped sections. - SkillDetail.tsx — name + semver + description; tabs for SKILL.md / Files / Metadata / History. History tab shows revisions with click-to-diff against the live body. - Proposals.tsx — queue with Pending/Approved/Rejected tabs. Pending count is highlighted in amber. - ProposalDetail.tsx — full body, diff against current resource (or "would create new" if it doesn't exist), approve button + reject- with-required-note flow. ### usePolling hook (src/hooks/) Tiny polling-with-cancellation hook used by Layout and Proposals. ### Layout rewrite (src/components/Layout.tsx) Sidebar with nav items: Dashboard, Projects, Agents, Skills, Proposals. Lucide icons. Active-route highlighting via NavLink. Pending-proposals warning badge on the Proposals item. ### Routes (src/App.tsx) New routes: /dashboard, /skills, /skills/:name, /proposals, /proposals/:id. Default redirects to /dashboard. ### API types (src/api.ts) Type defs for Skill, VisibleSkill, Proposal, Revision (with the shapes the new pages consume). ## Tests Existing 7 web tests still pass (Login + api). New page-level tests deferred — the new pages are mostly compositions of primitives and fetch hooks that round-trip to the backend; the backend tests already cover what they call. PR-7 polish can add render-and-click tests if coverage drift surfaces. ## Verification - `pnpm --filter @mcpctl/web build` clean, no warnings. - `pnpm test:run` whole monorepo: 162 test files / 2157 tests green. - Visual smoke deferred — needs a running mcpd to populate the fixtures. Manual smoke tested locally is the next step. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-07 17:54:55 +01:00
lucide-react:
specifier: ^0.487.0
version: 0.487.0(react@19.2.5)
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
react:
specifier: ^19.2.5
version: 19.2.5
react-dom:
specifier: ^19.2.5
version: 19.2.5(react@19.2.5)
react-router-dom:
specifier: ^7.7.0
version: 7.14.2(react-dom@19.2.5(react@19.2.5))(react@19.2.5)
feat(web): bold redesign — Tailwind v4 + shadcn-style primitives + Skills/Proposals/Revisions UI Phase 6 of the Skills + Revisions + Proposals work. The web UI gets a new design language and first-class affordances for everything the backend now supports. ## Visual direction - Tailwind v4 with custom @theme block (oklch tokens). Dark-mode-only (internal tool — light mode doubles QA surface). - Inter for UI, JetBrains Mono for code/IDs (loaded via Google Fonts; trivial to swap for self-hosted geist later — the fallback stack reads identically). - Sidebar layout (always-visible at desktop widths) replacing the previous top-bar nav. Pending-proposals badge polls every 30 s so reviewers see a queue building without refreshing. - Lucide icons throughout. - Spacing and radii on Tailwind defaults. Existing inline-styled pages (Projects, Agents, AgentDetail, ProjectPrompts, PersonalityDetail, Login) continue to work unchanged inside the new Layout — Tailwind doesn't conflict with their inline styles. A follow-up can migrate them incrementally. ## What's added ### Build infra (src/web/) - package.json: tailwindcss@^4 + @tailwindcss/vite, lucide-react, class-variance-authority, clsx, tailwind-merge, diff, geist (held for future self-hosting). - vite.config.ts: registers the @tailwindcss/vite plugin. - src/index.css: Tailwind import + @theme tokens + @layer base. - src/main.tsx: imports index.css. - src/lib/utils.ts: shadcn-style cn() helper. ### shadcn-style primitives (src/components/ui/) Hand-written rather than generated via `npx shadcn` so the repo doesn't depend on a CLI tool that needs interactive runtime: - button.tsx — variants: primary / secondary / ghost / danger / link; sizes: sm / md / lg / icon. - card.tsx — Card + Header/Title/Description/Content/Footer subparts. - badge.tsx — variants: default / info / success / warning / danger / outline. - input.tsx — Input + Textarea + Label. - tabs.tsx — no-dep accessible Tabs (no Radix needed for our use). - separator.tsx — h/v separator with role=separator. ### Diff component (src/components/Diff.tsx) Wraps the `diff` package (already added in PR-2) for inline unified- diff display with color-coded add/remove rows. Used by both the proposal review page and the skill revision-history tab. ### New pages (src/pages/) - Dashboard.tsx — at-a-glance home. Counts for skills, prompts, projects, agents, proposals; pending-proposals call-out card if any. - Skills.tsx — list view, separated into Global vs Project/Agent- scoped sections. - SkillDetail.tsx — name + semver + description; tabs for SKILL.md / Files / Metadata / History. History tab shows revisions with click-to-diff against the live body. - Proposals.tsx — queue with Pending/Approved/Rejected tabs. Pending count is highlighted in amber. - ProposalDetail.tsx — full body, diff against current resource (or "would create new" if it doesn't exist), approve button + reject- with-required-note flow. ### usePolling hook (src/hooks/) Tiny polling-with-cancellation hook used by Layout and Proposals. ### Layout rewrite (src/components/Layout.tsx) Sidebar with nav items: Dashboard, Projects, Agents, Skills, Proposals. Lucide icons. Active-route highlighting via NavLink. Pending-proposals warning badge on the Proposals item. ### Routes (src/App.tsx) New routes: /dashboard, /skills, /skills/:name, /proposals, /proposals/:id. Default redirects to /dashboard. ### API types (src/api.ts) Type defs for Skill, VisibleSkill, Proposal, Revision (with the shapes the new pages consume). ## Tests Existing 7 web tests still pass (Login + api). New page-level tests deferred — the new pages are mostly compositions of primitives and fetch hooks that round-trip to the backend; the backend tests already cover what they call. PR-7 polish can add render-and-click tests if coverage drift surfaces. ## Verification - `pnpm --filter @mcpctl/web build` clean, no warnings. - `pnpm test:run` whole monorepo: 162 test files / 2157 tests green. - Visual smoke deferred — needs a running mcpd to populate the fixtures. Manual smoke tested locally is the next step. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-07 17:54:55 +01:00
tailwind-merge:
specifier: ^3.3.1
version: 3.5.0
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
devDependencies:
feat(web): bold redesign — Tailwind v4 + shadcn-style primitives + Skills/Proposals/Revisions UI Phase 6 of the Skills + Revisions + Proposals work. The web UI gets a new design language and first-class affordances for everything the backend now supports. ## Visual direction - Tailwind v4 with custom @theme block (oklch tokens). Dark-mode-only (internal tool — light mode doubles QA surface). - Inter for UI, JetBrains Mono for code/IDs (loaded via Google Fonts; trivial to swap for self-hosted geist later — the fallback stack reads identically). - Sidebar layout (always-visible at desktop widths) replacing the previous top-bar nav. Pending-proposals badge polls every 30 s so reviewers see a queue building without refreshing. - Lucide icons throughout. - Spacing and radii on Tailwind defaults. Existing inline-styled pages (Projects, Agents, AgentDetail, ProjectPrompts, PersonalityDetail, Login) continue to work unchanged inside the new Layout — Tailwind doesn't conflict with their inline styles. A follow-up can migrate them incrementally. ## What's added ### Build infra (src/web/) - package.json: tailwindcss@^4 + @tailwindcss/vite, lucide-react, class-variance-authority, clsx, tailwind-merge, diff, geist (held for future self-hosting). - vite.config.ts: registers the @tailwindcss/vite plugin. - src/index.css: Tailwind import + @theme tokens + @layer base. - src/main.tsx: imports index.css. - src/lib/utils.ts: shadcn-style cn() helper. ### shadcn-style primitives (src/components/ui/) Hand-written rather than generated via `npx shadcn` so the repo doesn't depend on a CLI tool that needs interactive runtime: - button.tsx — variants: primary / secondary / ghost / danger / link; sizes: sm / md / lg / icon. - card.tsx — Card + Header/Title/Description/Content/Footer subparts. - badge.tsx — variants: default / info / success / warning / danger / outline. - input.tsx — Input + Textarea + Label. - tabs.tsx — no-dep accessible Tabs (no Radix needed for our use). - separator.tsx — h/v separator with role=separator. ### Diff component (src/components/Diff.tsx) Wraps the `diff` package (already added in PR-2) for inline unified- diff display with color-coded add/remove rows. Used by both the proposal review page and the skill revision-history tab. ### New pages (src/pages/) - Dashboard.tsx — at-a-glance home. Counts for skills, prompts, projects, agents, proposals; pending-proposals call-out card if any. - Skills.tsx — list view, separated into Global vs Project/Agent- scoped sections. - SkillDetail.tsx — name + semver + description; tabs for SKILL.md / Files / Metadata / History. History tab shows revisions with click-to-diff against the live body. - Proposals.tsx — queue with Pending/Approved/Rejected tabs. Pending count is highlighted in amber. - ProposalDetail.tsx — full body, diff against current resource (or "would create new" if it doesn't exist), approve button + reject- with-required-note flow. ### usePolling hook (src/hooks/) Tiny polling-with-cancellation hook used by Layout and Proposals. ### Layout rewrite (src/components/Layout.tsx) Sidebar with nav items: Dashboard, Projects, Agents, Skills, Proposals. Lucide icons. Active-route highlighting via NavLink. Pending-proposals warning badge on the Proposals item. ### Routes (src/App.tsx) New routes: /dashboard, /skills, /skills/:name, /proposals, /proposals/:id. Default redirects to /dashboard. ### API types (src/api.ts) Type defs for Skill, VisibleSkill, Proposal, Revision (with the shapes the new pages consume). ## Tests Existing 7 web tests still pass (Login + api). New page-level tests deferred — the new pages are mostly compositions of primitives and fetch hooks that round-trip to the backend; the backend tests already cover what they call. PR-7 polish can add render-and-click tests if coverage drift surfaces. ## Verification - `pnpm --filter @mcpctl/web build` clean, no warnings. - `pnpm test:run` whole monorepo: 162 test files / 2157 tests green. - Visual smoke deferred — needs a running mcpd to populate the fixtures. Manual smoke tested locally is the next step. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-07 17:54:55 +01:00
'@tailwindcss/vite':
specifier: ^4.1.16
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
version: 4.2.4(vite@7.3.1(@types/node@25.3.0)(jiti@2.7.0)(lightningcss@1.32.0)(tsx@4.21.0)(yaml@2.9.0))
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
'@testing-library/jest-dom':
specifier: ^6.7.0
version: 6.9.1
'@testing-library/react':
specifier: ^16.3.0
version: 16.3.2(@testing-library/dom@10.4.1)(@types/react-dom@19.2.3(@types/react@19.2.14))(@types/react@19.2.14)(react-dom@19.2.5(react@19.2.5))(react@19.2.5)
feat(web): bold redesign — Tailwind v4 + shadcn-style primitives + Skills/Proposals/Revisions UI Phase 6 of the Skills + Revisions + Proposals work. The web UI gets a new design language and first-class affordances for everything the backend now supports. ## Visual direction - Tailwind v4 with custom @theme block (oklch tokens). Dark-mode-only (internal tool — light mode doubles QA surface). - Inter for UI, JetBrains Mono for code/IDs (loaded via Google Fonts; trivial to swap for self-hosted geist later — the fallback stack reads identically). - Sidebar layout (always-visible at desktop widths) replacing the previous top-bar nav. Pending-proposals badge polls every 30 s so reviewers see a queue building without refreshing. - Lucide icons throughout. - Spacing and radii on Tailwind defaults. Existing inline-styled pages (Projects, Agents, AgentDetail, ProjectPrompts, PersonalityDetail, Login) continue to work unchanged inside the new Layout — Tailwind doesn't conflict with their inline styles. A follow-up can migrate them incrementally. ## What's added ### Build infra (src/web/) - package.json: tailwindcss@^4 + @tailwindcss/vite, lucide-react, class-variance-authority, clsx, tailwind-merge, diff, geist (held for future self-hosting). - vite.config.ts: registers the @tailwindcss/vite plugin. - src/index.css: Tailwind import + @theme tokens + @layer base. - src/main.tsx: imports index.css. - src/lib/utils.ts: shadcn-style cn() helper. ### shadcn-style primitives (src/components/ui/) Hand-written rather than generated via `npx shadcn` so the repo doesn't depend on a CLI tool that needs interactive runtime: - button.tsx — variants: primary / secondary / ghost / danger / link; sizes: sm / md / lg / icon. - card.tsx — Card + Header/Title/Description/Content/Footer subparts. - badge.tsx — variants: default / info / success / warning / danger / outline. - input.tsx — Input + Textarea + Label. - tabs.tsx — no-dep accessible Tabs (no Radix needed for our use). - separator.tsx — h/v separator with role=separator. ### Diff component (src/components/Diff.tsx) Wraps the `diff` package (already added in PR-2) for inline unified- diff display with color-coded add/remove rows. Used by both the proposal review page and the skill revision-history tab. ### New pages (src/pages/) - Dashboard.tsx — at-a-glance home. Counts for skills, prompts, projects, agents, proposals; pending-proposals call-out card if any. - Skills.tsx — list view, separated into Global vs Project/Agent- scoped sections. - SkillDetail.tsx — name + semver + description; tabs for SKILL.md / Files / Metadata / History. History tab shows revisions with click-to-diff against the live body. - Proposals.tsx — queue with Pending/Approved/Rejected tabs. Pending count is highlighted in amber. - ProposalDetail.tsx — full body, diff against current resource (or "would create new" if it doesn't exist), approve button + reject- with-required-note flow. ### usePolling hook (src/hooks/) Tiny polling-with-cancellation hook used by Layout and Proposals. ### Layout rewrite (src/components/Layout.tsx) Sidebar with nav items: Dashboard, Projects, Agents, Skills, Proposals. Lucide icons. Active-route highlighting via NavLink. Pending-proposals warning badge on the Proposals item. ### Routes (src/App.tsx) New routes: /dashboard, /skills, /skills/:name, /proposals, /proposals/:id. Default redirects to /dashboard. ### API types (src/api.ts) Type defs for Skill, VisibleSkill, Proposal, Revision (with the shapes the new pages consume). ## Tests Existing 7 web tests still pass (Login + api). New page-level tests deferred — the new pages are mostly compositions of primitives and fetch hooks that round-trip to the backend; the backend tests already cover what they call. PR-7 polish can add render-and-click tests if coverage drift surfaces. ## Verification - `pnpm --filter @mcpctl/web build` clean, no warnings. - `pnpm test:run` whole monorepo: 162 test files / 2157 tests green. - Visual smoke deferred — needs a running mcpd to populate the fixtures. Manual smoke tested locally is the next step. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-07 17:54:55 +01:00
'@types/diff':
specifier: ^5.2.3
version: 5.2.3
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
'@types/react':
specifier: ^19.2.14
version: 19.2.14
'@types/react-dom':
specifier: ^19.2.0
version: 19.2.3(@types/react@19.2.14)
'@vitejs/plugin-react':
specifier: ^5.1.0
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
version: 5.2.0(vite@7.3.1(@types/node@25.3.0)(jiti@2.7.0)(lightningcss@1.32.0)(tsx@4.21.0)(yaml@2.9.0))
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
jsdom:
specifier: ^28.0.0
version: 28.1.0
feat(web): bold redesign — Tailwind v4 + shadcn-style primitives + Skills/Proposals/Revisions UI Phase 6 of the Skills + Revisions + Proposals work. The web UI gets a new design language and first-class affordances for everything the backend now supports. ## Visual direction - Tailwind v4 with custom @theme block (oklch tokens). Dark-mode-only (internal tool — light mode doubles QA surface). - Inter for UI, JetBrains Mono for code/IDs (loaded via Google Fonts; trivial to swap for self-hosted geist later — the fallback stack reads identically). - Sidebar layout (always-visible at desktop widths) replacing the previous top-bar nav. Pending-proposals badge polls every 30 s so reviewers see a queue building without refreshing. - Lucide icons throughout. - Spacing and radii on Tailwind defaults. Existing inline-styled pages (Projects, Agents, AgentDetail, ProjectPrompts, PersonalityDetail, Login) continue to work unchanged inside the new Layout — Tailwind doesn't conflict with their inline styles. A follow-up can migrate them incrementally. ## What's added ### Build infra (src/web/) - package.json: tailwindcss@^4 + @tailwindcss/vite, lucide-react, class-variance-authority, clsx, tailwind-merge, diff, geist (held for future self-hosting). - vite.config.ts: registers the @tailwindcss/vite plugin. - src/index.css: Tailwind import + @theme tokens + @layer base. - src/main.tsx: imports index.css. - src/lib/utils.ts: shadcn-style cn() helper. ### shadcn-style primitives (src/components/ui/) Hand-written rather than generated via `npx shadcn` so the repo doesn't depend on a CLI tool that needs interactive runtime: - button.tsx — variants: primary / secondary / ghost / danger / link; sizes: sm / md / lg / icon. - card.tsx — Card + Header/Title/Description/Content/Footer subparts. - badge.tsx — variants: default / info / success / warning / danger / outline. - input.tsx — Input + Textarea + Label. - tabs.tsx — no-dep accessible Tabs (no Radix needed for our use). - separator.tsx — h/v separator with role=separator. ### Diff component (src/components/Diff.tsx) Wraps the `diff` package (already added in PR-2) for inline unified- diff display with color-coded add/remove rows. Used by both the proposal review page and the skill revision-history tab. ### New pages (src/pages/) - Dashboard.tsx — at-a-glance home. Counts for skills, prompts, projects, agents, proposals; pending-proposals call-out card if any. - Skills.tsx — list view, separated into Global vs Project/Agent- scoped sections. - SkillDetail.tsx — name + semver + description; tabs for SKILL.md / Files / Metadata / History. History tab shows revisions with click-to-diff against the live body. - Proposals.tsx — queue with Pending/Approved/Rejected tabs. Pending count is highlighted in amber. - ProposalDetail.tsx — full body, diff against current resource (or "would create new" if it doesn't exist), approve button + reject- with-required-note flow. ### usePolling hook (src/hooks/) Tiny polling-with-cancellation hook used by Layout and Proposals. ### Layout rewrite (src/components/Layout.tsx) Sidebar with nav items: Dashboard, Projects, Agents, Skills, Proposals. Lucide icons. Active-route highlighting via NavLink. Pending-proposals warning badge on the Proposals item. ### Routes (src/App.tsx) New routes: /dashboard, /skills, /skills/:name, /proposals, /proposals/:id. Default redirects to /dashboard. ### API types (src/api.ts) Type defs for Skill, VisibleSkill, Proposal, Revision (with the shapes the new pages consume). ## Tests Existing 7 web tests still pass (Login + api). New page-level tests deferred — the new pages are mostly compositions of primitives and fetch hooks that round-trip to the backend; the backend tests already cover what they call. PR-7 polish can add render-and-click tests if coverage drift surfaces. ## Verification - `pnpm --filter @mcpctl/web build` clean, no warnings. - `pnpm test:run` whole monorepo: 162 test files / 2157 tests green. - Visual smoke deferred — needs a running mcpd to populate the fixtures. Manual smoke tested locally is the next step. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-07 17:54:55 +01:00
tailwindcss:
specifier: ^4.1.16
version: 4.2.4
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
vite:
specifier: ^7.2.0
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
version: 7.3.1(@types/node@25.3.0)(jiti@2.7.0)(lightningcss@1.32.0)(tsx@4.21.0)(yaml@2.9.0)
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
first commit
2026-02-21 03:10:39 +00:00
packages:
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
'@acemir/cssom@0.9.31':
resolution: {integrity: sha512-ZnR3GSaH+/vJ0YlHau21FjfLYjMpYVIzTD8M8vIEQvIGxeOXyXdzCI140rrCY862p/C/BbzWsjc1dgnM9mkoTA==}
'@adobe/css-tools@4.4.4':
resolution: {integrity: sha512-Elp+iwUx5rN5+Y8xLt5/GRoG20WGoDCQ/1Fb+1LiGtvwbDavuSk0jhD/eZdckHAuzcDzccnkv+rEjyWfRx18gg==}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
'@ai-sdk/provider@3.0.8':
resolution: {integrity: sha512-oGMAgGoQdBXbZqNG0Ze56CHjDZ1IDYOwGYxYjO5KLSlz5HiNQ9udIXsPZ61VWaHGZ5XW/jyjmr6t2xz2jGVwbQ==}
engines: {node: '>=18'}
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
'@alcalzone/ansi-tokenize@0.2.5':
resolution: {integrity: sha512-3NX/MpTdroi0aKz134A6RC2Gb2iXVECN4QaAXnvCIxxIm3C3AVB1mkUe8NaaiyvOpDfsrqWhYtj+Q6a62RrTsw==}
engines: {node: '>=18'}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
'@ampproject/remapping@2.3.0':
resolution: {integrity: sha512-30iZtAPgz+LTIYoeivqYo853f02jBYSd5uGnGpkFV0M3xOt9aN73erkgYAmZU43x4VfqcnLxW9Kpg3R5LC4YYw==}
engines: {node: '>=6.0.0'}
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
'@anthropic-ai/sdk@0.91.1':
resolution: {integrity: sha512-LAmu761tSN9r66ixvmciswUj/ZC+1Q4iAfpedTfSVLeswRwnY3n2Nb6Tsk+cLPP28aLOPWeMgIuTuCcMC6W/iw==}
hasBin: true
peerDependencies:
zod: ^3.25.0 || ^4.0.0
peerDependenciesMeta:
zod:
optional: true
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
'@asamuzakjp/css-color@5.1.11':
resolution: {integrity: sha512-KVw6qIiCTUQhByfTd78h2yD1/00waTmm9uy/R7Ck/ctUyAPj+AEDLkQIdJW0T8+qGgj3j5bpNKK7Q3G+LedJWg==}
engines: {node: ^20.19.0 || ^22.12.0 || >=24.0.0}
'@asamuzakjp/dom-selector@6.8.1':
resolution: {integrity: sha512-MvRz1nCqW0fsy8Qz4dnLIvhOlMzqDVBabZx6lH+YywFDdjXhMY37SmpV1XFX3JzG5GWHn63j6HX6QPr3lZXHvQ==}
'@asamuzakjp/generational-cache@1.0.1':
resolution: {integrity: sha512-wajfB8KqzMCN2KGNFdLkReeHncd0AslUSrvHVvvYWuU8ghncRJoA50kT3zP9MVL0+9g4/67H+cdvBskj9THPzg==}
engines: {node: ^20.19.0 || ^22.12.0 || >=24.0.0}
'@asamuzakjp/nwsapi@2.3.9':
resolution: {integrity: sha512-n8GuYSrI9bF7FFZ/SjhwevlHc8xaVlb/7HmHelnc/PZXBD2ZR49NnN9sMMuDdEGPeeRQ5d0hqlSlEpgCX3Wl0Q==}
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
'@aws-crypto/sha256-browser@5.2.0':
resolution: {integrity: sha512-AXfN/lGotSQwu6HNcEsIASo7kWXZ5HYWvfOmSNKDsEqC4OashTp8alTmaz+F7TC2L083SFv5RdB+qU3Vs1kZqw==}
'@aws-crypto/sha256-js@5.2.0':
resolution: {integrity: sha512-FFQQyu7edu4ufvIZ+OadFpHHOt+eSTBaYaki44c+akjg7qZg9oOQeLlk77F6tSYqjDAFClrHJk9tMf0HdVyOvA==}
engines: {node: '>=16.0.0'}
'@aws-crypto/supports-web-crypto@5.2.0':
resolution: {integrity: sha512-iAvUotm021kM33eCdNfwIN//F77/IADDSs58i+MDaOqFrVjZo9bAal0NK7HurRuWLLpF1iLX7gbWrjHjeo+YFg==}
'@aws-crypto/util@5.2.0':
resolution: {integrity: sha512-4RkU9EsI6ZpBve5fseQlGNUWKMa1RLPQ1dnjnQoe07ldfIzcsGb5hC5W0Dm7u423KWzawlrpbjXBrXCEv9zazQ==}
'@aws-sdk/client-bedrock-runtime@3.1048.0':
resolution: {integrity: sha512-u+NT61JZEkRFtpL0CAw1N1dwxnaLgwVXQl/zjJxTGgLyS/jTIdg2SdoEoCTHxgDyCnqa1HEi9QOoE9/pYRNpOQ==}
engines: {node: '>=20.0.0'}
'@aws-sdk/core@3.977.6':
resolution: {integrity: sha512-QiaJV4/zDrB4ZY2mfeSXSzSTc36W16sZXcGz+SPFk0CJ26gziO0cS+4LjJUMAbdeeBOvS0k0Aq1cZpfGdUXxSw==}
engines: {node: '>=20.0.0'}
'@aws-sdk/credential-provider-env@3.972.67':
resolution: {integrity: sha512-rcIpk5kxUqDaaNa6Xk23pQ6ViY7jlqzmfFWCahQcBT97ddXaXYYwzCen9Tz1Jvo6aJft6wDl5bN44/Jw5B4oLA==}
engines: {node: '>=20.0.0'}
'@aws-sdk/credential-provider-http@3.972.69':
resolution: {integrity: sha512-nggwJtZ4eeNsUw5IeWBMXsi1ryct5idi0K+/SCRF3kybLubOMaNTb3XCihXpWMiVpyzyPeIrl0zTkzhBH9porA==}
engines: {node: '>=20.0.0'}
'@aws-sdk/credential-provider-ini@3.973.12':
resolution: {integrity: sha512-pNEf/OeyN5X3VmLKlgSO6TqaWmW10CvI3TfwL1XhsuhYjSLT2VDaxFnCPHnOeQXSaFisMX4jNhpETriqN8DOmg==}
engines: {node: '>=20.0.0'}
'@aws-sdk/credential-provider-login@3.972.74':
resolution: {integrity: sha512-0AQfDcf99TNmqVKv0owHrw/TQs6i4ZE5t9qmz6NvO53bE/sA/tpXhXL9AAcEP1qHc6Zzjd1UMb69+/9zdhvY3g==}
engines: {node: '>=20.0.0'}
'@aws-sdk/credential-provider-node@3.972.78':
resolution: {integrity: sha512-OgPAnfvbGAMWac6yvxJ1ihslrvDpPVwR68D2csospdNCCyPvHk9JLzYKwz48SNiS1T2znDwHauywRKRFfpyYng==}
engines: {node: '>=20.0.0'}
'@aws-sdk/credential-provider-process@3.972.67':
resolution: {integrity: sha512-IlUEejorGTWKb4/Dm7K5Yw4QxUmXLThLhrvBmzVBqZFTbW72cv9LTcITmo1dsnYriALE4h68mOq4LB99x6sQ7Q==}
engines: {node: '>=20.0.0'}
'@aws-sdk/credential-provider-sso@3.973.11':
resolution: {integrity: sha512-gAQBkBZxUB84d71+pPcI9L+jh2ujhuAVxc/4FgGiWFDjkPBlMKxzd5XDtkSXTFX8Ro7ansnT88+XadasxMeCRw==}
engines: {node: '>=20.0.0'}
'@aws-sdk/credential-provider-web-identity@3.972.73':
resolution: {integrity: sha512-SnlEmQa6SjOgs6iOPLUQl1Eyq4AKiAdPQlkOhFhqNfDtDCwibMGvL6QlkSmf3o6vAUSImzdPCxowT5dfQUZP1A==}
engines: {node: '>=20.0.0'}
'@aws-sdk/eventstream-handler-node@3.972.31':
resolution: {integrity: sha512-/BRzvkp46mF6eXBL/l9WKPQQfifLlUPaWli6n9/T/WDLUg8he7TCyuNFnk6RvHP5j5W/kMj5Gxw7W778LJaXDA==}
engines: {node: '>=20.0.0'}
'@aws-sdk/middleware-eventstream@3.972.26':
resolution: {integrity: sha512-2eIvouTZoxPu5ClHY6ij13De1yhY8Rmllt0dlGeBNXX3wmR7fU1pvMCGb50fKm1GxcuntWK0t1T0cMjTyDoUQA==}
engines: {node: '>=20.0.0'}
'@aws-sdk/middleware-websocket@3.972.49':
resolution: {integrity: sha512-wGYJvu1/stdWuzGcNyh44u8aY7ArU8XFrMesBlzIYn70HWVCRBNEngQexDuPIMu0NEgsKGKmTc0uvnS/bF7KWw==}
engines: {node: '>= 14.0.0'}
'@aws-sdk/nested-clients@3.997.41':
resolution: {integrity: sha512-RDHqPGQWlF6tatA/Tp3rg6oIwtgN9IVderxE+9av2Y93Dfyu+mO1hZ5Bu2jpfZg2rwdNbsssnwM+sLafIczMlQ==}
engines: {node: '>=20.0.0'}
'@aws-sdk/signature-v4-multi-region@3.996.43':
resolution: {integrity: sha512-lKekx8bLBXSv4O+cslk9Zfnw2XKSkWBs3uWL5QGhH2ZAQfNS7FE0vcSSN2vD/AhxX54ZTywWxR4STThoeOXlBA==}
engines: {node: '>=20.0.0'}
'@aws-sdk/token-providers@3.1048.0':
resolution: {integrity: sha512-k0y/GcuesuSfWyUM0WamrGyeZmltRYaPbHO82UDA6mZ/doB+FOHKutikPAtSXMn/hDz970cF+iRuuiYO9VEbAA==}
engines: {node: '>=20.0.0'}
'@aws-sdk/token-providers@3.1103.0':
resolution: {integrity: sha512-N4wy26MNn31ItGVHYHPrEuCIFY4MBBjC+C5v1lJKqIUSA7OZBdhleCY53zCCrXn27hsk7YNOaTuhQu807S4AfQ==}
engines: {node: '>=20.0.0'}
'@aws-sdk/types@3.974.2':
resolution: {integrity: sha512-3W6IUtSxFbH6X7Wb7DzGCV5QiFQsd0g8bOfntpmDxQlzBoKWUMBu/JPQR0DwkE+Hpnxd6db1tXbOwdeHddG6cA==}
engines: {node: '>=20.0.0'}
'@aws-sdk/util-locate-window@3.965.8':
resolution: {integrity: sha512-uUbMs1cBZPafD0ohUj6EwNf0fPZ534NvBxHox4hjX+0Rxq5paSYUem7+hi833pYrzrcnBATKIYpR02MDXT5M9g==}
engines: {node: '>=20.0.0'}
'@aws-sdk/xml-builder@3.972.37':
resolution: {integrity: sha512-zKq4HQum8JwDyEuyfuI4bbiAcU0KxP6qy+9PR/IsR92IyE/DaBAikzAS50tjxip4bqIIANpCcG+Yyj6CVhXupg==}
engines: {node: '>=20.0.0'}
'@aws/lambda-invoke-store@0.3.0':
resolution: {integrity: sha512-sl4Bm6yiMNYrZKkqqDFWN0UfnWhlS8ivKxrYl+6t0gCLrqr8y3B2IqZZbFRkfaVVp7C/baApyh71P+LeE1A2sQ==}
engines: {node: '>=18.0.0'}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
'@babel/code-frame@7.29.0':
resolution: {integrity: sha512-9NhCeYjq9+3uxgdtp20LSiJXJvN0FeCtNGpJxuMFZ1Kv3cWUNb6DOhJwUvcVCzKGR66cw4njwM6hrJLqgOwbcw==}
engines: {node: '>=6.9.0'}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
'@babel/code-frame@7.29.7':
resolution: {integrity: sha512-Aup7aUOfpbAUg2ROOJN6Iw5f9DMBlzu0mIkm/malLQFN/YQgO48wCj0Kxa3sEHJvPVFg7siR+qRInwXd2qhQKw==}
engines: {node: '>=6.9.0'}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
'@babel/compat-data@7.29.0':
resolution: {integrity: sha512-T1NCJqT/j9+cn8fvkt7jtwbLBfLC/1y1c7NtCeXFRgzGTsafi68MRv8yzkYSapBnFA6L3U2VSc02ciDzoAJhJg==}
engines: {node: '>=6.9.0'}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
'@babel/core@7.28.0':
resolution: {integrity: sha512-UlLAnTPrFdNGoFtbSXwcGFQBtQZJCNjaN6hQNP3UPvuNXT1i82N26KL3dZeIpNalWywr9IuQuncaAfUaS1g6sQ==}
engines: {node: '>=6.9.0'}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
'@babel/core@7.29.0':
resolution: {integrity: sha512-CGOfOJqWjg2qW/Mb6zNsDm+u5vFQ8DxXfbM09z69p5Z6+mE1ikP2jUXw+j42Pf1XTYED2Rni5f95npYeuwMDQA==}
engines: {node: '>=6.9.0'}
'@babel/generator@7.29.1':
resolution: {integrity: sha512-qsaF+9Qcm2Qv8SRIMMscAvG4O3lJ0F1GuMo5HR/Bp02LopNgnZBC/EkbevHFeGs4ls/oPz9v+Bsmzbkbe+0dUw==}
engines: {node: '>=6.9.0'}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
'@babel/generator@7.29.8':
resolution: {integrity: sha512-gZbepsdh3WDtgZKWL+vTPh71LSBrm/Y4/QDZBVCcYfmeTEEuoOYwlSy+G1StfJg+/Zy550u/3TATbm7qDbbMtg==}
engines: {node: '>=6.9.0'}
'@babel/helper-annotate-as-pure@7.29.7':
resolution: {integrity: sha512-OoK6239jHPuSQOoS0kfTVKn0b/rVTk0seKq4Gd2UMLtmOVLjDC0ki3e+c90Trqv2gMfvJFqkiljrr568+qddiw==}
engines: {node: '>=6.9.0'}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
'@babel/helper-compilation-targets@7.28.6':
resolution: {integrity: sha512-JYtls3hqi15fcx5GaSNL7SCTJ2MNmjrkHXg4FSpOA/grxK8KwyZ5bubHsCq8FXCkua6xhuaaBit+3b7+VZRfcA==}
engines: {node: '>=6.9.0'}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
'@babel/helper-create-class-features-plugin@7.29.7':
resolution: {integrity: sha512-IY3ZD9Tmooqr3TUhc3DUWxiuo8xx1DWLhd5M7hQ+ZWJamqM2BbalrBJb2MisSLoYorOj75U03qULCxQTY9r3hg==}
engines: {node: '>=6.9.0'}
peerDependencies:
'@babel/core': ^7.0.0
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
'@babel/helper-globals@7.28.0':
resolution: {integrity: sha512-+W6cISkXFa1jXsDEdYA8HeevQT/FULhxzR99pxphltZcVaugps53THCeiWA8SguxxpSp3gKPiuYfSWopkLQ4hw==}
engines: {node: '>=6.9.0'}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
'@babel/helper-globals@7.29.7':
resolution: {integrity: sha512-3nQVUAtvkKH9zahfWgw96Jc/uFOmjACE1kQz82E2lqWmHBgjzbNlsC22nuQTfahmWeQtTq5nQ/4Nnd2A1wj4zA==}
engines: {node: '>=6.9.0'}
'@babel/helper-member-expression-to-functions@7.29.7':
resolution: {integrity: sha512-j+7JYmk1JYDtACIGj0QJqqWZjoUpMoEikQGADMaHgCMCSDqd2+P32rfcibUNrGOMWrlzK1WJBdxrB3JJQZwWtg==}
engines: {node: '>=6.9.0'}
'@babel/helper-module-imports@7.18.6':
resolution: {integrity: sha512-0NFvs3VkuSYbFi1x2Vd6tKrywq+z/cLeYC/RJNFrIX/30Bf5aiGYbtvGXolEktzJH8o5E5KJ3tT+nkxuuZFVlA==}
engines: {node: '>=6.9.0'}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
'@babel/helper-module-imports@7.28.6':
resolution: {integrity: sha512-l5XkZK7r7wa9LucGw9LwZyyCUscb4x37JWTPz7swwFE/0FMQAGpiWUZn8u9DzkSBWEcK25jmvubfpw2dnAMdbw==}
engines: {node: '>=6.9.0'}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
'@babel/helper-module-imports@7.29.7':
resolution: {integrity: sha512-ejHwrQQYcm9xnTivShn2IDOlIzInN34AXskvq9QicvCtEzq1Vzclu/tKF8Jq1Cg8JG2GL6/EmjgsCT7lXepE3g==}
engines: {node: '>=6.9.0'}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
'@babel/helper-module-transforms@7.28.6':
resolution: {integrity: sha512-67oXFAYr2cDLDVGLXTEABjdBJZ6drElUSI7WKp70NrpyISso3plG9SAGEF6y7zbha/wOzUByWWTJvEDVNIUGcA==}
engines: {node: '>=6.9.0'}
peerDependencies:
'@babel/core': ^7.0.0
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
'@babel/helper-module-transforms@7.29.7':
resolution: {integrity: sha512-UPUVSyXbOh627KiCIGQSgwWzGeBKLkaJ9PJEdrngIwMSzxLR4jS4+f1f1jb7VzBbg8nFLaYotvVPFCTqdrmTAg==}
engines: {node: '>=6.9.0'}
peerDependencies:
'@babel/core': ^7.0.0
'@babel/helper-optimise-call-expression@7.29.7':
resolution: {integrity: sha512-+kmGVjcT9RGYzoDwdwEqEvGgKe3BYq+O1iGzjFubaNgZHwYHP6lsF2Yghf4kEuv9BV7tYDZ913aBW9am6YKong==}
engines: {node: '>=6.9.0'}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
'@babel/helper-plugin-utils@7.28.6':
resolution: {integrity: sha512-S9gzZ/bz83GRysI7gAD4wPT/AI3uCnY+9xn+Mx/KPs2JwHJIz1W8PZkg2cqyt3RNOBM8ejcXhV6y8Og7ly/Dug==}
engines: {node: '>=6.9.0'}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
'@babel/helper-plugin-utils@7.29.7':
resolution: {integrity: sha512-G7sHYigPY17oO5SYWnfD/0MTBwVR781S/JI643e/JhUYgVgWE/61SoW3NH9KWUKyKq5LVh3npif99Wkt6j86Jw==}
engines: {node: '>=6.9.0'}
'@babel/helper-replace-supers@7.29.7':
resolution: {integrity: sha512-atfGXWSeCiF4DnKZIfmJfQRkSw9b9gNNXR1kqKjbhG4pGYCOnkp8OcTB8E3NXjBu8NpheSnOeNKz8KT7UNFTmQ==}
engines: {node: '>=6.9.0'}
peerDependencies:
'@babel/core': ^7.0.0
'@babel/helper-skip-transparent-expression-wrappers@7.29.7':
resolution: {integrity: sha512-brcMGQaVzIeUb+6/bs1Av0f8YuNNjKY2JyvfRCsFuFsdKccEQ5Ges2y74D74NZ1Rz8lKJ9ksJkfqwQFJ/iNEyQ==}
engines: {node: '>=6.9.0'}
first commit
2026-02-21 03:10:39 +00:00
'@babel/helper-string-parser@7.27.1':
resolution: {integrity: sha512-qMlSxKbpRlAridDExk92nSobyDdpPijUq2DW6oDnUqd0iOGxmQjyqhMIihI9+zv4LPyZdRje2cavWPbCbWm3eA==}
engines: {node: '>=6.9.0'}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
'@babel/helper-string-parser@7.29.7':
resolution: {integrity: sha512-Pb5ijPrZ89GDH8223L4UP8i6QApWxs04RbPQJTeWDV0/keR2E36MeKnyr6LYmUUvqRRI+Iv87SuF1W6ErINzYw==}
engines: {node: '>=6.9.0'}
first commit
2026-02-21 03:10:39 +00:00
'@babel/helper-validator-identifier@7.28.5':
resolution: {integrity: sha512-qSs4ifwzKJSV39ucNjsvc6WVHs6b7S03sOh2OcHF9UHfVPqWWALUsNUVzhSBiItjRZoLHx7nIarVjqKVusUZ1Q==}
engines: {node: '>=6.9.0'}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
'@babel/helper-validator-identifier@7.29.7':
resolution: {integrity: sha512-qehxGkRj55h/ff8EMaJ+cYhyaKlHIxqYDn682wQD7RNp9UujOQsHog2uS0r2vzr4pW+sXf90NeeayjcNaX3fFg==}
engines: {node: '>=6.9.0'}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
'@babel/helper-validator-option@7.27.1':
resolution: {integrity: sha512-YvjJow9FxbhFFKDSuFnVCe2WxXk1zWc22fFePVNEaWJEu8IrZVlda6N0uHwzZrUM1il7NC9Mlp4MaJYbYd9JSg==}
engines: {node: '>=6.9.0'}
'@babel/helpers@7.29.2':
resolution: {integrity: sha512-HoGuUs4sCZNezVEKdVcwqmZN8GoHirLUcLaYVNBK2J0DadGtdcqgr3BCbvH8+XUo4NGjNl3VOtSjEKNzqfFgKw==}
engines: {node: '>=6.9.0'}
first commit
2026-02-21 03:10:39 +00:00
'@babel/parser@7.29.0':
resolution: {integrity: sha512-IyDgFV5GeDUVX4YdF/3CPULtVGSXXMLh1xVIgdCgxApktqnQV0r7/8Nqthg+8YLGaAtdyIlo2qIdZrbCv4+7ww==}
engines: {node: '>=6.0.0'}
hasBin: true
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
'@babel/parser@7.29.8':
resolution: {integrity: sha512-E8lTAYNB1KW+FH+VGJuZM1ioAx2E6oVlvQFRrf5P8ZZmsiJXYAD9vTFV7yyEURNzgh1dFqMZuO6tUwcARbqFCA==}
engines: {node: '>=6.0.0'}
hasBin: true
'@babel/plugin-syntax-jsx@7.29.7':
resolution: {integrity: sha512-TSu8+mHCoEaaCDEZ0I3+6mvTBYR4PCxQwf2z9/r5Tbztv6NaLR3B9thGTTxX2WGuGHJqRiAbKPeGTJ5XWXVg6A==}
engines: {node: '>=6.9.0'}
peerDependencies:
'@babel/core': ^7.0.0-0
'@babel/plugin-syntax-typescript@7.29.7':
resolution: {integrity: sha512-ngr+82Sh0xMz25TPCZi+nC2iTzjfCdWS2ONXTp/PtSCHCgaCNBpdMqgvJ2ccdLlClVZ7sisIgB914j/JFe+RZA==}
engines: {node: '>=6.9.0'}
peerDependencies:
'@babel/core': ^7.0.0-0
'@babel/plugin-transform-modules-commonjs@7.29.7':
resolution: {integrity: sha512-j0vCldybPC5b5dwCQOJ21uKtHzt7hxLygJTg9eF1ScfaikEDNfzn94XoW5Fi+seBR0nCyL23xaBFFkq7dTM8XQ==}
engines: {node: '>=6.9.0'}
peerDependencies:
'@babel/core': ^7.0.0-0
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
'@babel/plugin-transform-react-jsx-self@7.27.1':
resolution: {integrity: sha512-6UzkCs+ejGdZ5mFFC/OCUrv028ab2fp1znZmCZjAOBKiBK2jXD1O+BPSfX8X2qjJ75fZBMSnQn3Rq2mrBJK2mw==}
engines: {node: '>=6.9.0'}
peerDependencies:
'@babel/core': ^7.0.0-0
'@babel/plugin-transform-react-jsx-source@7.27.1':
resolution: {integrity: sha512-zbwoTsBruTeKB9hSq73ha66iFeJHuaFkUbwvqElnygoNbj/jHRsSeokowZFN3CZ64IvEqcmmkVe89OPXc7ldAw==}
engines: {node: '>=6.9.0'}
peerDependencies:
'@babel/core': ^7.0.0-0
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
'@babel/plugin-transform-typescript@7.29.7':
resolution: {integrity: sha512-jK52h8LaLc7JarhQV2ofeFMts4H7vnOXnqZNA6fYglBTZewRBE51KWt3BUltW1P+KoPsYkHoJeXePuz4zo2LMw==}
engines: {node: '>=6.9.0'}
peerDependencies:
'@babel/core': ^7.0.0-0
'@babel/preset-typescript@7.27.1':
resolution: {integrity: sha512-l7WfQfX0WK4M0v2RudjuQK4u99BS6yLHYEmdtVPP7lKV013zr9DygFuWNlnbvQ9LR+LS0Egz/XAvGx5U9MX0fQ==}
engines: {node: '>=6.9.0'}
peerDependencies:
'@babel/core': ^7.0.0-0
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
'@babel/runtime@7.29.2':
resolution: {integrity: sha512-JiDShH45zKHWyGe4ZNVRrCjBz8Nh9TMmZG1kh4QTK8hCBTWBi8Da+i7s1fJw7/lYpM4ccepSNfqzZ/QvABBi5g==}
engines: {node: '>=6.9.0'}
'@babel/template@7.28.6':
resolution: {integrity: sha512-YA6Ma2KsCdGb+WC6UpBVFJGXL58MDA6oyONbjyF/+5sBgxY/dwkhLogbMT2GXXyU84/IhRw/2D1Os1B/giz+BQ==}
engines: {node: '>=6.9.0'}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
'@babel/template@7.29.7':
resolution: {integrity: sha512-puq+Gf35oI24FeN11LkoUQFqv9uwNeWpxXZi/Ji3rRIoKAzKnxRaZ+Gkj0vKS9ZCiTESfng1N9LyOyXvo+m+Gg==}
engines: {node: '>=6.9.0'}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
'@babel/traverse@7.29.0':
resolution: {integrity: sha512-4HPiQr0X7+waHfyXPZpWPfWL/J7dcN1mx9gL6WdQVMbPnF3+ZhSMs8tCxN7oHddJE9fhNE7+lxdnlyemKfJRuA==}
engines: {node: '>=6.9.0'}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
'@babel/traverse@7.29.8':
resolution: {integrity: sha512-I5z7H3bf/41ktsNVLtpN0wAa336HkqIHQ5BuPLEhTkt1jVSyZpeNKIzTgEWmlxjdg81R0IgUCcaE+Ok3NvrfZg==}
engines: {node: '>=6.9.0'}
first commit
2026-02-21 03:10:39 +00:00
'@babel/types@7.29.0':
resolution: {integrity: sha512-LwdZHpScM4Qz8Xw2iKSzS+cfglZzJGvofQICy7W7v4caru4EaAmyUuO6BGrbyQ2mYV11W0U8j5mBhd14dd3B0A==}
engines: {node: '>=6.9.0'}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
'@babel/types@7.29.8':
resolution: {integrity: sha512-Vj1jF3cPfxg7OAfoI7QnVKLoILlm2JF9pnVHrX8qx7AHMiYWT+NDAA7jChlNgRS4WTLc/fD1lXLmPixluj+3Gg==}
engines: {node: '>=6.9.0'}
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
'@balena/dockerignore@1.0.2':
resolution: {integrity: sha512-wMue2Sy4GAVTk6Ic4tJVcnfdau+gx2EnG7S+uAEe+TWJFqE4YoWN4/H8MSLj4eYJKxGg26lZwboEniNiNwZQ6Q==}
first commit
2026-02-21 03:10:39 +00:00
'@bcoe/v8-coverage@1.0.2':
resolution: {integrity: sha512-6zABk/ECA/QYSCQ1NGiVwwbQerUCZ+TQbp64Q3AgmfNvurHH0j8TtXa1qbShXA6qqkpAj4V5W8pP6mLe1mcMqA==}
engines: {node: '>=18'}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
'@bramus/specificity@2.4.2':
resolution: {integrity: sha512-ctxtJ/eA+t+6q2++vj5j7FYX3nRu311q1wfYH3xjlLOsczhlhxAg2FWNUXhpGvAw3BWo1xBcvOV6/YLc2r5FJw==}
hasBin: true
'@csstools/color-helpers@6.0.2':
resolution: {integrity: sha512-LMGQLS9EuADloEFkcTBR3BwV/CGHV7zyDxVRtVDTwdI2Ca4it0CCVTT9wCkxSgokjE5Ho41hEPgb8OEUwoXr6Q==}
engines: {node: '>=20.19.0'}
'@csstools/css-calc@3.2.0':
resolution: {integrity: sha512-bR9e6o2BDB12jzN/gIbjHa5wLJ4UjD1CB9pM7ehlc0ddk6EBz+yYS1EV2MF55/HUxrHcB/hehAyt5vhsA3hx7w==}
engines: {node: '>=20.19.0'}
peerDependencies:
'@csstools/css-parser-algorithms': ^4.0.0
'@csstools/css-tokenizer': ^4.0.0
'@csstools/css-color-parser@4.1.0':
resolution: {integrity: sha512-U0KhLYmy2GVj6q4T3WaAe6NPuFYCPQoE3b0dRGxejWDgcPp8TP7S5rVdM5ZrFaqu4N67X8YaPBw14dQSYx3IyQ==}
engines: {node: '>=20.19.0'}
peerDependencies:
'@csstools/css-parser-algorithms': ^4.0.0
'@csstools/css-tokenizer': ^4.0.0
'@csstools/css-parser-algorithms@4.0.0':
resolution: {integrity: sha512-+B87qS7fIG3L5h3qwJ/IFbjoVoOe/bpOdh9hAjXbvx0o8ImEmUsGXN0inFOnk2ChCFgqkkGFQ+TpM5rbhkKe4w==}
engines: {node: '>=20.19.0'}
peerDependencies:
'@csstools/css-tokenizer': ^4.0.0
'@csstools/css-syntax-patches-for-csstree@1.1.3':
resolution: {integrity: sha512-SH60bMfrRCJF3morcdk57WklujF4Jr/EsQUzqkarfHXEFcAR1gg7fS/chAE922Sehgzc1/+Tz5H3Ypa1HiEKrg==}
peerDependencies:
css-tree: ^3.2.1
peerDependenciesMeta:
css-tree:
optional: true
'@csstools/css-tokenizer@4.0.0':
resolution: {integrity: sha512-QxULHAm7cNu72w97JUNCBFODFaXpbDg+dP8b/oWFAZ2MTRppA3U00Y2L1HqaS4J6yBqxwa/Y3nMBaxVKbB/NsA==}
engines: {node: '>=20.19.0'}
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
'@earendil-works/pi-agent-core@0.84.1':
resolution: {integrity: sha512-evyzXYWCLQGmcaBYHlmSku02r8qoN4SGI60GZABo6iV+H+nqX+P9ud8fEZ4GmRq9mUSREvvfX+w9dA9ThF9C6w==}
engines: {node: '>=22.19.0'}
'@earendil-works/pi-ai@0.84.1':
resolution: {integrity: sha512-wMsAdJMxuNri08vLqTyYVI201DQQezGhPSTkzYsHdw5dYX3rCNwEmSvpaAwhi7ELKI/2tE/CEgSWg/6iRxSgdQ==}
engines: {node: '>=22.19.0'}
hasBin: true
'@earendil-works/pi-client@0.84.1':
resolution: {integrity: sha512-/V5hGHE4Zq+jG0GtwIB9PyBUOGd6gBLZ7lkQYFKchKnxYHeH3rmWC5xw4kpnZKKBuBuFTdLVbU9vEjlAGMMb2A==}
engines: {node: '>=22.19.0'}
'@earendil-works/pi-coding-agent@0.84.1':
resolution: {integrity: sha512-ncAqFrG+iybuPGOhMiZoEHkEzTpJgz3guYD32pD+M7ucc0WeHmauP6wa7qwP8V/KWvsZDVNa5XGsdZ7fkC7w7A==}
engines: {node: '>=22.19.0'}
hasBin: true
'@earendil-works/pi-protocol@0.84.1':
resolution: {integrity: sha512-Ox1pciyeSPGEEUcxvR0/dJcrY7C6hrEGA8y71rOsvSIUlXN1Cbp/be/eoL71OGDBk5O97TeQPfWN6Ju/2Ehjww==}
engines: {node: '>=22.19.0'}
'@earendil-works/pi-telemetry@0.84.1':
resolution: {integrity: sha512-180/xGJtsq7IoR3p9EKWjRd0e9M4DkxInhlo9xyD7prDC7Qrhqq+nhvwrW0lFjPfXcEI2FSHmGCSyvSJE9GsaQ==}
engines: {node: '>=22.19.0'}
'@earendil-works/pi-tui@0.84.1':
resolution: {integrity: sha512-udeXFbgEhJ6JiB0uguwNVNkDy2FENfmtQwPcY+/iJ8GWeq18wkal1tKqa5YyeH0IqtX1vG0cGh8zfSYzyzVuLA==}
engines: {node: '>=22.19.0'}
feat(web): bold redesign — Tailwind v4 + shadcn-style primitives + Skills/Proposals/Revisions UI Phase 6 of the Skills + Revisions + Proposals work. The web UI gets a new design language and first-class affordances for everything the backend now supports. ## Visual direction - Tailwind v4 with custom @theme block (oklch tokens). Dark-mode-only (internal tool — light mode doubles QA surface). - Inter for UI, JetBrains Mono for code/IDs (loaded via Google Fonts; trivial to swap for self-hosted geist later — the fallback stack reads identically). - Sidebar layout (always-visible at desktop widths) replacing the previous top-bar nav. Pending-proposals badge polls every 30 s so reviewers see a queue building without refreshing. - Lucide icons throughout. - Spacing and radii on Tailwind defaults. Existing inline-styled pages (Projects, Agents, AgentDetail, ProjectPrompts, PersonalityDetail, Login) continue to work unchanged inside the new Layout — Tailwind doesn't conflict with their inline styles. A follow-up can migrate them incrementally. ## What's added ### Build infra (src/web/) - package.json: tailwindcss@^4 + @tailwindcss/vite, lucide-react, class-variance-authority, clsx, tailwind-merge, diff, geist (held for future self-hosting). - vite.config.ts: registers the @tailwindcss/vite plugin. - src/index.css: Tailwind import + @theme tokens + @layer base. - src/main.tsx: imports index.css. - src/lib/utils.ts: shadcn-style cn() helper. ### shadcn-style primitives (src/components/ui/) Hand-written rather than generated via `npx shadcn` so the repo doesn't depend on a CLI tool that needs interactive runtime: - button.tsx — variants: primary / secondary / ghost / danger / link; sizes: sm / md / lg / icon. - card.tsx — Card + Header/Title/Description/Content/Footer subparts. - badge.tsx — variants: default / info / success / warning / danger / outline. - input.tsx — Input + Textarea + Label. - tabs.tsx — no-dep accessible Tabs (no Radix needed for our use). - separator.tsx — h/v separator with role=separator. ### Diff component (src/components/Diff.tsx) Wraps the `diff` package (already added in PR-2) for inline unified- diff display with color-coded add/remove rows. Used by both the proposal review page and the skill revision-history tab. ### New pages (src/pages/) - Dashboard.tsx — at-a-glance home. Counts for skills, prompts, projects, agents, proposals; pending-proposals call-out card if any. - Skills.tsx — list view, separated into Global vs Project/Agent- scoped sections. - SkillDetail.tsx — name + semver + description; tabs for SKILL.md / Files / Metadata / History. History tab shows revisions with click-to-diff against the live body. - Proposals.tsx — queue with Pending/Approved/Rejected tabs. Pending count is highlighted in amber. - ProposalDetail.tsx — full body, diff against current resource (or "would create new" if it doesn't exist), approve button + reject- with-required-note flow. ### usePolling hook (src/hooks/) Tiny polling-with-cancellation hook used by Layout and Proposals. ### Layout rewrite (src/components/Layout.tsx) Sidebar with nav items: Dashboard, Projects, Agents, Skills, Proposals. Lucide icons. Active-route highlighting via NavLink. Pending-proposals warning badge on the Proposals item. ### Routes (src/App.tsx) New routes: /dashboard, /skills, /skills/:name, /proposals, /proposals/:id. Default redirects to /dashboard. ### API types (src/api.ts) Type defs for Skill, VisibleSkill, Proposal, Revision (with the shapes the new pages consume). ## Tests Existing 7 web tests still pass (Login + api). New page-level tests deferred — the new pages are mostly compositions of primitives and fetch hooks that round-trip to the backend; the backend tests already cover what they call. PR-7 polish can add render-and-click tests if coverage drift surfaces. ## Verification - `pnpm --filter @mcpctl/web build` clean, no warnings. - `pnpm test:run` whole monorepo: 162 test files / 2157 tests green. - Visual smoke deferred — needs a running mcpd to populate the fixtures. Manual smoke tested locally is the next step. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-07 17:54:55 +01:00
'@emnapi/runtime@1.10.0':
resolution: {integrity: sha512-ewvYlk86xUoGI0zQRNq/mC+16R1QeDlKQy21Ki3oSYXNgLb45GV1P6A0M+/s6nyCuNDqe5VpaY84BzXGwVbwFA==}
first commit
2026-02-21 03:10:39 +00:00
'@esbuild/aix-ppc64@0.27.3':
resolution: {integrity: sha512-9fJMTNFTWZMh5qwrBItuziu834eOCUcEqymSH7pY+zoMVEZg3gcPuBNxH1EvfVYe9h0x/Ptw8KBzv7qxb7l8dg==}
engines: {node: '>=18'}
cpu: [ppc64]
os: [aix]
'@esbuild/android-arm64@0.27.3':
resolution: {integrity: sha512-YdghPYUmj/FX2SYKJ0OZxf+iaKgMsKHVPF1MAq/P8WirnSpCStzKJFjOjzsW0QQ7oIAiccHdcqjbHmJxRb/dmg==}
engines: {node: '>=18'}
cpu: [arm64]
os: [android]
'@esbuild/android-arm@0.27.3':
resolution: {integrity: sha512-i5D1hPY7GIQmXlXhs2w8AWHhenb00+GxjxRncS2ZM7YNVGNfaMxgzSGuO8o8SJzRc/oZwU2bcScvVERk03QhzA==}
engines: {node: '>=18'}
cpu: [arm]
os: [android]
'@esbuild/android-x64@0.27.3':
resolution: {integrity: sha512-IN/0BNTkHtk8lkOM8JWAYFg4ORxBkZQf9zXiEOfERX/CzxW3Vg1ewAhU7QSWQpVIzTW+b8Xy+lGzdYXV6UZObQ==}
engines: {node: '>=18'}
cpu: [x64]
os: [android]
'@esbuild/darwin-arm64@0.27.3':
resolution: {integrity: sha512-Re491k7ByTVRy0t3EKWajdLIr0gz2kKKfzafkth4Q8A5n1xTHrkqZgLLjFEHVD+AXdUGgQMq+Godfq45mGpCKg==}
engines: {node: '>=18'}
cpu: [arm64]
os: [darwin]
'@esbuild/darwin-x64@0.27.3':
resolution: {integrity: sha512-vHk/hA7/1AckjGzRqi6wbo+jaShzRowYip6rt6q7VYEDX4LEy1pZfDpdxCBnGtl+A5zq8iXDcyuxwtv3hNtHFg==}
engines: {node: '>=18'}
cpu: [x64]
os: [darwin]
'@esbuild/freebsd-arm64@0.27.3':
resolution: {integrity: sha512-ipTYM2fjt3kQAYOvo6vcxJx3nBYAzPjgTCk7QEgZG8AUO3ydUhvelmhrbOheMnGOlaSFUoHXB6un+A7q4ygY9w==}
engines: {node: '>=18'}
cpu: [arm64]
os: [freebsd]
'@esbuild/freebsd-x64@0.27.3':
resolution: {integrity: sha512-dDk0X87T7mI6U3K9VjWtHOXqwAMJBNN2r7bejDsc+j03SEjtD9HrOl8gVFByeM0aJksoUuUVU9TBaZa2rgj0oA==}
engines: {node: '>=18'}
cpu: [x64]
os: [freebsd]
'@esbuild/linux-arm64@0.27.3':
resolution: {integrity: sha512-sZOuFz/xWnZ4KH3YfFrKCf1WyPZHakVzTiqji3WDc0BCl2kBwiJLCXpzLzUBLgmp4veFZdvN5ChW4Eq/8Fc2Fg==}
engines: {node: '>=18'}
cpu: [arm64]
os: [linux]
'@esbuild/linux-arm@0.27.3':
resolution: {integrity: sha512-s6nPv2QkSupJwLYyfS+gwdirm0ukyTFNl3KTgZEAiJDd+iHZcbTPPcWCcRYH+WlNbwChgH2QkE9NSlNrMT8Gfw==}
engines: {node: '>=18'}
cpu: [arm]
os: [linux]
'@esbuild/linux-ia32@0.27.3':
resolution: {integrity: sha512-yGlQYjdxtLdh0a3jHjuwOrxQjOZYD/C9PfdbgJJF3TIZWnm/tMd/RcNiLngiu4iwcBAOezdnSLAwQDPqTmtTYg==}
engines: {node: '>=18'}
cpu: [ia32]
os: [linux]
'@esbuild/linux-loong64@0.27.3':
resolution: {integrity: sha512-WO60Sn8ly3gtzhyjATDgieJNet/KqsDlX5nRC5Y3oTFcS1l0KWba+SEa9Ja1GfDqSF1z6hif/SkpQJbL63cgOA==}
engines: {node: '>=18'}
cpu: [loong64]
os: [linux]
'@esbuild/linux-mips64el@0.27.3':
resolution: {integrity: sha512-APsymYA6sGcZ4pD6k+UxbDjOFSvPWyZhjaiPyl/f79xKxwTnrn5QUnXR5prvetuaSMsb4jgeHewIDCIWljrSxw==}
engines: {node: '>=18'}
cpu: [mips64el]
os: [linux]
'@esbuild/linux-ppc64@0.27.3':
resolution: {integrity: sha512-eizBnTeBefojtDb9nSh4vvVQ3V9Qf9Df01PfawPcRzJH4gFSgrObw+LveUyDoKU3kxi5+9RJTCWlj4FjYXVPEA==}
engines: {node: '>=18'}
cpu: [ppc64]
os: [linux]
'@esbuild/linux-riscv64@0.27.3':
resolution: {integrity: sha512-3Emwh0r5wmfm3ssTWRQSyVhbOHvqegUDRd0WhmXKX2mkHJe1SFCMJhagUleMq+Uci34wLSipf8Lagt4LlpRFWQ==}
engines: {node: '>=18'}
cpu: [riscv64]
os: [linux]
'@esbuild/linux-s390x@0.27.3':
resolution: {integrity: sha512-pBHUx9LzXWBc7MFIEEL0yD/ZVtNgLytvx60gES28GcWMqil8ElCYR4kvbV2BDqsHOvVDRrOxGySBM9Fcv744hw==}
engines: {node: '>=18'}
cpu: [s390x]
os: [linux]
'@esbuild/linux-x64@0.27.3':
resolution: {integrity: sha512-Czi8yzXUWIQYAtL/2y6vogER8pvcsOsk5cpwL4Gk5nJqH5UZiVByIY8Eorm5R13gq+DQKYg0+JyQoytLQas4dA==}
engines: {node: '>=18'}
cpu: [x64]
os: [linux]
'@esbuild/netbsd-arm64@0.27.3':
resolution: {integrity: sha512-sDpk0RgmTCR/5HguIZa9n9u+HVKf40fbEUt+iTzSnCaGvY9kFP0YKBWZtJaraonFnqef5SlJ8/TiPAxzyS+UoA==}
engines: {node: '>=18'}
cpu: [arm64]
os: [netbsd]
'@esbuild/netbsd-x64@0.27.3':
resolution: {integrity: sha512-P14lFKJl/DdaE00LItAukUdZO5iqNH7+PjoBm+fLQjtxfcfFE20Xf5CrLsmZdq5LFFZzb5JMZ9grUwvtVYzjiA==}
engines: {node: '>=18'}
cpu: [x64]
os: [netbsd]
'@esbuild/openbsd-arm64@0.27.3':
resolution: {integrity: sha512-AIcMP77AvirGbRl/UZFTq5hjXK+2wC7qFRGoHSDrZ5v5b8DK/GYpXW3CPRL53NkvDqb9D+alBiC/dV0Fb7eJcw==}
engines: {node: '>=18'}
cpu: [arm64]
os: [openbsd]
'@esbuild/openbsd-x64@0.27.3':
resolution: {integrity: sha512-DnW2sRrBzA+YnE70LKqnM3P+z8vehfJWHXECbwBmH/CU51z6FiqTQTHFenPlHmo3a8UgpLyH3PT+87OViOh1AQ==}
engines: {node: '>=18'}
cpu: [x64]
os: [openbsd]
'@esbuild/openharmony-arm64@0.27.3':
resolution: {integrity: sha512-NinAEgr/etERPTsZJ7aEZQvvg/A6IsZG/LgZy+81wON2huV7SrK3e63dU0XhyZP4RKGyTm7aOgmQk0bGp0fy2g==}
engines: {node: '>=18'}
cpu: [arm64]
os: [openharmony]
'@esbuild/sunos-x64@0.27.3':
resolution: {integrity: sha512-PanZ+nEz+eWoBJ8/f8HKxTTD172SKwdXebZ0ndd953gt1HRBbhMsaNqjTyYLGLPdoWHy4zLU7bDVJztF5f3BHA==}
engines: {node: '>=18'}
cpu: [x64]
os: [sunos]
'@esbuild/win32-arm64@0.27.3':
resolution: {integrity: sha512-B2t59lWWYrbRDw/tjiWOuzSsFh1Y/E95ofKz7rIVYSQkUYBjfSgf6oeYPNWHToFRr2zx52JKApIcAS/D5TUBnA==}
engines: {node: '>=18'}
cpu: [arm64]
os: [win32]
'@esbuild/win32-ia32@0.27.3':
resolution: {integrity: sha512-QLKSFeXNS8+tHW7tZpMtjlNb7HKau0QDpwm49u0vUp9y1WOF+PEzkU84y9GqYaAVW8aH8f3GcBck26jh54cX4Q==}
engines: {node: '>=18'}
cpu: [ia32]
os: [win32]
'@esbuild/win32-x64@0.27.3':
resolution: {integrity: sha512-4uJGhsxuptu3OcpVAzli+/gWusVGwZZHTlS63hh++ehExkVT8SgiEf7/uC/PclrPPkLhZqGgCTjd0VWLo6xMqA==}
engines: {node: '>=18'}
cpu: [x64]
os: [win32]
'@eslint-community/eslint-utils@4.9.1':
resolution: {integrity: sha512-phrYmNiYppR7znFEdqgfWHXR6NCkZEK7hwWDHZUjit/2/U0r6XvkDl0SYnoM51Hq7FhCGdLDT6zxCCOY1hexsQ==}
engines: {node: ^12.22.0 || ^14.17.0 || >=16.0.0}
peerDependencies:
eslint: ^6.0.0 || ^7.0.0 || >=8.0.0
'@eslint-community/regexpp@4.12.2':
resolution: {integrity: sha512-EriSTlt5OC9/7SXkRSCAhfSxxoSUgBm33OH+IkwbdpgoqsSsUg7y3uh+IICI/Qg4BBWr3U2i39RpmycbxMq4ew==}
engines: {node: ^12.0.0 || ^14.0.0 || >=16.0.0}
'@eslint/config-array@0.23.2':
resolution: {integrity: sha512-YF+fE6LV4v5MGWRGj7G404/OZzGNepVF8fxk7jqmqo3lrza7a0uUcDnROGRBG1WFC1omYUS/Wp1f42i0M+3Q3A==}
engines: {node: ^20.19.0 || ^22.13.0 || >=24}
'@eslint/config-helpers@0.5.2':
resolution: {integrity: sha512-a5MxrdDXEvqnIq+LisyCX6tQMPF/dSJpCfBgBauY+pNZ28yCtSsTvyTYrMhaI+LK26bVyCJfJkT0u8KIj2i1dQ==}
engines: {node: ^20.19.0 || ^22.13.0 || >=24}
'@eslint/core@1.1.0':
resolution: {integrity: sha512-/nr9K9wkr3P1EzFTdFdMoLuo1PmIxjmwvPozwoSodjNBdefGujXQUF93u1DDZpEaTuDvMsIQddsd35BwtrW9Xw==}
engines: {node: ^20.19.0 || ^22.13.0 || >=24}
'@eslint/object-schema@3.0.2':
resolution: {integrity: sha512-HOy56KJt48Bx8KmJ+XGQNSUMT/6dZee/M54XyUyuvTvPXJmsERRvBchsUVx1UMe1WwIH49XLAczNC7V2INsuUw==}
engines: {node: ^20.19.0 || ^22.13.0 || >=24}
'@eslint/plugin-kit@0.6.0':
resolution: {integrity: sha512-bIZEUzOI1jkhviX2cp5vNyXQc6olzb2ohewQubuYlMXZ2Q/XjBO0x0XhGPvc9fjSIiUN0vw+0hq53BJ4eQSJKQ==}
engines: {node: ^20.19.0 || ^22.13.0 || >=24}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
'@exodus/bytes@1.15.0':
resolution: {integrity: sha512-UY0nlA+feH81UGSHv92sLEPLCeZFjXOuHhrIo0HQydScuQc8s0A7kL/UdgwgDq8g8ilksmuoF35YVTNphV2aBQ==}
engines: {node: ^20.19.0 || ^22.12.0 || >=24.0.0}
peerDependencies:
'@noble/hashes': ^1.8.0 || ^2.0.0
peerDependenciesMeta:
'@noble/hashes':
optional: true
feat(mcpd+deploy): serve web UI at /ui + smoke tests + docs (Stage 6) The closing stage. mcpd now hosts the Stage 5 SPA, the Docker image bundles the build artifact, a smoke test exercises the personality HTTP surface end-to-end, and the user-facing docs spell out the mental model. mcpd: - Add @fastify/static dep. - New routes/web-ui.ts: registers /ui/* against a static bundle. Looks for the bundle at $MCPD_WEB_ROOT, then /usr/share/mcpd/web (the Docker image path), then a dev-tree fallback. Logs and skips cleanly if missing — API-only deploys keep working. - SPA fallback: any /ui/<path> that doesn't match a file falls through to index.html so direct hits to react-router URLs work. - /ui/* falls through to `kind: skip` in mapUrlToPermission, so the static assets are served unauthenticated. Each API call from the SPA still carries the bearer token. Deploy: - Dockerfile.mcpd builds the @mcpctl/web bundle in the same builder stage and copies dist/ to /usr/share/mcpd/web in the runtime image. Smoke (personality.smoke.test.ts): - Live mcpd flow: create secret/llm/agent/personality, attach an agent-direct prompt, verify the binding listing, reject double- attach (409) + foreign-agent prompt (400), set defaultPersonality by name, detach + delete cleanup. Docs: - New docs/personalities.md: VLAN-on-ethernet model, system-block ordering table, three prompt scopes, CLI walkthrough, web UI walkthrough, full API surface, RBAC notes. - agents.md and chat.md cross-link. - README's Agents section gains a Personalities subsection. Test count after Stage 6: mcpd: 801/801 cli: 430/430 web: 7/7 db: 58/62 (4 pre-existing) Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:48:43 +01:00
'@fastify/accept-negotiator@2.0.1':
resolution: {integrity: sha512-/c/TW2bO/v9JeEgoD/g1G5GxGeCF1Hafdf79WPmUlgYiBXummY0oX3VVq4yFkKKVBKDNlaDUYoab7g38RpPqCQ==}
first commit
2026-02-21 03:10:39 +00:00
'@fastify/ajv-compiler@4.0.5':
resolution: {integrity: sha512-KoWKW+MhvfTRWL4qrhUwAAZoaChluo0m0vbiJlGMt2GXvL4LVPQEjt8kSpHI3IBq5Rez8fg+XeH3cneztq+C7A==}
'@fastify/cors@10.1.0':
resolution: {integrity: sha512-MZyBCBJtII60CU9Xme/iE4aEy8G7QpzGR8zkdXZkDFt7ElEMachbE61tfhAG/bvSaULlqlf0huMT12T7iqEmdQ==}
'@fastify/error@4.2.0':
resolution: {integrity: sha512-RSo3sVDXfHskiBZKBPRgnQTtIqpi/7zhJOEmAxCiBcM7d0uwdGdxLlsCaLzGs8v8NnxIRlfG0N51p5yFaOentQ==}
'@fastify/fast-json-stringify-compiler@5.0.3':
resolution: {integrity: sha512-uik7yYHkLr6fxd8hJSZ8c+xF4WafPK+XzneQDPU+D10r5X19GW8lJcom2YijX2+qtFF1ENJlHXKFM9ouXNJYgQ==}
'@fastify/forwarded@3.0.1':
resolution: {integrity: sha512-JqDochHFqXs3C3Ml3gOY58zM7OqO9ENqPo0UqAjAjH8L01fRZqwX9iLeX34//kiJubF7r2ZQHtBRU36vONbLlw==}
'@fastify/helmet@12.0.1':
resolution: {integrity: sha512-kkjBcedWwdflRThovGuvN9jB2QQLytBqArCFPdMIb7o2Fp0l/H3xxYi/6x/SSRuH/FFt9qpTGIfJz2bfnMrLqA==}
'@fastify/merge-json-schemas@0.2.1':
resolution: {integrity: sha512-OA3KGBCy6KtIvLf8DINC5880o5iBlDX4SxzLQS8HorJAbqluzLRn80UXU0bxZn7UOFhFgpRJDasfwn9nG4FG4A==}
'@fastify/proxy-addr@5.1.0':
resolution: {integrity: sha512-INS+6gh91cLUjB+PVHfu1UqcB76Sqtpyp7bnL+FYojhjygvOPA9ctiD/JDKsyD9Xgu4hUhCSJBPig/w7duNajw==}
'@fastify/rate-limit@10.3.0':
resolution: {integrity: sha512-eIGkG9XKQs0nyynatApA3EVrojHOuq4l6fhB4eeCk4PIOeadvOJz9/4w3vGI44Go17uaXOWEcPkaD8kuKm7g6Q==}
feat(mcpd+deploy): serve web UI at /ui + smoke tests + docs (Stage 6) The closing stage. mcpd now hosts the Stage 5 SPA, the Docker image bundles the build artifact, a smoke test exercises the personality HTTP surface end-to-end, and the user-facing docs spell out the mental model. mcpd: - Add @fastify/static dep. - New routes/web-ui.ts: registers /ui/* against a static bundle. Looks for the bundle at $MCPD_WEB_ROOT, then /usr/share/mcpd/web (the Docker image path), then a dev-tree fallback. Logs and skips cleanly if missing — API-only deploys keep working. - SPA fallback: any /ui/<path> that doesn't match a file falls through to index.html so direct hits to react-router URLs work. - /ui/* falls through to `kind: skip` in mapUrlToPermission, so the static assets are served unauthenticated. Each API call from the SPA still carries the bearer token. Deploy: - Dockerfile.mcpd builds the @mcpctl/web bundle in the same builder stage and copies dist/ to /usr/share/mcpd/web in the runtime image. Smoke (personality.smoke.test.ts): - Live mcpd flow: create secret/llm/agent/personality, attach an agent-direct prompt, verify the binding listing, reject double- attach (409) + foreign-agent prompt (400), set defaultPersonality by name, detach + delete cleanup. Docs: - New docs/personalities.md: VLAN-on-ethernet model, system-block ordering table, three prompt scopes, CLI walkthrough, web UI walkthrough, full API surface, RBAC notes. - agents.md and chat.md cross-link. - README's Agents section gains a Personalities subsection. Test count after Stage 6: mcpd: 801/801 cli: 430/430 web: 7/7 db: 58/62 (4 pre-existing) Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:48:43 +01:00
'@fastify/send@4.1.0':
resolution: {integrity: sha512-TMYeQLCBSy2TOFmV95hQWkiTYgC/SEx7vMdV+wnZVX4tt8VBLKzmH8vV9OzJehV0+XBfg+WxPMt5wp+JBUKsVw==}
'@fastify/static@8.3.0':
resolution: {integrity: sha512-yKxviR5PH1OKNnisIzZKmgZSus0r2OZb8qCSbqmw34aolT4g3UlzYfeBRym+HJ1J471CR8e2ldNub4PubD1coA==}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
'@gar/promise-retry@1.0.3':
resolution: {integrity: sha512-GmzA9ckNokPypTg10pgpeHNQe7ph+iIKKmhKu3Ob9ANkswreCx7R3cKmY781K8QK3AqVL3xVh9A42JvIAbkkSA==}
engines: {node: ^20.17.0 || >=22.9.0}
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
'@google/genai@1.52.0':
resolution: {integrity: sha512-gwSvbpiN/17O9TbsqSsE/OzZcpv5Fo4RQjdngGgogtuB9RsyJ8ZHhX5KjHj1bp5N9snN2eK8LDGXSaWW2hof8Q==}
engines: {node: '>=20.0.0'}
peerDependencies:
'@modelcontextprotocol/sdk': ^1.25.2
peerDependenciesMeta:
'@modelcontextprotocol/sdk':
optional: true
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
'@grpc/grpc-js@1.14.3':
resolution: {integrity: sha512-Iq8QQQ/7X3Sac15oB6p0FmUg/klxQvXLeileoqrTRGJYLV+/9tubbr9ipz0GKHjmXVsgFPo/+W+2cA8eNcR+XA==}
engines: {node: '>=12.10.0'}
'@grpc/proto-loader@0.7.15':
resolution: {integrity: sha512-tMXdRCfYVixjuFK+Hk0Q1s38gV9zDiDJfWL3h1rv4Qc39oILCu1TRTDt7+fGUI8K4G1Fj125Hx/ru3azECWTyQ==}
engines: {node: '>=6'}
hasBin: true
'@grpc/proto-loader@0.8.0':
resolution: {integrity: sha512-rc1hOQtjIWGxcxpb9aHAfLpIctjEnsDehj0DAiVfBlmT84uvR0uUtN2hEi/ecvWVjXUGf5qPF4qEgiLOx1YIMQ==}
engines: {node: '>=6'}
hasBin: true
first commit
2026-02-21 03:10:39 +00:00
'@hono/node-server@1.19.9':
resolution: {integrity: sha512-vHL6w3ecZsky+8P5MD+eFfaGTyCeOHUIFYMGpQGbrBTSmNNoxv0if69rEZ5giu36weC5saFuznL411gRX7bJDw==}
engines: {node: '>=18.14.1'}
peerDependencies:
hono: ^4
'@humanfs/core@0.19.1':
resolution: {integrity: sha512-5DyQ4+1JEUzejeK1JGICcideyfUbGixgS9jNgex5nqkW+cY7WZhxBigmieN5Qnw9ZosSNVC9KQKyb+GUaGyKUA==}
engines: {node: '>=18.18.0'}
'@humanfs/node@0.16.7':
resolution: {integrity: sha512-/zUx+yOsIrG4Y43Eh2peDeKCxlRt/gET6aHfaKpuq267qXdYDFViVHfMaLyygZOnl0kGWxFIgsBy8QFuTLUXEQ==}
engines: {node: '>=18.18.0'}
'@humanwhocodes/module-importer@1.0.1':
resolution: {integrity: sha512-bxveV4V8v5Yb4ncFTT3rPSgZBOpCkjfK0y4oVVVJwIuDVBRMDXrPyXRL988i5ap9m9bnyEEjWfm5WkBmtffLfA==}
engines: {node: '>=12.22'}
'@humanwhocodes/retry@0.4.3':
resolution: {integrity: sha512-bV0Tgo9K4hfPCek+aMAn81RppFKv2ySDQeMoSZuvTASywNTnVJCArCZE2FWqpvIatKu7VMRLWlR1EazvVhDyhQ==}
engines: {node: '>=18.18'}
feat(web): bold redesign — Tailwind v4 + shadcn-style primitives + Skills/Proposals/Revisions UI Phase 6 of the Skills + Revisions + Proposals work. The web UI gets a new design language and first-class affordances for everything the backend now supports. ## Visual direction - Tailwind v4 with custom @theme block (oklch tokens). Dark-mode-only (internal tool — light mode doubles QA surface). - Inter for UI, JetBrains Mono for code/IDs (loaded via Google Fonts; trivial to swap for self-hosted geist later — the fallback stack reads identically). - Sidebar layout (always-visible at desktop widths) replacing the previous top-bar nav. Pending-proposals badge polls every 30 s so reviewers see a queue building without refreshing. - Lucide icons throughout. - Spacing and radii on Tailwind defaults. Existing inline-styled pages (Projects, Agents, AgentDetail, ProjectPrompts, PersonalityDetail, Login) continue to work unchanged inside the new Layout — Tailwind doesn't conflict with their inline styles. A follow-up can migrate them incrementally. ## What's added ### Build infra (src/web/) - package.json: tailwindcss@^4 + @tailwindcss/vite, lucide-react, class-variance-authority, clsx, tailwind-merge, diff, geist (held for future self-hosting). - vite.config.ts: registers the @tailwindcss/vite plugin. - src/index.css: Tailwind import + @theme tokens + @layer base. - src/main.tsx: imports index.css. - src/lib/utils.ts: shadcn-style cn() helper. ### shadcn-style primitives (src/components/ui/) Hand-written rather than generated via `npx shadcn` so the repo doesn't depend on a CLI tool that needs interactive runtime: - button.tsx — variants: primary / secondary / ghost / danger / link; sizes: sm / md / lg / icon. - card.tsx — Card + Header/Title/Description/Content/Footer subparts. - badge.tsx — variants: default / info / success / warning / danger / outline. - input.tsx — Input + Textarea + Label. - tabs.tsx — no-dep accessible Tabs (no Radix needed for our use). - separator.tsx — h/v separator with role=separator. ### Diff component (src/components/Diff.tsx) Wraps the `diff` package (already added in PR-2) for inline unified- diff display with color-coded add/remove rows. Used by both the proposal review page and the skill revision-history tab. ### New pages (src/pages/) - Dashboard.tsx — at-a-glance home. Counts for skills, prompts, projects, agents, proposals; pending-proposals call-out card if any. - Skills.tsx — list view, separated into Global vs Project/Agent- scoped sections. - SkillDetail.tsx — name + semver + description; tabs for SKILL.md / Files / Metadata / History. History tab shows revisions with click-to-diff against the live body. - Proposals.tsx — queue with Pending/Approved/Rejected tabs. Pending count is highlighted in amber. - ProposalDetail.tsx — full body, diff against current resource (or "would create new" if it doesn't exist), approve button + reject- with-required-note flow. ### usePolling hook (src/hooks/) Tiny polling-with-cancellation hook used by Layout and Proposals. ### Layout rewrite (src/components/Layout.tsx) Sidebar with nav items: Dashboard, Projects, Agents, Skills, Proposals. Lucide icons. Active-route highlighting via NavLink. Pending-proposals warning badge on the Proposals item. ### Routes (src/App.tsx) New routes: /dashboard, /skills, /skills/:name, /proposals, /proposals/:id. Default redirects to /dashboard. ### API types (src/api.ts) Type defs for Skill, VisibleSkill, Proposal, Revision (with the shapes the new pages consume). ## Tests Existing 7 web tests still pass (Login + api). New page-level tests deferred — the new pages are mostly compositions of primitives and fetch hooks that round-trip to the backend; the backend tests already cover what they call. PR-7 polish can add render-and-click tests if coverage drift surfaces. ## Verification - `pnpm --filter @mcpctl/web build` clean, no warnings. - `pnpm test:run` whole monorepo: 162 test files / 2157 tests green. - Visual smoke deferred — needs a running mcpd to populate the fixtures. Manual smoke tested locally is the next step. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-07 17:54:55 +01:00
'@img/colour@1.1.0':
resolution: {integrity: sha512-Td76q7j57o/tLVdgS746cYARfSyxk8iEfRxewL9h4OMzYhbW4TAcppl0mT4eyqXddh6L/jwoM75mo7ixa/pCeQ==}
engines: {node: '>=18'}
'@img/sharp-darwin-arm64@0.34.5':
resolution: {integrity: sha512-imtQ3WMJXbMY4fxb/Ndp6HBTNVtWCUI0WdobyheGf5+ad6xX8VIDO8u2xE4qc/fr08CKG/7dDseFtn6M6g/r3w==}
engines: {node: ^18.17.0 || ^20.3.0 || >=21.0.0}
cpu: [arm64]
os: [darwin]
'@img/sharp-darwin-x64@0.34.5':
resolution: {integrity: sha512-YNEFAF/4KQ/PeW0N+r+aVVsoIY0/qxxikF2SWdp+NRkmMB7y9LBZAVqQ4yhGCm/H3H270OSykqmQMKLBhBJDEw==}
engines: {node: ^18.17.0 || ^20.3.0 || >=21.0.0}
cpu: [x64]
os: [darwin]
'@img/sharp-libvips-darwin-arm64@1.2.4':
resolution: {integrity: sha512-zqjjo7RatFfFoP0MkQ51jfuFZBnVE2pRiaydKJ1G/rHZvnsrHAOcQALIi9sA5co5xenQdTugCvtb1cuf78Vf4g==}
cpu: [arm64]
os: [darwin]
'@img/sharp-libvips-darwin-x64@1.2.4':
resolution: {integrity: sha512-1IOd5xfVhlGwX+zXv2N93k0yMONvUlANylbJw1eTah8K/Jtpi15KC+WSiaX/nBmbm2HxRM1gZ0nSdjSsrZbGKg==}
cpu: [x64]
os: [darwin]
'@img/sharp-libvips-linux-arm64@1.2.4':
resolution: {integrity: sha512-excjX8DfsIcJ10x1Kzr4RcWe1edC9PquDRRPx3YVCvQv+U5p7Yin2s32ftzikXojb1PIFc/9Mt28/y+iRklkrw==}
cpu: [arm64]
os: [linux]
'@img/sharp-libvips-linux-arm@1.2.4':
resolution: {integrity: sha512-bFI7xcKFELdiNCVov8e44Ia4u2byA+l3XtsAj+Q8tfCwO6BQ8iDojYdvoPMqsKDkuoOo+X6HZA0s0q11ANMQ8A==}
cpu: [arm]
os: [linux]
'@img/sharp-libvips-linux-ppc64@1.2.4':
resolution: {integrity: sha512-FMuvGijLDYG6lW+b/UvyilUWu5Ayu+3r2d1S8notiGCIyYU/76eig1UfMmkZ7vwgOrzKzlQbFSuQfgm7GYUPpA==}
cpu: [ppc64]
os: [linux]
'@img/sharp-libvips-linux-riscv64@1.2.4':
resolution: {integrity: sha512-oVDbcR4zUC0ce82teubSm+x6ETixtKZBh/qbREIOcI3cULzDyb18Sr/Wcyx7NRQeQzOiHTNbZFF1UwPS2scyGA==}
cpu: [riscv64]
os: [linux]
'@img/sharp-libvips-linux-s390x@1.2.4':
resolution: {integrity: sha512-qmp9VrzgPgMoGZyPvrQHqk02uyjA0/QrTO26Tqk6l4ZV0MPWIW6LTkqOIov+J1yEu7MbFQaDpwdwJKhbJvuRxQ==}
cpu: [s390x]
os: [linux]
'@img/sharp-libvips-linux-x64@1.2.4':
resolution: {integrity: sha512-tJxiiLsmHc9Ax1bz3oaOYBURTXGIRDODBqhveVHonrHJ9/+k89qbLl0bcJns+e4t4rvaNBxaEZsFtSfAdquPrw==}
cpu: [x64]
os: [linux]
'@img/sharp-libvips-linuxmusl-arm64@1.2.4':
resolution: {integrity: sha512-FVQHuwx1IIuNow9QAbYUzJ+En8KcVm9Lk5+uGUQJHaZmMECZmOlix9HnH7n1TRkXMS0pGxIJokIVB9SuqZGGXw==}
cpu: [arm64]
os: [linux]
'@img/sharp-libvips-linuxmusl-x64@1.2.4':
resolution: {integrity: sha512-+LpyBk7L44ZIXwz/VYfglaX/okxezESc6UxDSoyo2Ks6Jxc4Y7sGjpgU9s4PMgqgjj1gZCylTieNamqA1MF7Dg==}
cpu: [x64]
os: [linux]
'@img/sharp-linux-arm64@0.34.5':
resolution: {integrity: sha512-bKQzaJRY/bkPOXyKx5EVup7qkaojECG6NLYswgktOZjaXecSAeCWiZwwiFf3/Y+O1HrauiE3FVsGxFg8c24rZg==}
engines: {node: ^18.17.0 || ^20.3.0 || >=21.0.0}
cpu: [arm64]
os: [linux]
'@img/sharp-linux-arm@0.34.5':
resolution: {integrity: sha512-9dLqsvwtg1uuXBGZKsxem9595+ujv0sJ6Vi8wcTANSFpwV/GONat5eCkzQo/1O6zRIkh0m/8+5BjrRr7jDUSZw==}
engines: {node: ^18.17.0 || ^20.3.0 || >=21.0.0}
cpu: [arm]
os: [linux]
'@img/sharp-linux-ppc64@0.34.5':
resolution: {integrity: sha512-7zznwNaqW6YtsfrGGDA6BRkISKAAE1Jo0QdpNYXNMHu2+0dTrPflTLNkpc8l7MUP5M16ZJcUvysVWWrMefZquA==}
engines: {node: ^18.17.0 || ^20.3.0 || >=21.0.0}
cpu: [ppc64]
os: [linux]
'@img/sharp-linux-riscv64@0.34.5':
resolution: {integrity: sha512-51gJuLPTKa7piYPaVs8GmByo7/U7/7TZOq+cnXJIHZKavIRHAP77e3N2HEl3dgiqdD/w0yUfiJnII77PuDDFdw==}
engines: {node: ^18.17.0 || ^20.3.0 || >=21.0.0}
cpu: [riscv64]
os: [linux]
'@img/sharp-linux-s390x@0.34.5':
resolution: {integrity: sha512-nQtCk0PdKfho3eC5MrbQoigJ2gd1CgddUMkabUj+rBevs8tZ2cULOx46E7oyX+04WGfABgIwmMC0VqieTiR4jg==}
engines: {node: ^18.17.0 || ^20.3.0 || >=21.0.0}
cpu: [s390x]
os: [linux]
'@img/sharp-linux-x64@0.34.5':
resolution: {integrity: sha512-MEzd8HPKxVxVenwAa+JRPwEC7QFjoPWuS5NZnBt6B3pu7EG2Ge0id1oLHZpPJdn3OQK+BQDiw9zStiHBTJQQQQ==}
engines: {node: ^18.17.0 || ^20.3.0 || >=21.0.0}
cpu: [x64]
os: [linux]
'@img/sharp-linuxmusl-arm64@0.34.5':
resolution: {integrity: sha512-fprJR6GtRsMt6Kyfq44IsChVZeGN97gTD331weR1ex1c1rypDEABN6Tm2xa1wE6lYb5DdEnk03NZPqA7Id21yg==}
engines: {node: ^18.17.0 || ^20.3.0 || >=21.0.0}
cpu: [arm64]
os: [linux]
'@img/sharp-linuxmusl-x64@0.34.5':
resolution: {integrity: sha512-Jg8wNT1MUzIvhBFxViqrEhWDGzqymo3sV7z7ZsaWbZNDLXRJZoRGrjulp60YYtV4wfY8VIKcWidjojlLcWrd8Q==}
engines: {node: ^18.17.0 || ^20.3.0 || >=21.0.0}
cpu: [x64]
os: [linux]
'@img/sharp-wasm32@0.34.5':
resolution: {integrity: sha512-OdWTEiVkY2PHwqkbBI8frFxQQFekHaSSkUIJkwzclWZe64O1X4UlUjqqqLaPbUpMOQk6FBu/HtlGXNblIs0huw==}
engines: {node: ^18.17.0 || ^20.3.0 || >=21.0.0}
cpu: [wasm32]
'@img/sharp-win32-arm64@0.34.5':
resolution: {integrity: sha512-WQ3AgWCWYSb2yt+IG8mnC6Jdk9Whs7O0gxphblsLvdhSpSTtmu69ZG1Gkb6NuvxsNACwiPV6cNSZNzt0KPsw7g==}
engines: {node: ^18.17.0 || ^20.3.0 || >=21.0.0}
cpu: [arm64]
os: [win32]
'@img/sharp-win32-ia32@0.34.5':
resolution: {integrity: sha512-FV9m/7NmeCmSHDD5j4+4pNI8Cp3aW+JvLoXcTUo0IqyjSfAZJ8dIUmijx1qaJsIiU+Hosw6xM5KijAWRJCSgNg==}
engines: {node: ^18.17.0 || ^20.3.0 || >=21.0.0}
cpu: [ia32]
os: [win32]
'@img/sharp-win32-x64@0.34.5':
resolution: {integrity: sha512-+29YMsqY2/9eFEiW93eqWnuLcWcufowXewwSNIT6UwZdUUCrM3oFjMWH/Z6/TMmb4hlFenmfAVbpWeup2jryCw==}
engines: {node: ^18.17.0 || ^20.3.0 || >=21.0.0}
cpu: [x64]
os: [win32]
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
'@inkjs/ui@2.0.0':
resolution: {integrity: sha512-5+8fJmwtF9UvikzLfph9sA+LS+l37Ij/szQltkuXLOAXwNkBX9innfzh4pLGXIB59vKEQUtc6D4qGvhD7h3pAg==}
engines: {node: '>=18'}
peerDependencies:
ink: '>=5'
first commit
2026-02-21 03:10:39 +00:00
'@inquirer/ansi@1.0.2':
resolution: {integrity: sha512-S8qNSZiYzFd0wAcyG5AXCvUHC5Sr7xpZ9wZ2py9XR88jUz8wooStVx5M6dRzczbBWjic9NP7+rY0Xi7qqK/aMQ==}
engines: {node: '>=18'}
'@inquirer/checkbox@4.3.2':
resolution: {integrity: sha512-VXukHf0RR1doGe6Sm4F0Em7SWYLTHSsbGfJdS9Ja2bX5/D5uwVOEjr07cncLROdBvmnvCATYEWlHqYmXv2IlQA==}
engines: {node: '>=18'}
peerDependencies:
'@types/node': '>=18'
peerDependenciesMeta:
'@types/node':
optional: true
'@inquirer/confirm@5.1.21':
resolution: {integrity: sha512-KR8edRkIsUayMXV+o3Gv+q4jlhENF9nMYUZs9PA2HzrXeHI8M5uDag70U7RJn9yyiMZSbtF5/UexBtAVtZGSbQ==}
engines: {node: '>=18'}
peerDependencies:
'@types/node': '>=18'
peerDependenciesMeta:
'@types/node':
optional: true
'@inquirer/core@10.3.2':
resolution: {integrity: sha512-43RTuEbfP8MbKzedNqBrlhhNKVwoK//vUFNW3Q3vZ88BLcrs4kYpGg+B2mm5p2K/HfygoCxuKwJJiv8PbGmE0A==}
engines: {node: '>=18'}
peerDependencies:
'@types/node': '>=18'
peerDependenciesMeta:
'@types/node':
optional: true
'@inquirer/editor@4.2.23':
resolution: {integrity: sha512-aLSROkEwirotxZ1pBaP8tugXRFCxW94gwrQLxXfrZsKkfjOYC1aRvAZuhpJOb5cu4IBTJdsCigUlf2iCOu4ZDQ==}
engines: {node: '>=18'}
peerDependencies:
'@types/node': '>=18'
peerDependenciesMeta:
'@types/node':
optional: true
'@inquirer/expand@4.0.23':
resolution: {integrity: sha512-nRzdOyFYnpeYTTR2qFwEVmIWypzdAx/sIkCMeTNTcflFOovfqUk+HcFhQQVBftAh9gmGrpFj6QcGEqrDMDOiew==}
engines: {node: '>=18'}
peerDependencies:
'@types/node': '>=18'
peerDependenciesMeta:
'@types/node':
optional: true
'@inquirer/external-editor@1.0.3':
resolution: {integrity: sha512-RWbSrDiYmO4LbejWY7ttpxczuwQyZLBUyygsA9Nsv95hpzUWwnNTVQmAq3xuh7vNwCp07UTmE5i11XAEExx4RA==}
engines: {node: '>=18'}
peerDependencies:
'@types/node': '>=18'
peerDependenciesMeta:
'@types/node':
optional: true
'@inquirer/figures@1.0.15':
resolution: {integrity: sha512-t2IEY+unGHOzAaVM5Xx6DEWKeXlDDcNPeDyUpsRc6CUhBfU3VQOEl+Vssh7VNp1dR8MdUJBWhuObjXCsVpjN5g==}
engines: {node: '>=18'}
'@inquirer/input@4.3.1':
resolution: {integrity: sha512-kN0pAM4yPrLjJ1XJBjDxyfDduXOuQHrBB8aLDMueuwUGn+vNpF7Gq7TvyVxx8u4SHlFFj4trmj+a2cbpG4Jn1g==}
engines: {node: '>=18'}
peerDependencies:
'@types/node': '>=18'
peerDependenciesMeta:
'@types/node':
optional: true
'@inquirer/number@3.0.23':
resolution: {integrity: sha512-5Smv0OK7K0KUzUfYUXDXQc9jrf8OHo4ktlEayFlelCjwMXz0299Y8OrI+lj7i4gCBY15UObk76q0QtxjzFcFcg==}
engines: {node: '>=18'}
peerDependencies:
'@types/node': '>=18'
peerDependenciesMeta:
'@types/node':
optional: true
'@inquirer/password@4.0.23':
resolution: {integrity: sha512-zREJHjhT5vJBMZX/IUbyI9zVtVfOLiTO66MrF/3GFZYZ7T4YILW5MSkEYHceSii/KtRk+4i3RE7E1CUXA2jHcA==}
engines: {node: '>=18'}
peerDependencies:
'@types/node': '>=18'
peerDependenciesMeta:
'@types/node':
optional: true
'@inquirer/prompts@7.10.1':
resolution: {integrity: sha512-Dx/y9bCQcXLI5ooQ5KyvA4FTgeo2jYj/7plWfV5Ak5wDPKQZgudKez2ixyfz7tKXzcJciTxqLeK7R9HItwiByg==}
engines: {node: '>=18'}
peerDependencies:
'@types/node': '>=18'
peerDependenciesMeta:
'@types/node':
optional: true
'@inquirer/rawlist@4.1.11':
resolution: {integrity: sha512-+LLQB8XGr3I5LZN/GuAHo+GpDJegQwuPARLChlMICNdwW7OwV2izlCSCxN6cqpL0sMXmbKbFcItJgdQq5EBXTw==}
engines: {node: '>=18'}
peerDependencies:
'@types/node': '>=18'
peerDependenciesMeta:
'@types/node':
optional: true
'@inquirer/search@3.2.2':
resolution: {integrity: sha512-p2bvRfENXCZdWF/U2BXvnSI9h+tuA8iNqtUKb9UWbmLYCRQxd8WkvwWvYn+3NgYaNwdUkHytJMGG4MMLucI1kA==}
engines: {node: '>=18'}
peerDependencies:
'@types/node': '>=18'
peerDependenciesMeta:
'@types/node':
optional: true
'@inquirer/select@4.4.2':
resolution: {integrity: sha512-l4xMuJo55MAe+N7Qr4rX90vypFwCajSakx59qe/tMaC1aEHWLyw68wF4o0A4SLAY4E0nd+Vt+EyskeDIqu1M6w==}
engines: {node: '>=18'}
peerDependencies:
'@types/node': '>=18'
peerDependenciesMeta:
'@types/node':
optional: true
'@inquirer/type@3.0.10':
resolution: {integrity: sha512-BvziSRxfz5Ov8ch0z/n3oijRSEcEsHnhggm4xFZe93DHcUCTlutlq9Ox4SVENAfcRD22UQq7T/atg9Wr3k09eA==}
engines: {node: '>=18'}
peerDependencies:
'@types/node': '>=18'
peerDependenciesMeta:
'@types/node':
optional: true
feat(mcpd+deploy): serve web UI at /ui + smoke tests + docs (Stage 6) The closing stage. mcpd now hosts the Stage 5 SPA, the Docker image bundles the build artifact, a smoke test exercises the personality HTTP surface end-to-end, and the user-facing docs spell out the mental model. mcpd: - Add @fastify/static dep. - New routes/web-ui.ts: registers /ui/* against a static bundle. Looks for the bundle at $MCPD_WEB_ROOT, then /usr/share/mcpd/web (the Docker image path), then a dev-tree fallback. Logs and skips cleanly if missing — API-only deploys keep working. - SPA fallback: any /ui/<path> that doesn't match a file falls through to index.html so direct hits to react-router URLs work. - /ui/* falls through to `kind: skip` in mapUrlToPermission, so the static assets are served unauthenticated. Each API call from the SPA still carries the bearer token. Deploy: - Dockerfile.mcpd builds the @mcpctl/web bundle in the same builder stage and copies dist/ to /usr/share/mcpd/web in the runtime image. Smoke (personality.smoke.test.ts): - Live mcpd flow: create secret/llm/agent/personality, attach an agent-direct prompt, verify the binding listing, reject double- attach (409) + foreign-agent prompt (400), set defaultPersonality by name, detach + delete cleanup. Docs: - New docs/personalities.md: VLAN-on-ethernet model, system-block ordering table, three prompt scopes, CLI walkthrough, web UI walkthrough, full API surface, RBAC notes. - agents.md and chat.md cross-link. - README's Agents section gains a Personalities subsection. Test count after Stage 6: mcpd: 801/801 cli: 430/430 web: 7/7 db: 58/62 (4 pre-existing) Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:48:43 +01:00
'@isaacs/cliui@9.0.0':
resolution: {integrity: sha512-AokJm4tuBHillT+FpMtxQ60n8ObyXBatq7jD2/JA9dxbDDokKQm8KMht5ibGzLVU9IJDIKK4TPKgMHEYMn3lMg==}
engines: {node: '>=18'}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
'@isaacs/fs-minipass@4.0.1':
resolution: {integrity: sha512-wgm9Ehl2jpeqP3zw/7mo3kRHFp5MEDhqAdwy1fTGkHAwnkGOVsgpvQhL8B5n1qlb01jV3n/bI0ZfZp5lWA1k4w==}
engines: {node: '>=18.0.0'}
'@isaacs/string-locale-compare@1.1.0':
resolution: {integrity: sha512-SQ7Kzhh9+D+ZW9MA0zkYv3VXhIDNx+LzM6EJ+/65I3QY+enU6Itte7E5XX7EWrqLW2FN4n06GWzBnPoC3th2aQ==}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
'@jridgewell/gen-mapping@0.3.13':
resolution: {integrity: sha512-2kkt/7niJ6MgEPxF0bYdQ6etZaA+fQvDcLKckhy1yIQOzaoKjBBjSj63/aLVjYE3qhRt5dvM+uUyfCg6UKCBbA==}
'@jridgewell/remapping@2.3.5':
resolution: {integrity: sha512-LI9u/+laYG4Ds1TDKSJW2YPrIlcVYOwi2fUC6xB43lueCjgxV4lffOCZCtYFiH6TNOX+tQKXx97T4IKHbhyHEQ==}
first commit
2026-02-21 03:10:39 +00:00
'@jridgewell/resolve-uri@3.1.2':
resolution: {integrity: sha512-bRISgCIjP20/tbWSPWMEi54QVPRZExkuD9lJL+UIxUKtwVJA8wW1Trb1jMs1RFXo1CBTNZ/5hpC9QvmKWdopKw==}
engines: {node: '>=6.0.0'}
'@jridgewell/sourcemap-codec@1.5.5':
resolution: {integrity: sha512-cYQ9310grqxueWbl+WuIUIaiUaDcj7WOq5fVhEljNVgRfOUhY9fy2zTvfoqWsnebh8Sl70VScFbICvJnLKB0Og==}
'@jridgewell/trace-mapping@0.3.31':
resolution: {integrity: sha512-zzNR+SdQSDJzc8joaeP8QQoCQr8NuYx2dIIytl1QeBEZHJ9uW6hebsrYgbz8hJwUQao3TWCMtmfV8Nu1twOLAw==}
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
'@js-sdsl/ordered-map@4.4.2':
resolution: {integrity: sha512-iUKgm52T8HOE/makSxjqoWhe95ZJA1/G1sYsGev2JDKUSS14KAgg1LHb+Ba+IPow0xflbnSkOsZcO08C7w1gYw==}
feat: add Kubernetes orchestrator for MCP server pod management mcpd can now deploy MCP server instances as Kubernetes pods instead of Docker containers. Set MCPD_ORCHESTRATOR=kubernetes to enable. - Add @kubernetes/client-node with thin wrapper (context enforcement via MCPD_K8S_CONTEXT to prevent multi-cluster mishaps) - Rewrite KubernetesOrchestrator: pod CRUD, pod IP extraction, exec via SPDY (one-shot + interactive), log streaming - Manifest generator: stdin:true for STDIO servers, args (not command) to preserve runner image entrypoint, security hardening - Orchestrator selection in main.ts via MCPD_ORCHESTRATOR env var - 25 unit tests for k8s orchestrator, all 624 tests pass Tested end-to-end on local k3s: - mcpd deployed via Pulumi, creates pods in mcpctl-servers namespace - NetworkPolicy verified: only mcpd can reach MCP server pods - Python runner (uvx) successfully runs aws-documentation-mcp-server Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-08 01:55:13 +01:00
'@jsep-plugin/assignment@1.3.0':
resolution: {integrity: sha512-VVgV+CXrhbMI3aSusQyclHkenWSAm95WaiKrMxRFam3JSUiIaQjoMIw2sEs/OX4XifnqeQUN4DYbJjlA8EfktQ==}
engines: {node: '>= 10.16.0'}
peerDependencies:
jsep: ^0.4.0||^1.0.0
'@jsep-plugin/regex@1.0.4':
resolution: {integrity: sha512-q7qL4Mgjs1vByCaTnDFcBnV9HS7GVPJX5vyVoCgZHNSC9rjwIlmbXG5sUuorR5ndfHAIlJ8pVStxvjXHbNvtUg==}
engines: {node: '>= 10.16.0'}
peerDependencies:
jsep: ^0.4.0||^1.0.0
'@kubernetes/client-node@1.4.0':
resolution: {integrity: sha512-Zge3YvF7DJi264dU1b3wb/GmzR99JhUpqTvp+VGHfwZT+g7EOOYNScDJNZwXy9cszyIGPIs0VHr+kk8e95qqrA==}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
'@logdna/tail-file@2.2.0':
resolution: {integrity: sha512-XGSsWDweP80Fks16lwkAUIr54ICyBs6PsI4mpfTLQaWgEJRtY9xEV+PeyDpJ+sJEGZxqINlpmAwe/6tS1pP8Ng==}
engines: {node: '>=10.3.0'}
first commit
2026-02-21 03:10:39 +00:00
'@lukeed/ms@2.0.2':
resolution: {integrity: sha512-9I2Zn6+NJLfaGoz9jN3lpwDgAYvfGeNYdbAIjJOqzs4Tpc+VU3Jqq4IofSUBKajiDS8k9fZIg18/z13mpk1bsA==}
engines: {node: '>=8'}
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
'@mapbox/node-pre-gyp@1.0.11':
resolution: {integrity: sha512-Yhlar6v9WQgUp/He7BdgzOz8lqMQ8sU+jkCq7Wx8Myc5YFJLbEe7lgui/V7G1qB1DJykHSGwreceSaD60Y0PUQ==}
hasBin: true
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
'@mariozechner/clipboard-darwin-arm64@0.3.9':
resolution: {integrity: sha512-BfgV7vCEWZwJwZJw03r6bP5+tf0iI/ANuQYCxi9RNn7FrWB3yzGuMKCrNLRl6V761vXRdL8+OqZ0wd4TqlsNOQ==}
engines: {node: '>= 10'}
cpu: [arm64]
os: [darwin]
'@mariozechner/clipboard-darwin-universal@0.3.9':
resolution: {integrity: sha512-BGGR4iA9Z2shAjI65eI5xtyb3LYNlDW9X3gxKxDbqtbnREohsrqznov6zpKoIrsRWpzlYVEdKphS7ksJ0/ndSQ==}
engines: {node: '>= 10'}
os: [darwin]
'@mariozechner/clipboard-darwin-x64@0.3.9':
resolution: {integrity: sha512-4kURmCbS6nt8uYhtmWpUcJWyPHfmAr5dTpXD1nO3pIfa+TSQ9DbrGOYCKH+aEFW47XhQ4Vp8ZTszie+wfFvDKg==}
engines: {node: '>= 10'}
cpu: [x64]
os: [darwin]
'@mariozechner/clipboard-linux-arm64-gnu@0.3.9':
resolution: {integrity: sha512-g59OkUGP2DDfCOIKypHeYgv2M55u/cKvXa5dSxFbEJ34XvIQMdcVmpKCkGUro3ZgefXiGVdwguvTMQGpHWzIXw==}
engines: {node: '>= 10'}
cpu: [arm64]
os: [linux]
'@mariozechner/clipboard-linux-arm64-musl@0.3.9':
resolution: {integrity: sha512-AGuJdgKsmJdm4Pych7kv3sqe591ERRaAHW3xjLooiFzn8J+PxUyof++7YZrB5Y5tpnTO+K18Og3taj2NpluCRQ==}
engines: {node: '>= 10'}
cpu: [arm64]
os: [linux]
'@mariozechner/clipboard-linux-riscv64-gnu@0.3.9':
resolution: {integrity: sha512-DXBEAiuMpk7dhS1a9NzNxVAFi1vaKoPu7rQNgY8LIDLGrK3lnIp3nT10DUum+PKVJoJppIP+NAA8IZe4DMNDPw==}
engines: {node: '>= 10'}
cpu: [riscv64]
os: [linux]
'@mariozechner/clipboard-linux-x64-gnu@0.3.9':
resolution: {integrity: sha512-WORrMLd6EpElEME7JRKfSaY34nW1P5LbdgK5YNCS1ncG2LqmITsSMEJ8nh2mpvxb3TxqbOOKgY7k9eMJYlW9Mw==}
engines: {node: '>= 10'}
cpu: [x64]
os: [linux]
'@mariozechner/clipboard-linux-x64-musl@0.3.9':
resolution: {integrity: sha512-/DHn+1DrfL6oRaPPWXaOKvonFFrni666fxd+zFqiQEfvBH0tsHVWjq9iqBk0oDp0qaPA72lIMy5BptxISBEhZQ==}
engines: {node: '>= 10'}
cpu: [x64]
os: [linux]
'@mariozechner/clipboard-win32-arm64-msvc@0.3.9':
resolution: {integrity: sha512-O5FHD3ErkMwMhNzAfu3ggy0ug4z7btZuoQgwwxlzPrwV2bxlD6WDpqBY4NCgICAgZdDKdp+loUEKVAVt8aYnhQ==}
engines: {node: '>= 10'}
cpu: [arm64]
os: [win32]
'@mariozechner/clipboard-win32-x64-msvc@0.3.9':
resolution: {integrity: sha512-ihQC3EufqEY81vhXBgVBtK4prL+wc62zJsSvxrgz7K1hsdt6OObz6v9p3Rn1OG3GJksTTKMJF0u/guMISHPhSA==}
engines: {node: '>= 10'}
cpu: [x64]
os: [win32]
'@mariozechner/clipboard@0.3.9':
resolution: {integrity: sha512-ABnA53mdfkGZwOFUdZNv2S0CWGO/EIuPj8Vv9xmBFmSYg/qFc7ihO6q5FcQjvoE67kZpWkEc4AhD6B/os04yuA==}
engines: {node: '>= 10'}
'@mistralai/mistralai@2.2.6':
resolution: {integrity: sha512-W8pX7zHxjJvMIpw8JMxeJEleapXX0Q9NPszdNzqkM3MIEoIGPObdodujj+WHteXEvGfaP/AMwlNyRfEzSY6dQQ==}
peerDependencies:
'@opentelemetry/api': ^1.9.0
peerDependenciesMeta:
'@opentelemetry/api':
optional: true
first commit
2026-02-21 03:10:39 +00:00
'@modelcontextprotocol/sdk@1.26.0':
resolution: {integrity: sha512-Y5RmPncpiDtTXDbLKswIJzTqu2hyBKxTNsgKqKclDbhIgg1wgtf1fRuvxgTnRfcnxtvvgbIEcqUOzZrJ6iSReg==}
engines: {node: '>=18'}
peerDependencies:
'@cfworker/json-schema': ^4.1.1
zod: ^3.25 || ^4.0
peerDependenciesMeta:
'@cfworker/json-schema':
optional: true
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
'@monaco-editor/loader@1.7.0':
resolution: {integrity: sha512-gIwR1HrJrrx+vfyOhYmCZ0/JcWqG5kbfG7+d3f/C1LXk2EvzAbHSg3MQ5lO2sMlo9izoAZ04shohfKLVT6crVA==}
'@monaco-editor/react@4.7.0':
resolution: {integrity: sha512-cyzXQCtO47ydzxpQtCGSQGOC8Gk3ZUeBXFAxD+CWXYFo5OqZyZUonFl0DwUlTyAfRHntBfw2p3w4s9R6oe1eCA==}
peerDependencies:
monaco-editor: '>= 0.25.0 < 1'
react: ^16.8.0 || ^17.0.0 || ^18.0.0 || ^19.0.0
react-dom: ^16.8.0 || ^17.0.0 || ^18.0.0 || ^19.0.0
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
'@msgpackr-extract/msgpackr-extract-darwin-arm64@3.0.4':
resolution: {integrity: sha512-LCkGo6JDfaBhgST7UpPWgNgLINpcpabaHfyz5OBx75nUYxBsaEPxjnyNjWpeb/xBup/682QnBfRBy2/LvPutZQ==}
cpu: [arm64]
os: [darwin]
'@msgpackr-extract/msgpackr-extract-darwin-x64@3.0.4':
resolution: {integrity: sha512-zExlW9zUJKZH/tOtVMttwjKa4Xm/3KcNjnE3dPN92uCktwavMxpgCA3MoJK/DOnTWsQgo224OaST27/mPNAf+w==}
cpu: [x64]
os: [darwin]
'@msgpackr-extract/msgpackr-extract-linux-arm64@3.0.4':
resolution: {integrity: sha512-dgX0P/9wGPJeHFBG+ZmhgE6bmtMt7NP5CRBGyyktpopdk/mW4POnrpQsSLtKI1dwpc+pPLuXHDh6vvskyQE/sw==}
cpu: [arm64]
os: [linux]
'@msgpackr-extract/msgpackr-extract-linux-arm@3.0.4':
resolution: {integrity: sha512-Tg3yX65f5GbtXLkrYEHE5oibZG9epyYWas7FogTTEJeDEF9JlXJzKgXaNhT3UXlTOeA+AfZpYZYZ0uPj7Cfquw==}
cpu: [arm]
os: [linux]
'@msgpackr-extract/msgpackr-extract-linux-x64@3.0.4':
resolution: {integrity: sha512-8TNXMEjJc3QEy7R/x1INhgiU+XakDAFUzBhaz7+Rbrs8NH5UQeHQxxmzsSBJGyV6I1jW79undiQm8tOI+D+8FQ==}
cpu: [x64]
os: [linux]
'@msgpackr-extract/msgpackr-extract-win32-x64@3.0.4':
resolution: {integrity: sha512-CmCXPQrkbwExx3j946/PtHWHbYJiCRBRDl4BlkRQcJB/YOwQxJRTpoo7aTsortjgoJ1x7opzTSxn7C+ASSLVjQ==}
cpu: [x64]
os: [win32]
feat(web): bold redesign — Tailwind v4 + shadcn-style primitives + Skills/Proposals/Revisions UI Phase 6 of the Skills + Revisions + Proposals work. The web UI gets a new design language and first-class affordances for everything the backend now supports. ## Visual direction - Tailwind v4 with custom @theme block (oklch tokens). Dark-mode-only (internal tool — light mode doubles QA surface). - Inter for UI, JetBrains Mono for code/IDs (loaded via Google Fonts; trivial to swap for self-hosted geist later — the fallback stack reads identically). - Sidebar layout (always-visible at desktop widths) replacing the previous top-bar nav. Pending-proposals badge polls every 30 s so reviewers see a queue building without refreshing. - Lucide icons throughout. - Spacing and radii on Tailwind defaults. Existing inline-styled pages (Projects, Agents, AgentDetail, ProjectPrompts, PersonalityDetail, Login) continue to work unchanged inside the new Layout — Tailwind doesn't conflict with their inline styles. A follow-up can migrate them incrementally. ## What's added ### Build infra (src/web/) - package.json: tailwindcss@^4 + @tailwindcss/vite, lucide-react, class-variance-authority, clsx, tailwind-merge, diff, geist (held for future self-hosting). - vite.config.ts: registers the @tailwindcss/vite plugin. - src/index.css: Tailwind import + @theme tokens + @layer base. - src/main.tsx: imports index.css. - src/lib/utils.ts: shadcn-style cn() helper. ### shadcn-style primitives (src/components/ui/) Hand-written rather than generated via `npx shadcn` so the repo doesn't depend on a CLI tool that needs interactive runtime: - button.tsx — variants: primary / secondary / ghost / danger / link; sizes: sm / md / lg / icon. - card.tsx — Card + Header/Title/Description/Content/Footer subparts. - badge.tsx — variants: default / info / success / warning / danger / outline. - input.tsx — Input + Textarea + Label. - tabs.tsx — no-dep accessible Tabs (no Radix needed for our use). - separator.tsx — h/v separator with role=separator. ### Diff component (src/components/Diff.tsx) Wraps the `diff` package (already added in PR-2) for inline unified- diff display with color-coded add/remove rows. Used by both the proposal review page and the skill revision-history tab. ### New pages (src/pages/) - Dashboard.tsx — at-a-glance home. Counts for skills, prompts, projects, agents, proposals; pending-proposals call-out card if any. - Skills.tsx — list view, separated into Global vs Project/Agent- scoped sections. - SkillDetail.tsx — name + semver + description; tabs for SKILL.md / Files / Metadata / History. History tab shows revisions with click-to-diff against the live body. - Proposals.tsx — queue with Pending/Approved/Rejected tabs. Pending count is highlighted in amber. - ProposalDetail.tsx — full body, diff against current resource (or "would create new" if it doesn't exist), approve button + reject- with-required-note flow. ### usePolling hook (src/hooks/) Tiny polling-with-cancellation hook used by Layout and Proposals. ### Layout rewrite (src/components/Layout.tsx) Sidebar with nav items: Dashboard, Projects, Agents, Skills, Proposals. Lucide icons. Active-route highlighting via NavLink. Pending-proposals warning badge on the Proposals item. ### Routes (src/App.tsx) New routes: /dashboard, /skills, /skills/:name, /proposals, /proposals/:id. Default redirects to /dashboard. ### API types (src/api.ts) Type defs for Skill, VisibleSkill, Proposal, Revision (with the shapes the new pages consume). ## Tests Existing 7 web tests still pass (Login + api). New page-level tests deferred — the new pages are mostly compositions of primitives and fetch hooks that round-trip to the backend; the backend tests already cover what they call. PR-7 polish can add render-and-click tests if coverage drift surfaces. ## Verification - `pnpm --filter @mcpctl/web build` clean, no warnings. - `pnpm test:run` whole monorepo: 162 test files / 2157 tests green. - Visual smoke deferred — needs a running mcpd to populate the fixtures. Manual smoke tested locally is the next step. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-07 17:54:55 +01:00
'@next/env@16.2.5':
resolution: {integrity: sha512-Lb9ElHD2klcyeVD25vW+siPFqz9QMzDUSgvFZNO+dZEKoMHex4viJhVuzBhrXKqb+UKnih7mVYbt50/7KLsSCA==}
'@next/swc-darwin-arm64@16.2.5':
resolution: {integrity: sha512-BW+8PGVmsruomXHsitD8JG6gny9lEdobctjBwvtPF8AKtxGDR7nR35FOl/oK9UAPXBOBm+vx0k8qtpeHOXQMGQ==}
engines: {node: '>= 10'}
cpu: [arm64]
os: [darwin]
'@next/swc-darwin-x64@16.2.5':
resolution: {integrity: sha512-ZoCGnCl9LlQJWmqXrZAUlNxvuNmclvE+7zUif+nDydkkehl9FKxHJ+wxSQMj+C37BYFerKiEdX9s9o02ir975Q==}
engines: {node: '>= 10'}
cpu: [x64]
os: [darwin]
'@next/swc-linux-arm64-gnu@16.2.5':
resolution: {integrity: sha512-AwcZzMChaWkOTZt3vu+2ZMIj8g4dYQY+B8VUVhlFSQ2JtvyZpefyYHTe00D6b6L7BysYw7vl3zsvs9jix8tl5Q==}
engines: {node: '>= 10'}
cpu: [arm64]
os: [linux]
'@next/swc-linux-arm64-musl@16.2.5':
resolution: {integrity: sha512-QqMgqWbCBFsfiQ7BF3dUlW8HJy1LWhpcqbTpoHMWA9IV+TnWwDKozQJA5NdIAHjQ00yX2Q7AUkLr/XK4n77q8A==}
engines: {node: '>= 10'}
cpu: [arm64]
os: [linux]
'@next/swc-linux-x64-gnu@16.2.5':
resolution: {integrity: sha512-3hzeiFGZtyATVx9pCeuzTshXmh50vHZitqaeZiyJZaUmjQyrfjsVUgS8apOj1vEJCIpKJM/55F45yPAV2kpjsA==}
engines: {node: '>= 10'}
cpu: [x64]
os: [linux]
'@next/swc-linux-x64-musl@16.2.5':
resolution: {integrity: sha512-0mzZV/mAt7Qj2tYNdTB6AqrS8dwng/AQLSYC5Z1YLpZdi2wxqKDPK7RY2RvjB1fXyJfOfdA3l/yTF5yLi+WfuQ==}
engines: {node: '>= 10'}
cpu: [x64]
os: [linux]
'@next/swc-win32-arm64-msvc@16.2.5':
resolution: {integrity: sha512-f/H4nZ2zJBvA8/+HpsB9mNonF9zfQoAU6D0WxJrfzhJDvJLfngVN85oqxUyrDVK99DIFfFYhLpGa5K+c5uotSw==}
engines: {node: '>= 10'}
cpu: [arm64]
os: [win32]
'@next/swc-win32-x64-msvc@16.2.5':
resolution: {integrity: sha512-nuP7DHs4koAojsIxVPkihNgKiRUKtCU65j5X6DAbSy8VBrfT/o90bCLLHPf51JEdOZwZMFzM6e0NiGWfIWjVAg==}
engines: {node: '>= 10'}
cpu: [x64]
os: [win32]
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
'@npmcli/agent@4.0.2':
resolution: {integrity: sha512-EUEuWAxnL07Sp5/iC/1X6Xj+XThUvnbei9zfRWZdEXa7lss9RTHMhAHBeg+MZ5To9s/gGaSI+UwZTPdYMvKSeg==}
engines: {node: ^20.17.0 || >=22.9.0}
'@npmcli/arborist@9.9.0':
resolution: {integrity: sha512-tzsb1R8mx0k0drlLDT/9ckEYaQHIaWnUti/ci2IaFoQqdqwfmYrc+90p1+QEZC/ocvqcxBep9P1xKw4FbAGofw==}
engines: {node: ^20.17.0 || >=22.9.0}
hasBin: true
'@npmcli/fs@5.0.0':
resolution: {integrity: sha512-7OsC1gNORBEawOa5+j2pXN9vsicaIOH5cPXxoR6fJOmH6/EXpJB2CajXOu1fPRFun2m1lktEFX11+P89hqO/og==}
engines: {node: ^20.17.0 || >=22.9.0}
'@npmcli/git@7.0.2':
resolution: {integrity: sha512-oeolHDjExNAJAnlYP2qzNjMX/Xi9bmu78C9dIGr4xjobrSKbuMYCph8lTzn4vnW3NjIqVmw/f8BCfouqyJXlRg==}
engines: {node: ^20.17.0 || >=22.9.0}
'@npmcli/installed-package-contents@4.0.0':
resolution: {integrity: sha512-yNyAdkBxB72gtZ4GrwXCM0ZUedo9nIbOMKfGjt6Cu6DXf0p8y1PViZAKDC8q8kv/fufx0WTjRBdSlyrvnP7hmA==}
engines: {node: ^20.17.0 || >=22.9.0}
hasBin: true
'@npmcli/map-workspaces@5.0.3':
resolution: {integrity: sha512-o2grssXo1e774E5OtEwwrgoszYRh0lqkJH+Pb9r78UcqdGJRDRfhpM8DvZPjzNLLNYeD/rNbjOKM3Ss5UABROw==}
engines: {node: ^20.17.0 || >=22.9.0}
'@npmcli/metavuln-calculator@9.0.3':
resolution: {integrity: sha512-94GLSYhLXF2t2LAC7pDwLaM4uCARzxShyAQKsirmlNcpidH89VA4/+K1LbJmRMgz5gy65E/QBBWQdUvGLe2Frg==}
engines: {node: ^20.17.0 || >=22.9.0}
'@npmcli/name-from-folder@4.0.0':
resolution: {integrity: sha512-qfrhVlOSqmKM8i6rkNdZzABj8MKEITGFAY+4teqBziksCQAOLutiAxM1wY2BKEd8KjUSpWmWCYxvXr0y4VTlPg==}
engines: {node: ^20.17.0 || >=22.9.0}
'@npmcli/node-gyp@5.0.0':
resolution: {integrity: sha512-uuG5HZFXLfyFKqg8QypsmgLQW7smiRjVc45bqD/ofZZcR/uxEjgQU8qDPv0s9TEeMUiAAU/GC5bR6++UdTirIQ==}
engines: {node: ^20.17.0 || >=22.9.0}
'@npmcli/package-json@7.0.5':
resolution: {integrity: sha512-iVuTlG3ORq2iaVa1IWUxAO/jIp77tUKBhoMjuzYW2kL4MLN1bi/ofqkZ7D7OOwh8coAx1/S2ge0rMdGv8sLSOQ==}
engines: {node: ^20.17.0 || >=22.9.0}
'@npmcli/promise-spawn@9.0.1':
resolution: {integrity: sha512-OLUaoqBuyxeTqUvjA3FZFiXUfYC1alp3Sa99gW3EUDz3tZ3CbXDdcZ7qWKBzicrJleIgucoWamWH1saAmH/l2Q==}
engines: {node: ^20.17.0 || >=22.9.0}
'@npmcli/query@5.0.0':
resolution: {integrity: sha512-8TZWfTQOsODpLqo9SVhVjHovmKXNpevHU0gO9e+y4V4fRIOneiXy0u0sMP9LmS71XivrEWfZWg50ReH4WRT4aQ==}
engines: {node: ^20.17.0 || >=22.9.0}
'@npmcli/redact@4.0.0':
resolution: {integrity: sha512-gOBg5YHMfZy+TfHArfVogwgfBeQnKbbGo3pSUyK/gSI0AVu+pEiDVcKlQb0D8Mg1LNRZILZ6XG8I5dJ4KuAd9Q==}
engines: {node: ^20.17.0 || >=22.9.0}
'@npmcli/run-script@10.0.4':
resolution: {integrity: sha512-mGUWr1uMnf0le2TwfOZY4SFxZGXGfm4Jtay/nwAa2FLNAKXUoUwaGwBMNH36UHPtinWfTSJ3nqFQr0091CxVGg==}
engines: {node: ^20.17.0 || >=22.9.0}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
'@opencode-ai/plugin@1.18.15':
resolution: {integrity: sha512-AY10RtFbzLkf951dbLkSGJdBeddeS6ojbjvqCpWnINedqlj2cK/GF91xWjWnlHpqQZ4GABLPCuoSJx8mGmKvCw==}
peerDependencies:
'@opentui/core': '>=0.4.5'
'@opentui/keymap': '>=0.4.5'
'@opentui/solid': '>=0.4.5'
peerDependenciesMeta:
'@opentui/core':
optional: true
'@opentui/keymap':
optional: true
'@opentui/solid':
optional: true
'@opencode-ai/sdk@1.18.15':
resolution: {integrity: sha512-8sfo9nGiVwesAZW9Wqkvynyn7w4wYaHx1O9qOHpYL65+Bs2XUpHP3kBbZe58gjQydFgk6I74kUF7sqCiPu2arQ==}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
'@opentelemetry/api-logs@0.220.0':
resolution: {integrity: sha512-CmVa4ImJ+ynfrPMNaAXHET6Bhb44SwzmfyVJFq9ni2jgXJR/l7C6gfVFddNmHP+ZOkP9cf4f9DBe68qVLTHc9w==}
engines: {node: '>=8.0.0'}
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
'@opentelemetry/api@1.9.0':
resolution: {integrity: sha512-3giAOQvZiH5F9bMlMiv8+GSPMeqg0dbaeo58/0SlA9sxSqZhnUtxzX9/2FzyhS9sWQf5S0GJE0AKBrFqjpeYcg==}
engines: {node: '>=8.0.0'}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
'@opentelemetry/api@1.9.1':
resolution: {integrity: sha512-gLyJlPHPZYdAk1JENA9LeHejZe1Ti77/pTeFm/nMXmQH/HFZlcS/O2XJB+L8fkbrNSqhdtlvjBVjxwUYanNH5Q==}
engines: {node: '>=8.0.0'}
'@opentelemetry/context-async-hooks@2.9.0':
resolution: {integrity: sha512-OQ0vzvbZBiUhjqLnUaoNfYmP8553Crr3aggB4y0ZUi815mZ7idpdJXQmoKdeBKJelYttoBlLSSHubmyw3wvX4w==}
engines: {node: ^18.19.0 || >=20.6.0}
peerDependencies:
'@opentelemetry/api': '>=1.0.0 <1.10.0'
'@opentelemetry/core@2.9.0':
resolution: {integrity: sha512-m2nckMT80NnmjTYSPjJQObBJ+8dgkoajEOUbznL8AHZ3T3yHRk2P7gI1PhEBc1+lOnrYE9UWrWHqJDsmqjmNbw==}
engines: {node: ^18.19.0 || >=20.6.0}
peerDependencies:
'@opentelemetry/api': '>=1.0.0 <1.10.0'
'@opentelemetry/exporter-trace-otlp-grpc@0.220.0':
resolution: {integrity: sha512-bv1xmNhmNwIM6MdUBw4yYuJeVcEViVLk3uD69vOQMwueHBnfyl/u0HnBlB1FNY/Te0UOzJzvcbyR8wN6b+iGbA==}
engines: {node: ^18.19.0 || >=20.6.0}
peerDependencies:
'@opentelemetry/api': ^1.3.0
'@opentelemetry/exporter-zipkin@2.9.0':
resolution: {integrity: sha512-RwINoce2BH8T4obT5pMcAla2sWma1YZvYuaktWmTluQ0PkQdvv5D060rWI1+kawX+J2qBRcMbwrZJJNcMJUauQ==}
engines: {node: ^18.19.0 || >=20.6.0}
peerDependencies:
'@opentelemetry/api': ^1.0.0
'@opentelemetry/instrumentation-grpc@0.220.0':
resolution: {integrity: sha512-U1EF8KKu52XwH2ybUkjVDmaVQZGf3mXirRSw1KJQrOV5aymgJgkPJV7+kRPqawZe0rpVc/BK+pPSyMWuQoyJJQ==}
engines: {node: ^18.19.0 || >=20.6.0}
peerDependencies:
'@opentelemetry/api': ^1.3.0
'@opentelemetry/instrumentation@0.220.0':
resolution: {integrity: sha512-xQx3E2WxP1mDvKzxLxX+CTCtNLa560YJZ3087qYHerl2YmiKpv7AH+dAy7vmx+eVrZ5BwhfWUAVoKOoxCNHcpw==}
engines: {node: ^18.19.0 || >=20.6.0}
peerDependencies:
'@opentelemetry/api': ^1.3.0
'@opentelemetry/otlp-exporter-base@0.220.0':
resolution: {integrity: sha512-CXYo8UD5Mn9YbgebO2EL4wejtA+gxLmLiu6HCk2KH2BR7XhFN6/6p1UlCb23DYCjeYkndevLHuejCCN1yx4+OQ==}
engines: {node: ^18.19.0 || >=20.6.0}
peerDependencies:
'@opentelemetry/api': ^1.3.0
'@opentelemetry/otlp-grpc-exporter-base@0.220.0':
resolution: {integrity: sha512-/eIkBPMBTIvM3x/0mDX4aJeSkYifYClnBPr68PL1h5LV4VQv4+SV6CGrpiZ4fIWDnobVmhTWCm1J/QRdAWUfvA==}
engines: {node: ^18.19.0 || >=20.6.0}
peerDependencies:
'@opentelemetry/api': ^1.3.0
'@opentelemetry/otlp-transformer@0.220.0':
resolution: {integrity: sha512-lXGrv7KXZ0gNH9SVNUaa6vv6phVYGvJxfXAlMbzbakiXru75f5MZl8Z7oqiMMQD77riVHJCFlQvbZs/VVN2/4A==}
engines: {node: ^18.19.0 || >=20.6.0}
peerDependencies:
'@opentelemetry/api': ^1.3.0
'@opentelemetry/resources@2.9.0':
resolution: {integrity: sha512-jyA5MBLQ+Dkl3+JsZkUoUvL7yHvU64kLsvpXKarWm6347Sl1t1bXFTFykUePNpT5WH5pm9a2Qtt03iIYQhZ1Fg==}
engines: {node: ^18.19.0 || >=20.6.0}
peerDependencies:
'@opentelemetry/api': '>=1.3.0 <1.10.0'
'@opentelemetry/sdk-logs@0.220.0':
resolution: {integrity: sha512-WywcTkQtv2iNmt+6y5Kcd4rzvx9bLVsBa2Nwcmg01IUaBTkTow3W4d9KE5vNBpEDtb9tp21WcRBY/lANRrApYA==}
engines: {node: ^18.19.0 || >=20.6.0}
peerDependencies:
'@opentelemetry/api': '>=1.4.0 <1.10.0'
'@opentelemetry/sdk-metrics@2.9.0':
resolution: {integrity: sha512-Xx8RGS4H5XEBl01WuCreMIpiah9cCXMbSkeuIePPdD2cUpq/vUzYmj8E/MK1OsbOc93FuAD4jfn2WOacKwLn7Q==}
engines: {node: ^18.19.0 || >=20.6.0}
peerDependencies:
'@opentelemetry/api': '>=1.9.0 <1.10.0'
'@opentelemetry/sdk-trace-base@2.9.0':
resolution: {integrity: sha512-cp9zmTl62R8PJrpvFcmc8N2JQU/xfa0S+61q511Nji+QxCfZ8Ifvg7H27G8cANe4crg4RTrWsVvanHiXjSp6ag==}
engines: {node: ^18.19.0 || >=20.6.0}
peerDependencies:
'@opentelemetry/api': '>=1.3.0 <1.10.0'
'@opentelemetry/sdk-trace-node@2.9.0':
resolution: {integrity: sha512-ec9a7ps37huy5itYk0MalaZdSLlM6AXWp/FhtEjgMpp5leEGojBDvAl/UWttQnkMZOvFHKzRESn8TD3yKTF5nQ==}
engines: {node: ^18.19.0 || >=20.6.0}
peerDependencies:
'@opentelemetry/api': '>=1.0.0 <1.10.0'
'@opentelemetry/sdk-trace@2.9.0':
resolution: {integrity: sha512-sGA19HvtrrSKYsseHphluH6j3p6Xa3fqc7c7y8f/7mYWejc1lyDFcpSdD1kYa50HCLUeEo4zA5bW0pniaPszuw==}
engines: {node: ^18.19.0 || >=20.6.0}
peerDependencies:
'@opentelemetry/api': '>=1.3.0 <1.10.0'
'@opentelemetry/semantic-conventions@1.43.0':
resolution: {integrity: sha512-eSYWTm620tTk45EKSedaUL8MFYI8hW164hIXsgIHyxu3VobUB3fFCu5t0hQby6OoWRPsG1KkKUG2M5UadiLiVg==}
engines: {node: '>=14'}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
'@opentui/core-darwin-arm64@0.4.5':
resolution: {integrity: sha512-8KUG0oRidnR+oW1RSZJ72/PhZLl+qRRMk5U/mieF4c0SJ5V3tYACpBZAKzQfHNd1f7QzD8FHZct1lPpQgtmkWg==}
cpu: [arm64]
os: [darwin]
'@opentui/core-darwin-x64@0.4.5':
resolution: {integrity: sha512-R2bocsg55gwjOqCp/MWFgFYzRmsduKegB6nzgFAPCvAD/L5Jf30xpWJWFlSg3x8vxe1L9WJ84dfqa4M7mZZ3wA==}
cpu: [x64]
os: [darwin]
'@opentui/core-linux-arm64-musl@0.4.5':
resolution: {integrity: sha512-ieqdyKI6EIYPalYAETB2wsdP83hr5Ifi+dFnBFUmdEEFHsoKwBmn2S7bsTOYlX7Bg03F4/YPIg+IvRpeC+cUJw==}
cpu: [arm64]
os: [linux]
'@opentui/core-linux-arm64@0.4.5':
resolution: {integrity: sha512-R4MZ25a4CzOAGVjW9aj1hUfzQGVfCJwrwBDbNs2SXaIvzcZqkxCVtU4FoQ5LsaD0j/BdNQVg2CIfFkFsm1fDuQ==}
cpu: [arm64]
os: [linux]
'@opentui/core-linux-x64-musl@0.4.5':
resolution: {integrity: sha512-mKVKcIcPiSVVZZsdPSBoWwoa2/TCeQAaMDeHF7PFw2kt5bTXZPP7xxWfRQLCNIcA1eaGl59UuwUWHDR2Ve548Q==}
cpu: [x64]
os: [linux]
'@opentui/core-linux-x64@0.4.5':
resolution: {integrity: sha512-SNyuQoxMKI1vuJhgxSSW96adWM6LqFl2SoS3GM4tGeneGOanVVG2Y06PvlytXvF4cKik97t0rqkVMRetmOs93w==}
cpu: [x64]
os: [linux]
'@opentui/core-win32-arm64@0.4.5':
resolution: {integrity: sha512-GHTTsqeR45q2Iek9Rb7ty+x/hAKn2jZ1ujlCgPR8LBKyF7h0E1dNFryoZ7ehMc3kJndP1sKn836IemKFqxuDdQ==}
cpu: [arm64]
os: [win32]
'@opentui/core-win32-x64@0.4.5':
resolution: {integrity: sha512-Y8T/yXCDGagRGiQrtmuB6AhRcPucKFs/Dre3v8kJwNYqDccI4FzUPKclZ7djfmRZNjl7JUqPhZZP/PwDpQocMg==}
cpu: [x64]
os: [win32]
'@opentui/core@0.4.5':
resolution: {integrity: sha512-JsgRTPkA6e+Vxmumxai6SElOSlRQkbzNKHlCfemlArRiLhfC1IZ9RXJo2QH4xSu+uBOWAM90uss73/pPlkdEig==}
peerDependencies:
web-tree-sitter: 0.25.10
'@opentui/keymap@0.4.5':
resolution: {integrity: sha512-S1wzKHhF70zT6bH+VBFY+lSeTImLcIFW28JNQiME8MoPcy6KGPs7rKFSHrb/U7P8rsTJeRfW5A4d1Cy6PKodDg==}
peerDependencies:
'@opentui/react': 0.4.5
'@opentui/solid': 0.4.5
react: '>=19.2.0'
solid-js: 1.9.12
peerDependenciesMeta:
'@opentui/react':
optional: true
'@opentui/solid':
optional: true
react:
optional: true
solid-js:
optional: true
'@opentui/solid@0.4.5':
resolution: {integrity: sha512-B0RSkXnrtPVfEJOX+Hj+axjLJ3lzbG1BZw5I7Pvb9OPp48Vzg2cW2a3cSa86/q48ndLt647i/XwFPIw/jqnI5g==}
peerDependencies:
solid-js: 1.9.12
first commit
2026-02-21 03:10:39 +00:00
'@pinojs/redact@0.4.0':
resolution: {integrity: sha512-k2ENnmBugE/rzQfEcdWHcCY+/FM3VLzH9cYEsbdsoqrvzAKRhUZeRNhAZvB8OitQJ1TBed3yqWtdjzS6wJKBwg==}
'@prisma/client@6.19.2':
resolution: {integrity: sha512-gR2EMvfK/aTxsuooaDA32D8v+us/8AAet+C3J1cc04SW35FPdZYgLF+iN4NDLUgAaUGTKdAB0CYenu1TAgGdMg==}
engines: {node: '>=18.18'}
peerDependencies:
prisma: '*'
typescript: '>=5.1.0'
peerDependenciesMeta:
prisma:
optional: true
typescript:
optional: true
'@prisma/config@6.19.2':
resolution: {integrity: sha512-kadBGDl+aUswv/zZMk9Mx0C8UZs1kjao8H9/JpI4Wh4SHZaM7zkTwiKn/iFLfRg+XtOAo/Z/c6pAYhijKl0nzQ==}
'@prisma/debug@6.19.2':
resolution: {integrity: sha512-lFnEZsLdFLmEVCVNdskLDCL8Uup41GDfU0LUfquw+ercJC8ODTuL0WNKgOKmYxCJVvFwf0OuZBzW99DuWmoH2A==}
'@prisma/engines-version@7.1.1-3.c2990dca591cba766e3b7ef5d9e8a84796e47ab7':
resolution: {integrity: sha512-03bgb1VD5gvuumNf+7fVGBzfpJPjmqV423l/WxsWk2cNQ42JD0/SsFBPhN6z8iAvdHs07/7ei77SKu7aZfq8bA==}
'@prisma/engines@6.19.2':
resolution: {integrity: sha512-TTkJ8r+uk/uqczX40wb+ODG0E0icVsMgwCTyTHXehaEfb0uo80M9g1aW1tEJrxmFHeOZFXdI2sTA1j1AgcHi4A==}
'@prisma/fetch-engine@6.19.2':
resolution: {integrity: sha512-h4Ff4Pho+SR1S8XerMCC12X//oY2bG3Iug/fUnudfcXEUnIeRiBdXHFdGlGOgQ3HqKgosTEhkZMvGM9tWtYC+Q==}
'@prisma/get-platform@6.19.2':
resolution: {integrity: sha512-PGLr06JUSTqIvztJtAzIxOwtWKtJm5WwOG6xpsgD37Rc84FpfUBGLKz65YpJBGtkRQGXTYEFie7pYALocC3MtA==}
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
'@protobufjs/aspromise@1.1.2':
resolution: {integrity: sha512-j+gKExEuLmKwvz3OgROXtrJ2UG2x8Ch2YZUxahh+s1F2HZ+wAceUNLkvy6zKCPVRkU++ZWQrdxsUeQXmcg4uoQ==}
'@protobufjs/base64@1.1.2':
resolution: {integrity: sha512-AZkcAA5vnN/v4PDqKyMR5lx7hZttPDgClv83E//FMNhR2TMcLUhfRUBHCmSl0oi9zMgDDqRUJkSxO3wm85+XLg==}
'@protobufjs/codegen@2.0.4':
resolution: {integrity: sha512-YyFaikqM5sH0ziFZCN3xDC7zeGaB/d0IUb9CATugHWbd1FRFwWwt4ld4OYMPWu5a3Xe01mGAULCdqhMlPl29Jg==}
'@protobufjs/eventemitter@1.1.0':
resolution: {integrity: sha512-j9ednRT81vYJ9OfVuXG6ERSTdEL1xVsNgqpkxMsbIabzSo3goCjDIveeGv5d03om39ML71RdmrGNjG5SReBP/Q==}
'@protobufjs/fetch@1.1.0':
resolution: {integrity: sha512-lljVXpqXebpsijW71PZaCYeIcE5on1w5DlQy5WH6GLbFryLUrBD4932W/E2BSpfRJWseIL4v/KPgBFxDOIdKpQ==}
'@protobufjs/float@1.0.2':
resolution: {integrity: sha512-Ddb+kVXlXst9d+R9PfTIxh1EdNkgoRe5tOX6t01f1lYWOvJnSPDBlG241QLzcyPdoNTsblLUdujGSE4RzrTZGQ==}
'@protobufjs/inquire@1.1.0':
resolution: {integrity: sha512-kdSefcPdruJiFMVSbn801t4vFK7KB/5gd2fYvrxhuJYg8ILrmn9SKSX2tZdV6V+ksulWqS7aXjBcRXl3wHoD9Q==}
'@protobufjs/path@1.1.2':
resolution: {integrity: sha512-6JOcJ5Tm08dOHAbdR3GrvP+yUUfkjG5ePsHYczMFLq3ZmMkAD98cDgcT2iA1lJ9NVwFd4tH/iSSoe44YWkltEA==}
'@protobufjs/pool@1.1.0':
resolution: {integrity: sha512-0kELaGSIDBKvcgS4zkjz1PeddatrjYcmMWOlAuAPwAeccUrPHdUqo/J6LiymHHEiJT5NrF1UVwxY14f+fy4WQw==}
'@protobufjs/utf8@1.1.0':
resolution: {integrity: sha512-Vvn3zZrhQZkkBE8LSuW3em98c0FwgO4nxzv6OdSxPKJIEKY2bGbHn+mhGIPerzI4twdxaP8/0+06HBpwf345Lw==}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
'@pulumi/pulumi@3.253.0':
resolution: {integrity: sha512-riKRPh6hNCA/NRxVSOLx08c0rqgWYnKGOTgPZzCXi5TglVEERiXHW4U2DaTpP978FI1Xo1iqCa41vmd5i/PNHw==}
engines: {node: '>=22'}
peerDependencies:
ts-node: '>= 7.0.1 < 12'
typescript: '>= 3.8.3 < 7'
peerDependenciesMeta:
ts-node:
optional: true
typescript:
optional: true
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
'@rolldown/pluginutils@1.0.0-rc.3':
resolution: {integrity: sha512-eybk3TjzzzV97Dlj5c+XrBFW57eTNhzod66y9HrBlzJ6NsCrWCp/2kaPS3K9wJmurBC0Tdw4yPjXKZqlznim3Q==}
first commit
2026-02-21 03:10:39 +00:00
'@rollup/rollup-android-arm-eabi@4.58.0':
resolution: {integrity: sha512-mr0tmS/4FoVk1cnaeN244A/wjvGDNItZKR8hRhnmCzygyRXYtKF5jVDSIILR1U97CTzAYmbgIj/Dukg62ggG5w==}
cpu: [arm]
os: [android]
'@rollup/rollup-android-arm64@4.58.0':
resolution: {integrity: sha512-+s++dbp+/RTte62mQD9wLSbiMTV+xr/PeRJEc/sFZFSBRlHPNPVaf5FXlzAL77Mr8FtSfQqCN+I598M8U41ccQ==}
cpu: [arm64]
os: [android]
'@rollup/rollup-darwin-arm64@4.58.0':
resolution: {integrity: sha512-MFWBwTcYs0jZbINQBXHfSrpSQJq3IUOakcKPzfeSznONop14Pxuqa0Kg19GD0rNBMPQI2tFtu3UzapZpH0Uc1Q==}
cpu: [arm64]
os: [darwin]
'@rollup/rollup-darwin-x64@4.58.0':
resolution: {integrity: sha512-yiKJY7pj9c9JwzuKYLFaDZw5gma3fI9bkPEIyofvVfsPqjCWPglSHdpdwXpKGvDeYDms3Qal8qGMEHZ1M/4Udg==}
cpu: [x64]
os: [darwin]
'@rollup/rollup-freebsd-arm64@4.58.0':
resolution: {integrity: sha512-x97kCoBh5MOevpn/CNK9W1x8BEzO238541BGWBc315uOlN0AD/ifZ1msg+ZQB05Ux+VF6EcYqpiagfLJ8U3LvQ==}
cpu: [arm64]
os: [freebsd]
'@rollup/rollup-freebsd-x64@4.58.0':
resolution: {integrity: sha512-Aa8jPoZ6IQAG2eIrcXPpjRcMjROMFxCt1UYPZZtCxRV68WkuSigYtQ/7Zwrcr2IvtNJo7T2JfDXyMLxq5L4Jlg==}
cpu: [x64]
os: [freebsd]
'@rollup/rollup-linux-arm-gnueabihf@4.58.0':
resolution: {integrity: sha512-Ob8YgT5kD/lSIYW2Rcngs5kNB/44Q2RzBSPz9brf2WEtcGR7/f/E9HeHn1wYaAwKBni+bdXEwgHvUd0x12lQSA==}
cpu: [arm]
os: [linux]
'@rollup/rollup-linux-arm-musleabihf@4.58.0':
resolution: {integrity: sha512-K+RI5oP1ceqoadvNt1FecL17Qtw/n9BgRSzxif3rTL2QlIu88ccvY+Y9nnHe/cmT5zbH9+bpiJuG1mGHRVwF4Q==}
cpu: [arm]
os: [linux]
'@rollup/rollup-linux-arm64-gnu@4.58.0':
resolution: {integrity: sha512-T+17JAsCKUjmbopcKepJjHWHXSjeW7O5PL7lEFaeQmiVyw4kkc5/lyYKzrv6ElWRX/MrEWfPiJWqbTvfIvjM1Q==}
cpu: [arm64]
os: [linux]
'@rollup/rollup-linux-arm64-musl@4.58.0':
resolution: {integrity: sha512-cCePktb9+6R9itIJdeCFF9txPU7pQeEHB5AbHu/MKsfH/k70ZtOeq1k4YAtBv9Z7mmKI5/wOLYjQ+B9QdxR6LA==}
cpu: [arm64]
os: [linux]
'@rollup/rollup-linux-loong64-gnu@4.58.0':
resolution: {integrity: sha512-iekUaLkfliAsDl4/xSdoCJ1gnnIXvoNz85C8U8+ZxknM5pBStfZjeXgB8lXobDQvvPRCN8FPmmuTtH+z95HTmg==}
cpu: [loong64]
os: [linux]
'@rollup/rollup-linux-loong64-musl@4.58.0':
resolution: {integrity: sha512-68ofRgJNl/jYJbxFjCKE7IwhbfxOl1muPN4KbIqAIe32lm22KmU7E8OPvyy68HTNkI2iV/c8y2kSPSm2mW/Q9Q==}
cpu: [loong64]
os: [linux]
'@rollup/rollup-linux-ppc64-gnu@4.58.0':
resolution: {integrity: sha512-dpz8vT0i+JqUKuSNPCP5SYyIV2Lh0sNL1+FhM7eLC457d5B9/BC3kDPp5BBftMmTNsBarcPcoz5UGSsnCiw4XQ==}
cpu: [ppc64]
os: [linux]
'@rollup/rollup-linux-ppc64-musl@4.58.0':
resolution: {integrity: sha512-4gdkkf9UJ7tafnweBCR/mk4jf3Jfl0cKX9Np80t5i78kjIH0ZdezUv/JDI2VtruE5lunfACqftJ8dIMGN4oHew==}
cpu: [ppc64]
os: [linux]
'@rollup/rollup-linux-riscv64-gnu@4.58.0':
resolution: {integrity: sha512-YFS4vPnOkDTD/JriUeeZurFYoJhPf9GQQEF/v4lltp3mVcBmnsAdjEWhr2cjUCZzZNzxCG0HZOvJU44UGHSdzw==}
cpu: [riscv64]
os: [linux]
'@rollup/rollup-linux-riscv64-musl@4.58.0':
resolution: {integrity: sha512-x2xgZlFne+QVNKV8b4wwaCS8pwq3y14zedZ5DqLzjdRITvreBk//4Knbcvm7+lWmms9V9qFp60MtUd0/t/PXPw==}
cpu: [riscv64]
os: [linux]
'@rollup/rollup-linux-s390x-gnu@4.58.0':
resolution: {integrity: sha512-jIhrujyn4UnWF8S+DHSkAkDEO3hLX0cjzxJZPLF80xFyzyUIYgSMRcYQ3+uqEoyDD2beGq7Dj7edi8OnJcS/hg==}
cpu: [s390x]
os: [linux]
'@rollup/rollup-linux-x64-gnu@4.58.0':
resolution: {integrity: sha512-+410Srdoh78MKSJxTQ+hZ/Mx+ajd6RjjPwBPNd0R3J9FtL6ZA0GqiiyNjCO9In0IzZkCNrpGymSfn+kgyPQocg==}
cpu: [x64]
os: [linux]
'@rollup/rollup-linux-x64-musl@4.58.0':
resolution: {integrity: sha512-ZjMyby5SICi227y1MTR3VYBpFTdZs823Rs/hpakufleBoufoOIB6jtm9FEoxn/cgO7l6PM2rCEl5Kre5vX0QrQ==}
cpu: [x64]
os: [linux]
'@rollup/rollup-openbsd-x64@4.58.0':
resolution: {integrity: sha512-ds4iwfYkSQ0k1nb8LTcyXw//ToHOnNTJtceySpL3fa7tc/AsE+UpUFphW126A6fKBGJD5dhRvg8zw1rvoGFxmw==}
cpu: [x64]
os: [openbsd]
'@rollup/rollup-openharmony-arm64@4.58.0':
resolution: {integrity: sha512-fd/zpJniln4ICdPkjWFhZYeY/bpnaN9pGa6ko+5WD38I0tTqk9lXMgXZg09MNdhpARngmxiCg0B0XUamNw/5BQ==}
cpu: [arm64]
os: [openharmony]
'@rollup/rollup-win32-arm64-msvc@4.58.0':
resolution: {integrity: sha512-YpG8dUOip7DCz3nr/JUfPbIUo+2d/dy++5bFzgi4ugOGBIox+qMbbqt/JoORwvI/C9Kn2tz6+Bieoqd5+B1CjA==}
cpu: [arm64]
os: [win32]
'@rollup/rollup-win32-ia32-msvc@4.58.0':
resolution: {integrity: sha512-b9DI8jpFQVh4hIXFr0/+N/TzLdpBIoPzjt0Rt4xJbW3mzguV3mduR9cNgiuFcuL/TeORejJhCWiAXe3E/6PxWA==}
cpu: [ia32]
os: [win32]
'@rollup/rollup-win32-x64-gnu@4.58.0':
resolution: {integrity: sha512-CSrVpmoRJFN06LL9xhkitkwUcTZtIotYAF5p6XOR2zW0Zz5mzb3IPpcoPhB02frzMHFNo1reQ9xSF5fFm3hUsQ==}
cpu: [x64]
os: [win32]
'@rollup/rollup-win32-x64-msvc@4.58.0':
resolution: {integrity: sha512-QFsBgQNTnh5K0t/sBsjJLq24YVqEIVkGpfN2VHsnN90soZyhaiA9UUHufcctVNL4ypJY0wrwad0wslx2KJQ1/w==}
cpu: [x64]
os: [win32]
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
'@sigstore/bundle@4.0.0':
resolution: {integrity: sha512-NwCl5Y0V6Di0NexvkTqdoVfmjTaQwoLM236r89KEojGmq/jMls8S+zb7yOwAPdXvbwfKDlP+lmXgAL4vKSQT+A==}
engines: {node: ^20.17.0 || >=22.9.0}
'@sigstore/core@3.2.1':
resolution: {integrity: sha512-qRsxPnCrbC/puegGxKuynfnxgLiHqWStrSjxkoB4YKqq3Z3s4cyZyj42ZdWFAEblNP65C+rBH8EuREHIXoi83g==}
engines: {node: ^20.17.0 || >=22.9.0}
'@sigstore/protobuf-specs@0.5.1':
resolution: {integrity: sha512-/ScWUhhoFasJsSRGTVBwId1loQjjnjAfE4djL6ZhrXRpNCmPTnUKF5Jokd58ILseOMjzET3UrMOtJPS9sYeI0g==}
engines: {node: ^18.17.0 || >=20.5.0}
'@sigstore/sign@4.1.1':
resolution: {integrity: sha512-Hf4xglukg0XXQ2RiD5vSoLjdPe8OBUPA8XeVjUObheuDcWdYWrnH/BNmxZCzkAy68MzmNCxXLeurJvs6hcP2OQ==}
engines: {node: ^20.17.0 || >=22.9.0}
'@sigstore/tuf@4.0.2':
resolution: {integrity: sha512-TCAzTy0xzdP79EnxSjq9KQ3eaR7+FmudLC6eRKknVKZbV7ZNlGLClAAQb/HMNJ5n2OBNk2GT1tEmU0xuPr+SLQ==}
engines: {node: ^20.17.0 || >=22.9.0}
'@sigstore/verify@3.1.1':
resolution: {integrity: sha512-qv7+G3J2cc6wwFj3yKvXOamzqhMwSk1ogPGmhpS8iXllcPrJaIIBA+4HbttlHVu1pqWTdmaCH/WE7UOC51kdoA==}
engines: {node: ^20.17.0 || >=22.9.0}
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
'@silvia-odwyer/photon-node@0.3.4':
resolution: {integrity: sha512-bnly4BKB3KDTFxrUIcgCLbaeVVS8lrAkri1pEzskpmxu9MdfGQTy8b8EgcD83ywD3RPMsIulY8xJH5Awa+t9fA==}
'@smithy/core@3.31.1':
resolution: {integrity: sha512-CyogUINxvi7C7LDsh8Syo6hVJOT9ckz4rG8dRZfTJ8r91HkMY59PnNooaj7WcHyxEkxPfBAmbgztZU+xTo76lg==}
engines: {node: '>=18.0.0'}
'@smithy/credential-provider-imds@4.4.16':
resolution: {integrity: sha512-QfuLWAkLzptffFW980AFeHZFdqds2B64rpEd3uJ6lgs3xVn9QegGMUgUcj+4d7dRrAsya3r58ZKpku97WcFb4w==}
engines: {node: '>=18.0.0'}
'@smithy/fetch-http-handler@5.6.13':
resolution: {integrity: sha512-4fW86pEUOMbrD5nkbyl/tTvPHHWJFbuB2odl6ps9lWfHoXf9HWh3Q/Smh59qH1g7+c/BSZghX6bbUk4gsiMs8A==}
engines: {node: '>=18.0.0'}
'@smithy/is-array-buffer@2.2.0':
resolution: {integrity: sha512-GGP3O9QFD24uGeAXYUjwSTXARoqpZykHadOmA8G5vfJPK0/DC67qa//0qvqrJzL1xc8WQWX7/yc7fwudjPHPhA==}
engines: {node: '>=14.0.0'}
'@smithy/node-http-handler@4.7.3':
resolution: {integrity: sha512-/jPhevcTFPMVl6KNjbaI47iOg1zxC7IsnX4PQDGVZKMFceOXtB8IEYaB7a9VvkP/3oC60WzTeKocvSI7vLT0vA==}
engines: {node: '>=18.0.0'}
'@smithy/node-http-handler@4.9.13':
resolution: {integrity: sha512-Nmd/Nl35zfYrd+a6OO2cDJb3GPh9bgTjIUhcM+JFfjpp8/osCgboDV5nCT1I01Pv6R13eSKDKLSoVa5ZB6Zsfw==}
engines: {node: '>=18.0.0'}
'@smithy/signature-v4@5.6.12':
resolution: {integrity: sha512-I6KLtq3H0qqSuV9vLglfi8puHqzygzWHOnI4z/Rdoo+q50vvo18vBRdPAvvEtcaKROz7Zn6qnPa14kRfPH6PcQ==}
engines: {node: '>=18.0.0'}
'@smithy/types@4.16.1':
resolution: {integrity: sha512-0JFs3V2y2M9tKW5na/qxe69Zv+uxLMO7QBbhxF/FHu/Gp2NFZAAL9tWl9PU02xxo07pb3G9FTyjNc6D5uZrJIg==}
engines: {node: '>=18.0.0'}
'@smithy/util-buffer-from@2.2.0':
resolution: {integrity: sha512-IJdWBbTcMQ6DA0gdNhh/BwrLkDR+ADW5Kr1aZmd4k3DIF6ezMV4R2NIAmT08wQJ3yUK82thHWmC/TnK/wpMMIA==}
engines: {node: '>=14.0.0'}
'@smithy/util-utf8@2.3.0':
resolution: {integrity: sha512-R8Rdn8Hy72KKcebgLiv8jQcQkXoLMOGGv5uI1/k0l+snqkOzQ1R0ChUBCxWMlBsFMekWjq0wRudIweFs7sKT5A==}
engines: {node: '>=14.0.0'}
first commit
2026-02-21 03:10:39 +00:00
'@standard-schema/spec@1.1.0':
resolution: {integrity: sha512-l2aFy5jALhniG5HgqrD6jXLi/rUWrKvqN/qJx6yoJsgKhblVd+iqqU4RCXavm/jPityDo5TCvKMnpjKnOriy0w==}
feat(web): bold redesign — Tailwind v4 + shadcn-style primitives + Skills/Proposals/Revisions UI Phase 6 of the Skills + Revisions + Proposals work. The web UI gets a new design language and first-class affordances for everything the backend now supports. ## Visual direction - Tailwind v4 with custom @theme block (oklch tokens). Dark-mode-only (internal tool — light mode doubles QA surface). - Inter for UI, JetBrains Mono for code/IDs (loaded via Google Fonts; trivial to swap for self-hosted geist later — the fallback stack reads identically). - Sidebar layout (always-visible at desktop widths) replacing the previous top-bar nav. Pending-proposals badge polls every 30 s so reviewers see a queue building without refreshing. - Lucide icons throughout. - Spacing and radii on Tailwind defaults. Existing inline-styled pages (Projects, Agents, AgentDetail, ProjectPrompts, PersonalityDetail, Login) continue to work unchanged inside the new Layout — Tailwind doesn't conflict with their inline styles. A follow-up can migrate them incrementally. ## What's added ### Build infra (src/web/) - package.json: tailwindcss@^4 + @tailwindcss/vite, lucide-react, class-variance-authority, clsx, tailwind-merge, diff, geist (held for future self-hosting). - vite.config.ts: registers the @tailwindcss/vite plugin. - src/index.css: Tailwind import + @theme tokens + @layer base. - src/main.tsx: imports index.css. - src/lib/utils.ts: shadcn-style cn() helper. ### shadcn-style primitives (src/components/ui/) Hand-written rather than generated via `npx shadcn` so the repo doesn't depend on a CLI tool that needs interactive runtime: - button.tsx — variants: primary / secondary / ghost / danger / link; sizes: sm / md / lg / icon. - card.tsx — Card + Header/Title/Description/Content/Footer subparts. - badge.tsx — variants: default / info / success / warning / danger / outline. - input.tsx — Input + Textarea + Label. - tabs.tsx — no-dep accessible Tabs (no Radix needed for our use). - separator.tsx — h/v separator with role=separator. ### Diff component (src/components/Diff.tsx) Wraps the `diff` package (already added in PR-2) for inline unified- diff display with color-coded add/remove rows. Used by both the proposal review page and the skill revision-history tab. ### New pages (src/pages/) - Dashboard.tsx — at-a-glance home. Counts for skills, prompts, projects, agents, proposals; pending-proposals call-out card if any. - Skills.tsx — list view, separated into Global vs Project/Agent- scoped sections. - SkillDetail.tsx — name + semver + description; tabs for SKILL.md / Files / Metadata / History. History tab shows revisions with click-to-diff against the live body. - Proposals.tsx — queue with Pending/Approved/Rejected tabs. Pending count is highlighted in amber. - ProposalDetail.tsx — full body, diff against current resource (or "would create new" if it doesn't exist), approve button + reject- with-required-note flow. ### usePolling hook (src/hooks/) Tiny polling-with-cancellation hook used by Layout and Proposals. ### Layout rewrite (src/components/Layout.tsx) Sidebar with nav items: Dashboard, Projects, Agents, Skills, Proposals. Lucide icons. Active-route highlighting via NavLink. Pending-proposals warning badge on the Proposals item. ### Routes (src/App.tsx) New routes: /dashboard, /skills, /skills/:name, /proposals, /proposals/:id. Default redirects to /dashboard. ### API types (src/api.ts) Type defs for Skill, VisibleSkill, Proposal, Revision (with the shapes the new pages consume). ## Tests Existing 7 web tests still pass (Login + api). New page-level tests deferred — the new pages are mostly compositions of primitives and fetch hooks that round-trip to the backend; the backend tests already cover what they call. PR-7 polish can add render-and-click tests if coverage drift surfaces. ## Verification - `pnpm --filter @mcpctl/web build` clean, no warnings. - `pnpm test:run` whole monorepo: 162 test files / 2157 tests green. - Visual smoke deferred — needs a running mcpd to populate the fixtures. Manual smoke tested locally is the next step. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-07 17:54:55 +01:00
'@swc/helpers@0.5.15':
resolution: {integrity: sha512-JQ5TuMi45Owi4/BIMAJBoSQoOJu12oOk/gADqlcUL9JEdHB8vyjUSsxqeNXnmXHjYKMi2WcYtezGEEhqUI/E2g==}
'@tailwindcss/node@4.2.4':
resolution: {integrity: sha512-Ai7+yQPxz3ddrDQzFfBKdHEVBg0w3Zl83jnjuwxnZOsnH9pGn93QHQtpU0p/8rYWxvbFZHneni6p1BSLK4DkGA==}
'@tailwindcss/oxide-android-arm64@4.2.4':
resolution: {integrity: sha512-e7MOr1SAn9U8KlZzPi1ZXGZHeC5anY36qjNwmZv9pOJ8E4Q6jmD1vyEHkQFmNOIN7twGPEMXRHmitN4zCMN03g==}
engines: {node: '>= 20'}
cpu: [arm64]
os: [android]
'@tailwindcss/oxide-darwin-arm64@4.2.4':
resolution: {integrity: sha512-tSC/Kbqpz/5/o/C2sG7QvOxAKqyd10bq+ypZNf+9Fi2TvbVbv1zNpcEptcsU7DPROaSbVgUXmrzKhurFvo5eDg==}
engines: {node: '>= 20'}
cpu: [arm64]
os: [darwin]
'@tailwindcss/oxide-darwin-x64@4.2.4':
resolution: {integrity: sha512-yPyUXn3yO/ufR6+Kzv0t4fCg2qNr90jxXc5QqBpjlPNd0NqyDXcmQb/6weunH/MEDXW5dhyEi+agTDiqa3WsGg==}
engines: {node: '>= 20'}
cpu: [x64]
os: [darwin]
'@tailwindcss/oxide-freebsd-x64@4.2.4':
resolution: {integrity: sha512-BoMIB4vMQtZsXdGLVc2z+P9DbETkiopogfWZKbWwM8b/1Vinbs4YcUwo+kM/KeLkX3Ygrf4/PsRndKaYhS8Eiw==}
engines: {node: '>= 20'}
cpu: [x64]
os: [freebsd]
'@tailwindcss/oxide-linux-arm-gnueabihf@4.2.4':
resolution: {integrity: sha512-7pIHBLTHYRAlS7V22JNuTh33yLH4VElwKtB3bwchK/UaKUPpQ0lPQiOWcbm4V3WP2I6fNIJ23vABIvoy2izdwA==}
engines: {node: '>= 20'}
cpu: [arm]
os: [linux]
'@tailwindcss/oxide-linux-arm64-gnu@4.2.4':
resolution: {integrity: sha512-+E4wxJ0ZGOzSH325reXTWB48l42i93kQqMvDyz5gqfRzRZ7faNhnmvlV4EPGJU3QJM/3Ab5jhJ5pCRUsKn6OQw==}
engines: {node: '>= 20'}
cpu: [arm64]
os: [linux]
'@tailwindcss/oxide-linux-arm64-musl@4.2.4':
resolution: {integrity: sha512-bBADEGAbo4ASnppIziaQJelekCxdMaxisrk+fB7Thit72IBnALp9K6ffA2G4ruj90G9XRS2VQ6q2bCKbfFV82g==}
engines: {node: '>= 20'}
cpu: [arm64]
os: [linux]
'@tailwindcss/oxide-linux-x64-gnu@4.2.4':
resolution: {integrity: sha512-7Mx25E4WTfnht0TVRTyC00j3i0M+EeFe7wguMDTlX4mRxafznw0CA8WJkFjWYH5BlgELd1kSjuU2JiPnNZbJDA==}
engines: {node: '>= 20'}
cpu: [x64]
os: [linux]
'@tailwindcss/oxide-linux-x64-musl@4.2.4':
resolution: {integrity: sha512-2wwJRF7nyhOR0hhHoChc04xngV3iS+akccHTGtz965FwF0up4b2lOdo6kI1EbDaEXKgvcrFBYcYQQ/rrnWFVfA==}
engines: {node: '>= 20'}
cpu: [x64]
os: [linux]
'@tailwindcss/oxide-wasm32-wasi@4.2.4':
resolution: {integrity: sha512-FQsqApeor8Fo6gUEklzmaa9994orJZZDBAlQpK2Mq+DslRKFJeD6AjHpBQ0kZFQohVr8o85PPh8eOy86VlSCmw==}
engines: {node: '>=14.0.0'}
cpu: [wasm32]
bundledDependencies:
- '@napi-rs/wasm-runtime'
- '@emnapi/core'
- '@emnapi/runtime'
- '@tybys/wasm-util'
- '@emnapi/wasi-threads'
- tslib
'@tailwindcss/oxide-win32-arm64-msvc@4.2.4':
resolution: {integrity: sha512-L9BXqxC4ToVgwMFqj3pmZRqyHEztulpUJzCxUtLjobMCzTPsGt1Fa9enKbOpY2iIyVtaHNeNvAK8ERP/64sqGQ==}
engines: {node: '>= 20'}
cpu: [arm64]
os: [win32]
'@tailwindcss/oxide-win32-x64-msvc@4.2.4':
resolution: {integrity: sha512-ESlKG0EpVJQwRjXDDa9rLvhEAh0mhP1sF7sap9dNZT0yyl9SAG6T7gdP09EH0vIv0UNTlo6jPWyujD6559fZvw==}
engines: {node: '>= 20'}
cpu: [x64]
os: [win32]
'@tailwindcss/oxide@4.2.4':
resolution: {integrity: sha512-9El/iI069DKDSXwTvB9J4BwdO5JhRrOweGaK25taBAvBXyXqJAX+Jqdvs8r8gKpsI/1m0LeJLyQYTf/WLrBT1Q==}
engines: {node: '>= 20'}
'@tailwindcss/vite@4.2.4':
resolution: {integrity: sha512-pCvohwOCspk3ZFn6eJzrrX3g4n2JY73H6MmYC87XfGPyTty4YsCjYTMArRZm/zOI8dIt3+EcrLHAFPe5A4bgtw==}
peerDependencies:
vite: ^5.2.0 || ^6 || ^7 || ^8
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
'@testing-library/dom@10.4.1':
resolution: {integrity: sha512-o4PXJQidqJl82ckFaXUeoAW+XysPLauYI43Abki5hABd853iMhitooc6znOnczgbTYmEP6U6/y1ZyKAIsvMKGg==}
engines: {node: '>=18'}
'@testing-library/jest-dom@6.9.1':
resolution: {integrity: sha512-zIcONa+hVtVSSep9UT3jZ5rizo2BsxgyDYU7WFD5eICBE7no3881HGeb/QkGfsJs6JTkY1aQhT7rIPC7e+0nnA==}
engines: {node: '>=14', npm: '>=6', yarn: '>=1'}
'@testing-library/react@16.3.2':
resolution: {integrity: sha512-XU5/SytQM+ykqMnAnvB2umaJNIOsLF3PVv//1Ew4CTcpz0/BRyy/af40qqrt7SjKpDdT1saBMc42CUok5gaw+g==}
engines: {node: '>=18'}
peerDependencies:
'@testing-library/dom': ^10.0.0
'@types/react': ^18.0.0 || ^19.0.0
'@types/react-dom': ^18.0.0 || ^19.0.0
react: ^18.0.0 || ^19.0.0
react-dom: ^18.0.0 || ^19.0.0
peerDependenciesMeta:
'@types/react':
optional: true
'@types/react-dom':
optional: true
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
'@tufjs/canonical-json@2.0.0':
resolution: {integrity: sha512-yVtV8zsdo8qFHe+/3kw81dSLyF7D576A5cCFCi4X7B39tWT7SekaEFUnvnWJHz+9qO7qJTah1JbrDjWKqFtdWA==}
engines: {node: ^16.14.0 || >=18.0.0}
'@tufjs/models@4.1.0':
resolution: {integrity: sha512-Y8cK9aggNRsqJVaKUlEYs4s7CvQ1b1ta2DVPyAimb0I2qhzjNk+A+mxvll/klL0RlfuIUei8BF7YWiua4kQqww==}
engines: {node: ^20.17.0 || >=22.9.0}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
'@types/aria-query@5.0.4':
resolution: {integrity: sha512-rfT93uj5s0PRL7EzccGMs3brplhcrghnDoV26NqKhCAS1hVo+WdNsPvE/yb6ilfr5hi2MEk6d5EWJTKdxg8jVw==}
'@types/babel__core@7.20.5':
resolution: {integrity: sha512-qoQprZvz5wQFJwMDqeseRXWv3rqMvhgpbXFfVyWhbx9X47POIA6i/+dXefEmZKoAgOaTdaIgNSMqMIU61yRyzA==}
'@types/babel__generator@7.27.0':
resolution: {integrity: sha512-ufFd2Xi92OAVPYsy+P4n7/U7e68fex0+Ee8gSG9KX7eo084CWiQ4sdxktvdl0bOPupXtVJPY19zk6EwWqUQ8lg==}
'@types/babel__template@7.4.4':
resolution: {integrity: sha512-h/NUaSyG5EyxBIp8YRxo4RMe2/qQgvyowRwVMzhYhBCONbW8PUsg4lkFMrhgZhUe5z3L3MiLDuvyJ/CaPa2A8A==}
'@types/babel__traverse@7.28.0':
resolution: {integrity: sha512-8PvcXf70gTDZBgt9ptxJ8elBeBjcLOAcOtoO/mPJjtji1+CdGbHgm77om1GrsPxsiE+uXIpNSK64UYaIwQXd4Q==}
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
'@types/bcrypt@5.0.2':
resolution: {integrity: sha512-6atioO8Y75fNcbmj0G7UjI9lXN2pQ/IGJ2FWT4a/btd0Lk9lQalHLKhkgKVZ3r+spnmWUKfbMi1GEe9wyHQfNQ==}
first commit
2026-02-21 03:10:39 +00:00
'@types/chai@5.2.3':
resolution: {integrity: sha512-Mw558oeA9fFbv65/y4mHtXDs9bPnFMZAL/jxdPFUpOHHIXX91mcgEHbS5Lahr+pwZFR8A7GQleRWeI6cGFC2UA==}
'@types/deep-eql@4.0.2':
resolution: {integrity: sha512-c9h9dVVMigMPc4bwTvC5dxqtqJZwQPePsWjPlpSOnojbor6pGqdk541lfA7AqFQr5pB1BRdq0juY9db81BwyFw==}
feat(mcpd): ResourceRevision + ResourceProposal services + Prompt revision integration Phase 2 of the Skills + Revisions + Proposals work. Stands up the generic revision/proposal layer and wires Prompt into it. Skills will plug into the same infrastructure in PR-3 with no further service changes required. This PR is intentionally additive: PromptRequest table and routes are unchanged. The /api/v1/proposals API runs side-by-side with the legacy /api/v1/promptrequests API. The PromptRequest cutover (rename + backfill + mcplocal rewire) is deferred to a later PR so this one stays reviewable. ## What's added ### Repositories (src/mcpd/src/repositories/) - resource-revision.repository.ts — append-only revision log keyed by (resourceType, resourceId). Soft FK; no relations declared. Supports history listing, semver lookup, and contentHash cross-resource search. - resource-proposal.repository.ts — generic propose queue. Status lifecycle pending → approved | rejected. Mirrors Prompt's `?? ''` workaround for nullable-FK compound lookups. ### Services (src/mcpd/src/services/) - resource-revision.service.ts — record() inserts a revision with a stable sha256 contentHash computed from canonicalised JSON (key-sorted at every level so reordered objects produce the same hash). Caller passes a pre-computed semver; service does NOT decide bump policy. - resource-proposal.service.ts — propose / approve / reject / list, with a per-resourceType handler registry. PromptService registers the 'prompt' handler at construction; the SkillService will register 'skill' in PR-3. approve() runs in a Prisma $transaction so the resource update + revision insert + proposal status flip are atomic. ### Pure utility (src/mcpd/src/utils/semver.ts) - bumpSemver(current, kind) for major / minor / patch - compareSemver(a, b) — numeric, not lex (10 > 9) - isValidSemver(s) - Invalid input falls back to '0.1.0' rather than throwing — keeps the audit-write path from blowing up the prompt update if a row's semver ever drifts out of MAJOR.MINOR.PATCH shape. ### Routes (src/mcpd/src/routes/) - revisions.ts — GET /api/v1/revisions?resourceType=&resourceId=, GET /api/v1/revisions/:id, GET /api/v1/revisions/:id/diff?against=<id|live> (unified-format diff via the `diff` package), and POST /api/v1/prompts/:id/restore-revision { revisionId, note? }. - proposals.ts — GET / POST /api/v1/proposals, GET /api/v1/proposals/:id, PUT for body updates, POST .../approve and POST .../reject, plus DELETE. ## What's changed - PromptService.create / update now record a ResourceRevision when the revision service is wired. Update auto-bumps patch on content change; authors can override via `--bump major|minor|patch` or `--semver X.Y.Z` on the CLI (forwarded into the PUT body). Best-effort: revision write failures are swallowed so the prompt save still succeeds (revision is audit, not source of truth). - PromptService.setProposalService registers a 'prompt' approval handler with the proposal service. Approval runs in a Prisma transaction: upsert prompt → record revision → update currentRevisionId → flip proposal status. semver bumps to 0.1.0 on first approval, patch thereafter. - New CLI flags on `mcpctl edit prompt`: --bump, --semver, --note. They're prompt-only (validated client-side); other resources reject them. - Aliases in shared.ts: `proposal`/`prop` → proposals, `revision`/`rev` → revisions. - diff dependency added to mcpd. ## Tests - src/mcpd/tests/utils/semver.test.ts — covers bump/compare/validate including numeric (not lex) semver compare and invalid-input fallback. - prompt-service.test.ts updated: makePrompt fixture now sets semver + agentId + currentRevisionId; updatePrompt assertion expects the auto-bumped patch in the same update call. - prompt-routes.test.ts updated symmetrically. ## RBAC `proposals` and `revisions` URL segments map to the existing `prompts` permission for now. PR-7 may split if a "reviewer" role becomes useful. ## Verification Full suite: 158 test files / 2127 tests green. `pnpm build` clean across all 6 workspace packages. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-07 00:38:35 +01:00
'@types/diff@5.2.3':
resolution: {integrity: sha512-K0Oqlrq3kQMaO2RhfrNQX5trmt+XLyom88zS0u84nnIcLvFnRUMRRHmrGny5GSM+kNO9IZLARsdQHDzkhAgmrQ==}
feat: eager vLLM warmup and smart page titles in paginate stage - Add warmup() to LlmProvider interface for eager subprocess startup - ManagedVllmProvider.warmup() starts vLLM in background on project load - ProviderRegistry.warmupAll() triggers all managed providers - NamedProvider proxies warmup() to inner provider - paginate stage generates LLM-powered descriptive page titles when available, cached by content hash, falls back to generic "Page N" - project-mcp-endpoint calls warmupAll() on router creation so vLLM is loading while the session initializes Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-03-03 19:07:39 +00:00
'@types/diff@8.0.0':
resolution: {integrity: sha512-o7jqJM04gfaYrdCecCVMbZhNdG6T1MHg/oQoRFdERLV+4d+V7FijhiEAbFu0Usww84Yijk9yH58U4Jk4HbtzZw==}
deprecated: This is a stub types definition. diff provides its own type definitions, so you do not need this installed.
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
'@types/docker-modem@3.0.6':
resolution: {integrity: sha512-yKpAGEuKRSS8wwx0joknWxsmLha78wNMe9R2S3UNsVOkZded8UqOrV8KoeDXoXsjndxwyF3eIhyClGbO1SEhEg==}
'@types/dockerode@4.0.1':
resolution: {integrity: sha512-cmUpB+dPN955PxBEuXE3f6lKO1hHiIGYJA46IVF3BJpNsZGvtBDcRnlrHYHtOH/B6vtDOyl2kZ2ShAu3mgc27Q==}
first commit
2026-02-21 03:10:39 +00:00
'@types/esrecurse@4.3.1':
resolution: {integrity: sha512-xJBAbDifo5hpffDBuHl0Y8ywswbiAp/Wi7Y/GtAgSlZyIABppyurxVueOPE8LUQOxdlgi6Zqce7uoEpqNTeiUw==}
'@types/estree@1.0.8':
resolution: {integrity: sha512-dWHzHa2WqEXI/O1E9OjrocMTKJl2mSrEolh1Iomrv6U+JuNwaHXsXx9bLu5gG7BUWFIN0skIQJQ/L1rIex4X6w==}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
'@types/google-protobuf@3.15.12':
resolution: {integrity: sha512-40um9QqwHjRS92qnOaDpL7RmDK15NuZYo9HihiJRbYkMQZlWnuH8AdvbMy8/o6lgLmKbDUKa+OALCltHdbOTpQ==}
feat: build CLI core framework with Commander.js Add CLI entry point with Commander.js, config management (~/.mcpctl/config.json with Zod validation), output formatters (table/json/yaml), config and status commands with dependency injection for testing. Fix sanitizeString regex ordering. 67 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:17:31 +00:00
'@types/js-yaml@4.0.9':
resolution: {integrity: sha512-k4MGaQl5TGo/iipqb2UDG2UwjXziSWkh0uysQelTlJpX1qGlpUZYm8PnO4DxG1qBomtJUdYJ6qR6xdIah10JLg==}
first commit
2026-02-21 03:10:39 +00:00
'@types/json-schema@7.0.15':
resolution: {integrity: sha512-5+fP8P8MFNC+AyZCDxrB2pkZFPGzqQWUzpSeuuVLvm8VMcorNYavBqoFcxK8bQz4Qsbn4oUEEem4wDLfcysGHA==}
feat: add Kubernetes orchestrator for MCP server pod management mcpd can now deploy MCP server instances as Kubernetes pods instead of Docker containers. Set MCPD_ORCHESTRATOR=kubernetes to enable. - Add @kubernetes/client-node with thin wrapper (context enforcement via MCPD_K8S_CONTEXT to prevent multi-cluster mishaps) - Rewrite KubernetesOrchestrator: pod CRUD, pod IP extraction, exec via SPDY (one-shot + interactive), log streaming - Manifest generator: stdin:true for STDIO servers, args (not command) to preserve runner image entrypoint, security hardening - Orchestrator selection in main.ts via MCPD_ORCHESTRATOR env var - 25 unit tests for k8s orchestrator, all 624 tests pass Tested end-to-end on local k3s: - mcpd deployed via Pulumi, creates pods in mcpctl-servers namespace - NetworkPolicy verified: only mcpd can reach MCP server pods - Python runner (uvx) successfully runs aws-documentation-mcp-server Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-08 01:55:13 +01:00
'@types/node-fetch@2.6.13':
resolution: {integrity: sha512-QGpRVpzSaUs30JBSGPjOg4Uveu384erbHBoT1zeONvyCfwQxIkUshLAOqN/k9EjGviPRmWTTe6aH2qySWKTVSw==}
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
'@types/node@18.19.130':
resolution: {integrity: sha512-GRaXQx6jGfL8sKfaIDD6OupbIHBr9jv7Jnaml9tB7l4v068PAOXqfcujMMo5PhbIs6ggR1XODELqahT2R8v0fg==}
feat: add Kubernetes orchestrator for MCP server pod management mcpd can now deploy MCP server instances as Kubernetes pods instead of Docker containers. Set MCPD_ORCHESTRATOR=kubernetes to enable. - Add @kubernetes/client-node with thin wrapper (context enforcement via MCPD_K8S_CONTEXT to prevent multi-cluster mishaps) - Rewrite KubernetesOrchestrator: pod CRUD, pod IP extraction, exec via SPDY (one-shot + interactive), log streaming - Manifest generator: stdin:true for STDIO servers, args (not command) to preserve runner image entrypoint, security hardening - Orchestrator selection in main.ts via MCPD_ORCHESTRATOR env var - 25 unit tests for k8s orchestrator, all 624 tests pass Tested end-to-end on local k3s: - mcpd deployed via Pulumi, creates pods in mcpctl-servers namespace - NetworkPolicy verified: only mcpd can reach MCP server pods - Python runner (uvx) successfully runs aws-documentation-mcp-server Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-08 01:55:13 +01:00
'@types/node@24.12.2':
resolution: {integrity: sha512-A1sre26ke7HDIuY/M23nd9gfB+nrmhtYyMINbjI1zHJxYteKR6qSMX56FsmjMcDb3SMcjJg5BiRRgOCC/yBD0g==}
feat: implement MCP registry client with multi-source search Add registry client that queries Official, Glama, and Smithery MCP registries with caching, request deduplication, retry logic, and result ranking/dedup. Includes 53 tests covering all components. Also fix null priority values in cancelled tasks (19-21) that broke Task Master, and add new tasks 25-27 for registry completion and CLI discover/install commands. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 03:46:14 +00:00
'@types/node@25.3.0':
resolution: {integrity: sha512-4K3bqJpXpqfg2XKGK9bpDTc6xO/xoUP/RBWS7AtRMug6zZFaRekiLzjVtAoZMquxoAbzBvy5nxQ7veS5eYzf8A==}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
'@types/react-dom@19.2.3':
resolution: {integrity: sha512-jp2L/eY6fn+KgVVQAOqYItbF0VY/YApe5Mz2F0aykSO8gx31bYCZyvSeYxCHKvzHG5eZjc+zyaS5BrBWya2+kQ==}
peerDependencies:
'@types/react': ^19.2.0
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
'@types/react@19.2.14':
resolution: {integrity: sha512-ilcTH/UniCkMdtexkoCN0bI7pMcJDvmQFPvuPvmEaYA/NSfFTAgdUSLAoVjaRJm7+6PvcM+q1zYOwS4wTYMF9w==}
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
'@types/retry@0.12.0':
resolution: {integrity: sha512-wWKOClTTiizcZhXnPY4wikVAwmdYHp8q6DmC+EJUzAMsycb7HB32Kh9RN4+0gExjmPmZSAQjgURXIGATPegAvA==}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
'@types/semver@7.7.1':
resolution: {integrity: sha512-FmgJfu+MOcQ370SD0ev7EI8TlCAfKYU+B4m5T3yXc1CiRN94g/SZPtsCkk506aUDtlMnFZvasDwHHUcZUEaYuA==}
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
'@types/ssh2@1.15.5':
resolution: {integrity: sha512-N1ASjp/nXH3ovBHddRJpli4ozpk6UdDYIX4RJWFa9L1YKnzdhTlVmiGHm4DZnj/jLbqZpes4aeR30EFGQtvhQQ==}
feat: add Kubernetes orchestrator for MCP server pod management mcpd can now deploy MCP server instances as Kubernetes pods instead of Docker containers. Set MCPD_ORCHESTRATOR=kubernetes to enable. - Add @kubernetes/client-node with thin wrapper (context enforcement via MCPD_K8S_CONTEXT to prevent multi-cluster mishaps) - Rewrite KubernetesOrchestrator: pod CRUD, pod IP extraction, exec via SPDY (one-shot + interactive), log streaming - Manifest generator: stdin:true for STDIO servers, args (not command) to preserve runner image entrypoint, security hardening - Orchestrator selection in main.ts via MCPD_ORCHESTRATOR env var - 25 unit tests for k8s orchestrator, all 624 tests pass Tested end-to-end on local k3s: - mcpd deployed via Pulumi, creates pods in mcpctl-servers namespace - NetworkPolicy verified: only mcpd can reach MCP server pods - Python runner (uvx) successfully runs aws-documentation-mcp-server Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-08 01:55:13 +01:00
'@types/stream-buffers@3.0.8':
resolution: {integrity: sha512-J+7VaHKNvlNPJPEJXX/fKa9DZtR/xPMwuIbe+yNOwp1YB+ApUOBv2aUpEoBJEi8nJgbgs1x8e73ttg0r1rSUdw==}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
'@types/trusted-types@2.0.7':
resolution: {integrity: sha512-ScaPdn1dQczgbl0QFTeTOmVHFULt394XJgOQNoyVhZ6r2vLnMLJfBPd53SB52T/3G36VI1/g2MZaX0cwDuXsfw==}
first commit
2026-02-21 03:10:39 +00:00
'@typescript-eslint/eslint-plugin@8.56.0':
resolution: {integrity: sha512-lRyPDLzNCuae71A3t9NEINBiTn7swyOhvUj3MyUOxb8x6g6vPEFoOU+ZRmGMusNC3X3YMhqMIX7i8ShqhT74Pw==}
engines: {node: ^18.18.0 || ^20.9.0 || >=21.1.0}
peerDependencies:
'@typescript-eslint/parser': ^8.56.0
eslint: ^8.57.0 || ^9.0.0 || ^10.0.0
typescript: '>=4.8.4 <6.0.0'
'@typescript-eslint/parser@8.56.0':
resolution: {integrity: sha512-IgSWvLobTDOjnaxAfDTIHaECbkNlAlKv2j5SjpB2v7QHKv1FIfjwMy8FsDbVfDX/KjmCmYICcw7uGaXLhtsLNg==}
engines: {node: ^18.18.0 || ^20.9.0 || >=21.1.0}
peerDependencies:
eslint: ^8.57.0 || ^9.0.0 || ^10.0.0
typescript: '>=4.8.4 <6.0.0'
'@typescript-eslint/project-service@8.56.0':
resolution: {integrity: sha512-M3rnyL1vIQOMeWxTWIW096/TtVP+8W3p/XnaFflhmcFp+U4zlxUxWj4XwNs6HbDeTtN4yun0GNTTDBw/SvufKg==}
engines: {node: ^18.18.0 || ^20.9.0 || >=21.1.0}
peerDependencies:
typescript: '>=4.8.4 <6.0.0'
'@typescript-eslint/scope-manager@8.56.0':
resolution: {integrity: sha512-7UiO/XwMHquH+ZzfVCfUNkIXlp/yQjjnlYUyYz7pfvlK3/EyyN6BK+emDmGNyQLBtLGaYrTAI6KOw8tFucWL2w==}
engines: {node: ^18.18.0 || ^20.9.0 || >=21.1.0}
'@typescript-eslint/tsconfig-utils@8.56.0':
resolution: {integrity: sha512-bSJoIIt4o3lKXD3xmDh9chZcjCz5Lk8xS7Rxn+6l5/pKrDpkCwtQNQQwZ2qRPk7TkUYhrq3WPIHXOXlbXP0itg==}
engines: {node: ^18.18.0 || ^20.9.0 || >=21.1.0}
peerDependencies:
typescript: '>=4.8.4 <6.0.0'
'@typescript-eslint/type-utils@8.56.0':
resolution: {integrity: sha512-qX2L3HWOU2nuDs6GzglBeuFXviDODreS58tLY/BALPC7iu3Fa+J7EOTwnX9PdNBxUI7Uh0ntP0YWGnxCkXzmfA==}
engines: {node: ^18.18.0 || ^20.9.0 || >=21.1.0}
peerDependencies:
eslint: ^8.57.0 || ^9.0.0 || ^10.0.0
typescript: '>=4.8.4 <6.0.0'
'@typescript-eslint/types@8.56.0':
resolution: {integrity: sha512-DBsLPs3GsWhX5HylbP9HNG15U0bnwut55Lx12bHB9MpXxQ+R5GC8MwQe+N1UFXxAeQDvEsEDY6ZYwX03K7Z6HQ==}
engines: {node: ^18.18.0 || ^20.9.0 || >=21.1.0}
'@typescript-eslint/typescript-estree@8.56.0':
resolution: {integrity: sha512-ex1nTUMWrseMltXUHmR2GAQ4d+WjkZCT4f+4bVsps8QEdh0vlBsaCokKTPlnqBFqqGaxilDNJG7b8dolW2m43Q==}
engines: {node: ^18.18.0 || ^20.9.0 || >=21.1.0}
peerDependencies:
typescript: '>=4.8.4 <6.0.0'
'@typescript-eslint/utils@8.56.0':
resolution: {integrity: sha512-RZ3Qsmi2nFGsS+n+kjLAYDPVlrzf7UhTffrDIKr+h2yzAlYP/y5ZulU0yeDEPItos2Ph46JAL5P/On3pe7kDIQ==}
engines: {node: ^18.18.0 || ^20.9.0 || >=21.1.0}
peerDependencies:
eslint: ^8.57.0 || ^9.0.0 || ^10.0.0
typescript: '>=4.8.4 <6.0.0'
'@typescript-eslint/visitor-keys@8.56.0':
resolution: {integrity: sha512-q+SL+b+05Ud6LbEE35qe4A99P+htKTKVbyiNEe45eCbJFyh/HVK9QXwlrbz+Q4L8SOW4roxSVwXYj4DMBT7Ieg==}
engines: {node: ^18.18.0 || ^20.9.0 || >=21.1.0}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
'@vitejs/plugin-react@5.2.0':
resolution: {integrity: sha512-YmKkfhOAi3wsB1PhJq5Scj3GXMn3WvtQ/JC0xoopuHoXSdmtdStOpFrYaT1kie2YgFBcIe64ROzMYRjCrYOdYw==}
engines: {node: ^20.19.0 || >=22.12.0}
peerDependencies:
vite: ^4.2.0 || ^5.0.0 || ^6.0.0 || ^7.0.0 || ^8.0.0
first commit
2026-02-21 03:10:39 +00:00
'@vitest/coverage-v8@4.0.18':
resolution: {integrity: sha512-7i+N2i0+ME+2JFZhfuz7Tg/FqKtilHjGyGvoHYQ6iLV0zahbsJ9sljC9OcFcPDbhYKCet+sG8SsVqlyGvPflZg==}
peerDependencies:
'@vitest/browser': 4.0.18
vitest: 4.0.18
peerDependenciesMeta:
'@vitest/browser':
optional: true
'@vitest/expect@4.0.18':
resolution: {integrity: sha512-8sCWUyckXXYvx4opfzVY03EOiYVxyNrHS5QxX3DAIi5dpJAAkyJezHCP77VMX4HKA2LDT/Jpfo8i2r5BE3GnQQ==}
'@vitest/mocker@4.0.18':
resolution: {integrity: sha512-HhVd0MDnzzsgevnOWCBj5Otnzobjy5wLBe4EdeeFGv8luMsGcYqDuFRMcttKWZA5vVO8RFjexVovXvAM4JoJDQ==}
peerDependencies:
msw: ^2.4.9
vite: ^6.0.0 || ^7.0.0-0
peerDependenciesMeta:
msw:
optional: true
vite:
optional: true
'@vitest/pretty-format@4.0.18':
resolution: {integrity: sha512-P24GK3GulZWC5tz87ux0m8OADrQIUVDPIjjj65vBXYG17ZeU3qD7r+MNZ1RNv4l8CGU2vtTRqixrOi9fYk/yKw==}
'@vitest/runner@4.0.18':
resolution: {integrity: sha512-rpk9y12PGa22Jg6g5M3UVVnTS7+zycIGk9ZNGN+m6tZHKQb7jrP7/77WfZy13Y/EUDd52NDsLRQhYKtv7XfPQw==}
'@vitest/snapshot@4.0.18':
resolution: {integrity: sha512-PCiV0rcl7jKQjbgYqjtakly6T1uwv/5BQ9SwBLekVg/EaYeQFPiXcgrC2Y7vDMA8dM1SUEAEV82kgSQIlXNMvA==}
'@vitest/spy@4.0.18':
resolution: {integrity: sha512-cbQt3PTSD7P2OARdVW3qWER5EGq7PHlvE+QfzSC0lbwO+xnt7+XH06ZzFjFRgzUX//JmpxrCu92VdwvEPlWSNw==}
'@vitest/utils@4.0.18':
resolution: {integrity: sha512-msMRKLMVLWygpK3u2Hybgi4MNjcYJvwTb0Ru09+fOyCXIgT5raYP041DRRdiJiI3k/2U6SEbAETB3YtBrUkCFA==}
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
abbrev@1.1.1:
resolution: {integrity: sha512-nne9/IiQ/hzIhY6pdDnbBtz7DjPTKrY00P/zvPSm5pOFkl6xuGrGnXn/VtTNNfNtAfZ9/1RtehkszU9qcTii0Q==}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
abbrev@4.0.0:
resolution: {integrity: sha512-a1wflyaL0tHtJSmLSOVybYhy22vRih4eduhhrkcjgrWGnRfrZtovJ2FRjxuTtkkj47O/baf0R86QU5OuYpz8fA==}
engines: {node: ^20.17.0 || >=22.9.0}
first commit
2026-02-21 03:10:39 +00:00
abstract-logging@2.0.1:
resolution: {integrity: sha512-2BjRTZxTPvheOvGbBslFSYOUkr+SjPtOnrLP33f+VIWLzezQpZcqVg7ja3L4dBXmzzgwT+a029jRx5PCi3JuiA==}
accepts@2.0.0:
resolution: {integrity: sha512-5cvg6CtKwfgdmVqY1WIiXKc3Q1bkRqGLi+2W/6ao+6Y7gu/RCwRuAhGEzh5B4KlszSuTLgZYuqFqo5bImjNKng==}
engines: {node: '>= 0.6'}
acorn-jsx@5.3.2:
resolution: {integrity: sha512-rq9s+JNhf0IChjtDXxllJ7g41oZk5SlXtp0LHwyA5cejwn7vKmKp4pPri6YEePv2PU65sAsegbXtIinmDFDXgQ==}
peerDependencies:
acorn: ^6.0.0 || ^7.0.0 || ^8.0.0
acorn@8.16.0:
resolution: {integrity: sha512-UVJyE9MttOsBQIDKw1skb9nAwQuR5wuGD3+82K6JgJlm/Y+KI92oNsMNGZCYdDsVtRHSak0pcV5Dno5+4jh9sw==}
engines: {node: '>=0.4.0'}
hasBin: true
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
agent-base@6.0.2:
resolution: {integrity: sha512-RZNwNclF7+MS/8bDg70amg32dyeZGZxiDuQmZxKLAlQjr3jGyLx+4Kkk58UO7D2QdgFIQCovuSuZESne6RG6XQ==}
engines: {node: '>= 6.0.0'}
feat: add Kubernetes orchestrator for MCP server pod management mcpd can now deploy MCP server instances as Kubernetes pods instead of Docker containers. Set MCPD_ORCHESTRATOR=kubernetes to enable. - Add @kubernetes/client-node with thin wrapper (context enforcement via MCPD_K8S_CONTEXT to prevent multi-cluster mishaps) - Rewrite KubernetesOrchestrator: pod CRUD, pod IP extraction, exec via SPDY (one-shot + interactive), log streaming - Manifest generator: stdin:true for STDIO servers, args (not command) to preserve runner image entrypoint, security hardening - Orchestrator selection in main.ts via MCPD_ORCHESTRATOR env var - 25 unit tests for k8s orchestrator, all 624 tests pass Tested end-to-end on local k3s: - mcpd deployed via Pulumi, creates pods in mcpctl-servers namespace - NetworkPolicy verified: only mcpd can reach MCP server pods - Python runner (uvx) successfully runs aws-documentation-mcp-server Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-08 01:55:13 +01:00
agent-base@7.1.4:
resolution: {integrity: sha512-MnA+YT8fwfJPgBx3m60MNqakm30XOkyIoH1y6huTQvC0PwZG7ki8NacLBcrPbNoo8vEZy7Jpuk7+jMO+CUovTQ==}
engines: {node: '>= 14'}
first commit
2026-02-21 03:10:39 +00:00
ajv-formats@3.0.1:
resolution: {integrity: sha512-8iUql50EUR+uUcdRQ3HDqa6EVyo3docL8g5WJ3FNcWmu62IbkGUue/pEyLBW8VGKKucTPgqeks4fIU1DA4yowQ==}
peerDependencies:
ajv: ^8.0.0
peerDependenciesMeta:
ajv:
optional: true
ajv@6.14.0:
resolution: {integrity: sha512-IWrosm/yrn43eiKqkfkHis7QioDleaXQHdDVPKg0FSwwd/DuvyX79TZnFOnYpB7dcsFAMmtFztZuXPDvSePkFw==}
ajv@8.18.0:
resolution: {integrity: sha512-PlXPeEWMXMZ7sPYOHqmDyCJzcfNrUr3fGNKtezX14ykXOEIvyK81d+qydx89KY5O71FKMPaQ2vBfBFI5NHR63A==}
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
ansi-escapes@7.3.0:
resolution: {integrity: sha512-BvU8nYgGQBxcmMuEeUEmNTvrMVjJNSH7RgW24vXexN4Ven6qCvy4TntnvlnwnMLTVlcRQQdbRY8NKnaIoeWDNg==}
engines: {node: '>=18'}
first commit
2026-02-21 03:10:39 +00:00
ansi-regex@5.0.1:
resolution: {integrity: sha512-quJQXlTSUGL2LH9SUXo8VwsY4soanhgo6LNSm84E1LBcE8s3O0wpdiRzyR9z/ZZJMlMWv37qOOb9pdJlMUEKFQ==}
engines: {node: '>=8'}
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
ansi-regex@6.2.2:
resolution: {integrity: sha512-Bq3SmSpyFHaWjPk8If9yc6svM8c56dB5BAtW4Qbw5jHTwwXXcTLoRMkpDJp6VL0XzlWaCHTXrkFURMYmD0sLqg==}
engines: {node: '>=12'}
first commit
2026-02-21 03:10:39 +00:00
ansi-styles@4.3.0:
resolution: {integrity: sha512-zbB9rCJAT1rbjiVDb2hqKFHNYLxgtk8NURxZ3IZwD3F6NtxbXZQCnnSi1Lkx+IDohdPlFp222wVALIheZJQSEg==}
engines: {node: '>=8'}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
ansi-styles@5.2.0:
resolution: {integrity: sha512-Cxwpt2SfTzTtXcfOlzGEee8O+c+MmUgGrNiBcXnuWxuFJHe6a5Hz7qwhwe5OgaSYI0IJvkLqWX1ASG+cJOkEiA==}
engines: {node: '>=10'}
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
ansi-styles@6.2.3:
resolution: {integrity: sha512-4Dj6M28JB+oAH8kFkTLUo+a2jwOFkuqb3yucU0CANcRRUbxS0cP0nZYCGjcc3BNXwRIsUVmDGgzawme7zvJHvg==}
engines: {node: '>=12'}
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
aproba@2.1.0:
resolution: {integrity: sha512-tLIEcj5GuR2RSTnxNKdkK0dJ/GrC7P38sUkiDmDuHfsHmbagTFAxDVIBltoklXEVIQ/f14IL8IMJ5pn9Hez1Ew==}
are-we-there-yet@2.0.0:
resolution: {integrity: sha512-Ci/qENmwHnsYo9xKIcUJN5LeDKdJ6R1Z1j9V/J5wyq8nh/mYPEpIKJbBZXtZjG04HiK7zV/p6Vs9952MrMeUIw==}
engines: {node: '>=10'}
deprecated: This package is no longer supported.
first commit
2026-02-21 03:10:39 +00:00
argparse@2.0.1:
resolution: {integrity: sha512-8+9WqebbFzpX9OR+Wa6O29asIogeRMzcGtAINdpMHHyAg10f05aSFVBbcEqGf/PXw1EjAZ+q2/bEBg3DvurK3Q==}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
aria-query@5.3.0:
resolution: {integrity: sha512-b0P0sZPKtyu8HkeRAfCq0IfURZK+SuwMjY1UXGBU27wpAiTwQAIlq56IbIO+ytk/JjS1fMR14ee5WBBfKi5J6A==}
aria-query@5.3.2:
resolution: {integrity: sha512-COROpnaoap1E2F000S62r6A60uHZnmlvomhfyT2DlTcrY1OrBKn2UhH7qn5wTC9zMvD0AY7csdPSNwKP+7WiQw==}
engines: {node: '>= 0.4'}
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
asn1@0.2.6:
resolution: {integrity: sha512-ix/FxPn0MDjeyJ7i/yoHGFt/EX6LyNbxSEhPPXODPL+KB0VPk86UYfL0lMdy+KCnv+fmvIzySwaK5COwqVbWTQ==}
first commit
2026-02-21 03:10:39 +00:00
assertion-error@2.0.1:
resolution: {integrity: sha512-Izi8RQcffqCeNVgFigKli1ssklIbpHnCYc6AknXGYoB6grJqyeby7jv12JUQgmTAnIDnbck1uxksT4dzN3PWBA==}
engines: {node: '>=12'}
ast-v8-to-istanbul@0.3.11:
resolution: {integrity: sha512-Qya9fkoofMjCBNVdWINMjB5KZvkYfaO9/anwkWnjxibpWUxo5iHl2sOdP7/uAqaRuUYuoo8rDwnbaaKVFxoUvw==}
feat: add Kubernetes orchestrator for MCP server pod management mcpd can now deploy MCP server instances as Kubernetes pods instead of Docker containers. Set MCPD_ORCHESTRATOR=kubernetes to enable. - Add @kubernetes/client-node with thin wrapper (context enforcement via MCPD_K8S_CONTEXT to prevent multi-cluster mishaps) - Rewrite KubernetesOrchestrator: pod CRUD, pod IP extraction, exec via SPDY (one-shot + interactive), log streaming - Manifest generator: stdin:true for STDIO servers, args (not command) to preserve runner image entrypoint, security hardening - Orchestrator selection in main.ts via MCPD_ORCHESTRATOR env var - 25 unit tests for k8s orchestrator, all 624 tests pass Tested end-to-end on local k3s: - mcpd deployed via Pulumi, creates pods in mcpctl-servers namespace - NetworkPolicy verified: only mcpd can reach MCP server pods - Python runner (uvx) successfully runs aws-documentation-mcp-server Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-08 01:55:13 +01:00
asynckit@0.4.0:
resolution: {integrity: sha512-Oei9OH4tRh0YqU3GxhX79dM/mwVgvbZJaSNaRk+bshkj0S5cfHcgYakreBjrHwatXKbz+IoIdYLxrKim2MjW0Q==}
first commit
2026-02-21 03:10:39 +00:00
atomic-sleep@1.0.0:
resolution: {integrity: sha512-kNOjDqAh7px0XWNI+4QbzoiR/nTkHAWNud2uvnJquD1/x5a7EQZMJT0AczqK0Qn67oY/TTQ1LbUKajZpp3I9tQ==}
engines: {node: '>=8.0.0'}
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
auto-bind@5.0.1:
resolution: {integrity: sha512-ooviqdwwgfIfNmDwo94wlshcdzfO64XV0Cg6oDsDYBJfITDz1EngD2z7DkbvCWn+XIMsIqW27sEVF6qcpJrRcg==}
engines: {node: ^12.20.0 || ^14.13.1 || >=16.0.0}
first commit
2026-02-21 03:10:39 +00:00
avvio@9.2.0:
resolution: {integrity: sha512-2t/sy01ArdHHE0vRH5Hsay+RtCZt3dLPji7W7/MMOCEgze5b7SNDC4j5H6FnVgPkI1MTNFGzHdHrVXDDl7QSSQ==}
feat: add Kubernetes orchestrator for MCP server pod management mcpd can now deploy MCP server instances as Kubernetes pods instead of Docker containers. Set MCPD_ORCHESTRATOR=kubernetes to enable. - Add @kubernetes/client-node with thin wrapper (context enforcement via MCPD_K8S_CONTEXT to prevent multi-cluster mishaps) - Rewrite KubernetesOrchestrator: pod CRUD, pod IP extraction, exec via SPDY (one-shot + interactive), log streaming - Manifest generator: stdin:true for STDIO servers, args (not command) to preserve runner image entrypoint, security hardening - Orchestrator selection in main.ts via MCPD_ORCHESTRATOR env var - 25 unit tests for k8s orchestrator, all 624 tests pass Tested end-to-end on local k3s: - mcpd deployed via Pulumi, creates pods in mcpctl-servers namespace - NetworkPolicy verified: only mcpd can reach MCP server pods - Python runner (uvx) successfully runs aws-documentation-mcp-server Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-08 01:55:13 +01:00
b4a@1.8.0:
resolution: {integrity: sha512-qRuSmNSkGQaHwNbM7J78Wwy+ghLEYF1zNrSeMxj4Kgw6y33O3mXcQ6Ie9fRvfU/YnxWkOchPXbaLb73TkIsfdg==}
peerDependencies:
react-native-b4a: '*'
peerDependenciesMeta:
react-native-b4a:
optional: true
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
babel-plugin-jsx-dom-expressions@0.40.7:
resolution: {integrity: sha512-/O6JWUmjv03OI9lL2ry9bUjpD5S3PclM55RRJEyCdcFZ5W2SEA/59d+l2hNsk3gI6kiWRdRPdOtqZmsQzFN1pQ==}
peerDependencies:
'@babel/core': ^7.20.12
babel-plugin-module-resolver@5.0.2:
resolution: {integrity: sha512-9KtaCazHee2xc0ibfqsDeamwDps6FZNo5S0Q81dUqEuFzVwPhcT4J5jOqIVvgCA3Q/wO9hKYxN/Ds3tIsp5ygg==}
babel-preset-solid@1.9.12:
resolution: {integrity: sha512-LLqnuKVDlKpyBlMPcH6qEvs/wmS9a+NczppxJ3ryS/c0O5IiSFOIBQi9GzyiGDSbcJpx4Gr87jyFTos1MyEuWg==}
peerDependencies:
'@babel/core': ^7.0.0
solid-js: ^1.9.12
peerDependenciesMeta:
solid-js:
optional: true
first commit
2026-02-21 03:10:39 +00:00
balanced-match@1.0.2:
resolution: {integrity: sha512-3oSeUO0TMV67hN1AmbXsK4yaqU7tjiHlbxRDZOpH0KW9+CeX4bRAaX0Anxt0tx2MrpRpWwQaPwIlISEJhYU5Pw==}
balanced-match@4.0.3:
resolution: {integrity: sha512-1pHv8LX9CpKut1Zp4EXey7Z8OfH11ONNH6Dhi2WDUt31VVZFXZzKwXcysBgqSumFCmR+0dqjMK5v5JiFHzi0+g==}
engines: {node: 20 || >=22}
feat: add Kubernetes orchestrator for MCP server pod management mcpd can now deploy MCP server instances as Kubernetes pods instead of Docker containers. Set MCPD_ORCHESTRATOR=kubernetes to enable. - Add @kubernetes/client-node with thin wrapper (context enforcement via MCPD_K8S_CONTEXT to prevent multi-cluster mishaps) - Rewrite KubernetesOrchestrator: pod CRUD, pod IP extraction, exec via SPDY (one-shot + interactive), log streaming - Manifest generator: stdin:true for STDIO servers, args (not command) to preserve runner image entrypoint, security hardening - Orchestrator selection in main.ts via MCPD_ORCHESTRATOR env var - 25 unit tests for k8s orchestrator, all 624 tests pass Tested end-to-end on local k3s: - mcpd deployed via Pulumi, creates pods in mcpctl-servers namespace - NetworkPolicy verified: only mcpd can reach MCP server pods - Python runner (uvx) successfully runs aws-documentation-mcp-server Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-08 01:55:13 +01:00
bare-events@2.8.2:
resolution: {integrity: sha512-riJjyv1/mHLIPX4RwiK+oW9/4c3TEUeORHKefKAKnZ5kyslbN+HXowtbaVEqt4IMUB7OXlfixcs6gsFeo/jhiQ==}
peerDependencies:
bare-abort-controller: '*'
peerDependenciesMeta:
bare-abort-controller:
optional: true
bare-fs@4.6.0:
resolution: {integrity: sha512-2YkS7NuiJceSEbyEOdSNLE9tsGd+f4+f7C+Nik/MCk27SYdwIMPT/yRKvg++FZhQXgk0KWJKJyXX9RhVV0RGqA==}
engines: {bare: '>=1.16.0'}
peerDependencies:
bare-buffer: '*'
peerDependenciesMeta:
bare-buffer:
optional: true
bare-os@3.8.7:
resolution: {integrity: sha512-G4Gr1UsGeEy2qtDTZwL7JFLo2wapUarz7iTMcYcMFdS89AIQuBoyjgXZz0Utv7uHs3xA9LckhVbeBi8lEQrC+w==}
engines: {bare: '>=1.14.0'}
bare-path@3.0.0:
resolution: {integrity: sha512-tyfW2cQcB5NN8Saijrhqn0Zh7AnFNsnczRcuWODH0eYAXBsJ5gVxAUuNr7tsHSC6IZ77cA0SitzT+s47kot8Mw==}
bare-stream@2.12.0:
resolution: {integrity: sha512-w28i8lkBgREV3rPXGbgK+BO66q+ZpKqRWrZLiCdmmUlLPrQ45CzkvRhN+7lnv00Gpi2zy5naRxnUFAxCECDm9g==}
peerDependencies:
bare-abort-controller: '*'
bare-buffer: '*'
bare-events: '*'
peerDependenciesMeta:
bare-abort-controller:
optional: true
bare-buffer:
optional: true
bare-events:
optional: true
bare-url@2.4.0:
resolution: {integrity: sha512-NSTU5WN+fy/L0DDenfE8SXQna4voXuW0FHM7wH8i3/q9khUSchfPbPezO4zSFMnDGIf9YE+mt/RWhZgNRKRIXA==}
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
base64-js@1.5.1:
resolution: {integrity: sha512-AKpaYlHn8t4SVbOHCy+b5+KKgvR4vrsD8vbvrbiQJps7fKDTkjkDry6ji0rUJjC0kzbNePLwzxq8iypo41qeWA==}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
baseline-browser-mapping@2.10.23:
resolution: {integrity: sha512-xwVXGqevyKPsiuQdLj+dZMVjidjJV508TBqexND5HrF89cGdCYCJFB3qhcxRHSeMctdCfbR1jrxBajhDy7o29g==}
engines: {node: '>=6.0.0'}
hasBin: true
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
bcrypt-pbkdf@1.0.2:
resolution: {integrity: sha512-qeFIXtP4MSoi6NLqO12WfqARWWuCKi2Rn/9hJLEmtB5yTNr9DqFWkJRCf2qShWzPeAMRnOgCrq0sg/KLv5ES9w==}
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
bcrypt@5.1.1:
resolution: {integrity: sha512-AGBHOG5hPYZ5Xl9KXzU5iKq9516yEmvCKDg3ecP5kX2aB6UqTeXZxk2ELnDgDm6BQSMlLt9rDB4LoSMx0rYwww==}
engines: {node: '>= 10.0.0'}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
bidi-js@1.0.3:
resolution: {integrity: sha512-RKshQI1R3YQ+n9YJz2QQ147P66ELpa1FQEg20Dk8oW9t2KgLbpDLLp9aGZ7y8WHSshDknG0bknqGw5/tyCs5tw==}
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
bignumber.js@9.3.1:
resolution: {integrity: sha512-Ko0uX15oIUS7wJ3Rb30Fs6SkVbLmPBAKdlm7q9+ak9bbIeFf0MwuBsQV6z7+X768/cHsfg+WlysDWJcmthjsjQ==}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
bin-links@6.0.2:
resolution: {integrity: sha512-frE1t78WOwJ45PKV2cF2tNPjTcs9L1J9s6VkrV59wanRP4GlaomuxYPVma7BwthMg8WnfSory4w5PTE6FZZ81w==}
engines: {node: ^20.17.0 || >=22.9.0}
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
bl@4.1.0:
resolution: {integrity: sha512-1W07cM9gS6DcLperZfFSj+bWLtaPGSOHWhPiGzXmvVJbRLdG82sH/Kn8EtW1VqWVA54AKf2h5k5BbnIbwF3h6w==}
first commit
2026-02-21 03:10:39 +00:00
body-parser@2.2.2:
resolution: {integrity: sha512-oP5VkATKlNwcgvxi0vM0p/D3n2C3EReYVX+DNYs5TjZFn/oQt2j+4sVJtSMr18pdRr8wjTcBl6LoV+FUwzPmNA==}
engines: {node: '>=18'}
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
bowser@2.14.1:
resolution: {integrity: sha512-tzPjzCxygAKWFOJP011oxFHs57HzIhOEracIgAePE4pqB3LikALKnSzUyU4MGs9/iCEUuHlAJTjTc5M+u7YEGg==}
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
brace-expansion@1.1.12:
resolution: {integrity: sha512-9T9UjW3r0UW5c1Q7GTwllptXwhvYmEzFhzMfZ9H7FQWt+uZePjZPjBP/W1ZEyZ1twGWom5/56TF4lPcqjnDHcg==}
first commit
2026-02-21 03:10:39 +00:00
brace-expansion@2.0.2:
resolution: {integrity: sha512-Jt0vHyM+jmUBqojB7E1NIYadt0vI0Qxjxd2TErW94wDz+E2LAm5vKMXXwg6ZZBTHPuUlDgQHKXvjGBdfcF1ZDQ==}
brace-expansion@5.0.2:
resolution: {integrity: sha512-Pdk8c9poy+YhOgVWw1JNN22/HcivgKWwpxKq04M/jTmHyCZn12WPJebZxdjSa5TmBqISrUSgNYU3eRORljfCCw==}
engines: {node: 20 || >=22}
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
brace-expansion@5.0.9:
resolution: {integrity: sha512-ScQ4IuvIEF1TMlP7Zt+vjJ//9zlPb2SDcxWxM3bk8s6t6GGdJ7KO1dCcTidOPJKePW30LE/2cT7wCyPho9/Wxg==}
engines: {node: 20 || >=22}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
browserslist@4.28.2:
resolution: {integrity: sha512-48xSriZYYg+8qXna9kwqjIVzuQxi+KYWp2+5nCYnYKPTr0LvD89Jqk2Or5ogxz0NUMfIjhh2lIUX/LyX9B4oIg==}
engines: {node: ^6 || ^7 || ^8 || ^9 || ^10 || ^11 || ^12 || >=13.7}
hasBin: true
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
buffer-equal-constant-time@1.0.1:
resolution: {integrity: sha512-zRpUiDwd/xk6ADqPMATG8vc9VPrkck7T07OIx0gnjmJAnHnTVXNQG3vfvWNuiZIkwu9KrKdA1iJKfsfTVxE6NA==}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
buffer-from@1.1.2:
resolution: {integrity: sha512-E+XQCRwSbaaiChtv6k6Dwgc+bx+Bs6vuKJHHl5kox/BaKbhiXzqQOwK4cO22yElGp2OCmjwVhT3HmxgyPGnJfQ==}
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
buffer@5.7.1:
resolution: {integrity: sha512-EHcyIPBQ4BSGlvjB16k5KgAJ27CIsHY/2JBmCRReo48y9rQ3MaUzWX3KVlBa4U7MyX02HdVj0K7C3WaB3ju7FQ==}
buildcheck@0.0.7:
resolution: {integrity: sha512-lHblz4ahamxpTmnsk+MNTRWsjYKv965MwOrSJyeD588rR3Jcu7swE+0wN5F+PbL5cjgu/9ObkhfzEPuofEMwLA==}
engines: {node: '>=10.0.0'}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
bun-ffi-structs@0.2.4:
resolution: {integrity: sha512-AJzsqoVFs1KBbJbWHIYrVZLDC3NhTqqh25awRXqzoLzmBAKr5oqk6+CwuYHAekKx+VBCYVohBoKuRq40dV+TYg==}
peerDependencies:
typescript: ^5
first commit
2026-02-21 03:10:39 +00:00
bytes@3.1.2:
resolution: {integrity: sha512-/Nf7TyzTx6S3yRJObOAV7956r8cr2+Oj8AC5dt8wSP3BQAoeX58NoHyCU8P8zGkNXStjTSi6fzO6F0pBdcYbEg==}
engines: {node: '>= 0.8'}
c12@3.1.0:
resolution: {integrity: sha512-uWoS8OU1MEIsOv8p/5a82c3H31LsWVR5qiyXVfBNOzfffjUWtPnhAb4BYI2uG2HfGmZmFjCtui5XNWaps+iFuw==}
peerDependencies:
magicast: ^0.3.5
peerDependenciesMeta:
magicast:
optional: true
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
cacache@20.0.4:
resolution: {integrity: sha512-M3Lab8NPYlZU2exsL3bMVvMrMqgwCnMWfdZbK28bn3pK6APT/Te/I8hjRPNu1uwORY9a1eEQoifXbKPQMfMTOA==}
engines: {node: ^20.17.0 || >=22.9.0}
first commit
2026-02-21 03:10:39 +00:00
call-bind-apply-helpers@1.0.2:
resolution: {integrity: sha512-Sp1ablJ0ivDkSzjcaJdxEunN5/XvksFJ2sMBFfq6x0ryhQV/2b/KwFe21cMpmHtPOSij8K99/wSfoEuTObmuMQ==}
engines: {node: '>= 0.4'}
call-bound@1.0.4:
resolution: {integrity: sha512-+ys997U96po4Kx/ABpBCqhA9EuxJaQWDQg7295H4hBphv3IZg0boBKuwYpt4YXp6MZ5AmZQnU/tyMTlRpaSejg==}
engines: {node: '>= 0.4'}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
caniuse-lite@1.0.30001791:
resolution: {integrity: sha512-yk0l/YSrOnFZk3UROpDLQD9+kC1l4meK/wed583AXrzoarMGJcbRi2Q4RaUYbKxYAsZ8sWmaSa/DsLmdBeI1vQ==}
first commit
2026-02-21 03:10:39 +00:00
chai@6.2.2:
resolution: {integrity: sha512-NUPRluOfOiTKBKvWPtSD4PhFvWCqOi0BGStNWs57X9js7XGTprSmFoz5F0tWhR4WPjNeR9jXqdC7/UpSJTnlRg==}
engines: {node: '>=18'}
chalk@5.6.2:
resolution: {integrity: sha512-7NzBL0rN6fMUW+f7A6Io4h40qQlG+xGmtMxfbnH/K7TAtt8JQWVQK+6g0UXKMeVJoyV5EkkNsErQ8pVD3bLHbA==}
engines: {node: ^12.17.0 || ^14.13 || >=16.0.0}
chardet@2.1.1:
resolution: {integrity: sha512-PsezH1rqdV9VvyNhxxOW32/d75r01NY7TQCmOqomRo15ZSOKbpTFVsfjghxo6JloQUCGnH4k1LGu0R4yCLlWQQ==}
chokidar@4.0.3:
resolution: {integrity: sha512-Qgzu8kfBvo+cA4962jnP1KkS6Dop5NS6g7R5LFYJr4b8Ub94PPQXUksCw9PvXoeXPRRddRNC5C1JQUR2SMGtnA==}
engines: {node: '>= 14.16.0'}
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
chownr@1.1.4:
resolution: {integrity: sha512-jJ0bqzaylmJtVnNgzTeSOs8DPavpbYgEr/b0YL8/2GO3xJEhInFmhKMUnEJQjZumK7KXGFhUy89PrsJWlakBVg==}
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
chownr@2.0.0:
resolution: {integrity: sha512-bIomtDF5KGpdogkLd9VspvFzk9KfpyyGlS8YFVZl7TGPBHL5snIOnxeshwVgPteQ9b4Eydl+pVbIyE1DcvCWgQ==}
engines: {node: '>=10'}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
chownr@3.0.0:
resolution: {integrity: sha512-+IxzY9BZOQd/XuYPRmrvEVjF/nqj5kgT4kEq7VofrDoM1MxoRjEWkrCC3EtLi59TVawxTAn+orJwFQcrqEN1+g==}
engines: {node: '>=18'}
first commit
2026-02-21 03:10:39 +00:00
citty@0.1.6:
resolution: {integrity: sha512-tskPPKEs8D2KPafUypv2gxwJP8h/OaJmC82QQGGDQcHvXX43xF2VDACcJVmZ0EuSxkpO9Kc4MlrA3q0+FG58AQ==}
citty@0.2.1:
resolution: {integrity: sha512-kEV95lFBhQgtogAPlQfJJ0WGVSokvLr/UEoFPiKKOXF7pl98HfUVUD0ejsuTCld/9xH9vogSywZ5KqHzXrZpqg==}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
cjs-module-lexer@2.2.0:
resolution: {integrity: sha512-4bHTS2YuzUvtoLjdy+98ykbNB5jS0+07EvFNXerqZQJ89F7DI6ET7OQo/HJuW6K0aVsKA9hj9/RVb2kQVOrPDQ==}
feat(web): bold redesign — Tailwind v4 + shadcn-style primitives + Skills/Proposals/Revisions UI Phase 6 of the Skills + Revisions + Proposals work. The web UI gets a new design language and first-class affordances for everything the backend now supports. ## Visual direction - Tailwind v4 with custom @theme block (oklch tokens). Dark-mode-only (internal tool — light mode doubles QA surface). - Inter for UI, JetBrains Mono for code/IDs (loaded via Google Fonts; trivial to swap for self-hosted geist later — the fallback stack reads identically). - Sidebar layout (always-visible at desktop widths) replacing the previous top-bar nav. Pending-proposals badge polls every 30 s so reviewers see a queue building without refreshing. - Lucide icons throughout. - Spacing and radii on Tailwind defaults. Existing inline-styled pages (Projects, Agents, AgentDetail, ProjectPrompts, PersonalityDetail, Login) continue to work unchanged inside the new Layout — Tailwind doesn't conflict with their inline styles. A follow-up can migrate them incrementally. ## What's added ### Build infra (src/web/) - package.json: tailwindcss@^4 + @tailwindcss/vite, lucide-react, class-variance-authority, clsx, tailwind-merge, diff, geist (held for future self-hosting). - vite.config.ts: registers the @tailwindcss/vite plugin. - src/index.css: Tailwind import + @theme tokens + @layer base. - src/main.tsx: imports index.css. - src/lib/utils.ts: shadcn-style cn() helper. ### shadcn-style primitives (src/components/ui/) Hand-written rather than generated via `npx shadcn` so the repo doesn't depend on a CLI tool that needs interactive runtime: - button.tsx — variants: primary / secondary / ghost / danger / link; sizes: sm / md / lg / icon. - card.tsx — Card + Header/Title/Description/Content/Footer subparts. - badge.tsx — variants: default / info / success / warning / danger / outline. - input.tsx — Input + Textarea + Label. - tabs.tsx — no-dep accessible Tabs (no Radix needed for our use). - separator.tsx — h/v separator with role=separator. ### Diff component (src/components/Diff.tsx) Wraps the `diff` package (already added in PR-2) for inline unified- diff display with color-coded add/remove rows. Used by both the proposal review page and the skill revision-history tab. ### New pages (src/pages/) - Dashboard.tsx — at-a-glance home. Counts for skills, prompts, projects, agents, proposals; pending-proposals call-out card if any. - Skills.tsx — list view, separated into Global vs Project/Agent- scoped sections. - SkillDetail.tsx — name + semver + description; tabs for SKILL.md / Files / Metadata / History. History tab shows revisions with click-to-diff against the live body. - Proposals.tsx — queue with Pending/Approved/Rejected tabs. Pending count is highlighted in amber. - ProposalDetail.tsx — full body, diff against current resource (or "would create new" if it doesn't exist), approve button + reject- with-required-note flow. ### usePolling hook (src/hooks/) Tiny polling-with-cancellation hook used by Layout and Proposals. ### Layout rewrite (src/components/Layout.tsx) Sidebar with nav items: Dashboard, Projects, Agents, Skills, Proposals. Lucide icons. Active-route highlighting via NavLink. Pending-proposals warning badge on the Proposals item. ### Routes (src/App.tsx) New routes: /dashboard, /skills, /skills/:name, /proposals, /proposals/:id. Default redirects to /dashboard. ### API types (src/api.ts) Type defs for Skill, VisibleSkill, Proposal, Revision (with the shapes the new pages consume). ## Tests Existing 7 web tests still pass (Login + api). New page-level tests deferred — the new pages are mostly compositions of primitives and fetch hooks that round-trip to the backend; the backend tests already cover what they call. PR-7 polish can add render-and-click tests if coverage drift surfaces. ## Verification - `pnpm --filter @mcpctl/web build` clean, no warnings. - `pnpm test:run` whole monorepo: 162 test files / 2157 tests green. - Visual smoke deferred — needs a running mcpd to populate the fixtures. Manual smoke tested locally is the next step. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-07 17:54:55 +01:00
class-variance-authority@0.7.1:
resolution: {integrity: sha512-Ka+9Trutv7G8M6WT6SeiRWz792K5qEqIGEGzXKhAE6xOWAY6pPH8U+9IY3oCMv6kqTmLsv7Xh/2w2RigkePMsg==}
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
cli-boxes@3.0.0:
resolution: {integrity: sha512-/lzGpEWL/8PfI0BmBOPRwp0c/wFNX1RdUML3jK/RcSBA9T8mZDdQpqYBKtCFTOfQbwPqWEOpjqW+Fnayc0969g==}
engines: {node: '>=10'}
cli-cursor@4.0.0:
resolution: {integrity: sha512-VGtlMu3x/4DOtIUwEkRezxUZ2lBacNJCHash0N0WeZDBS+7Ux1dm3XWAgWYxLJFMMdOeXMHXorshEFhbMSGelg==}
engines: {node: ^12.20.0 || ^14.13.1 || >=16.0.0}
cli-spinners@3.4.0:
resolution: {integrity: sha512-bXfOC4QcT1tKXGorxL3wbJm6XJPDqEnij2gQ2m7ESQuE+/z9YFIWnl/5RpTiKWbMq3EVKR4fRLJGn6DVfu0mpw==}
engines: {node: '>=18.20'}
cli-truncate@5.1.1:
resolution: {integrity: sha512-SroPvNHxUnk+vIW/dOSfNqdy1sPEFkrTk6TUtqLCnBlo3N7TNYYkzzN7uSD6+jVjrdO4+p8nH7JzH6cIvUem6A==}
engines: {node: '>=20'}
first commit
2026-02-21 03:10:39 +00:00
cli-width@4.1.0:
resolution: {integrity: sha512-ouuZd4/dm2Sw5Gmqy6bGyNNNe1qt9RpmxveLSO7KcgsTnU7RXfsw+/bukWGo1abgBiMAic068rclZsO4IWmmxQ==}
engines: {node: '>= 12'}
feat(web): bold redesign — Tailwind v4 + shadcn-style primitives + Skills/Proposals/Revisions UI Phase 6 of the Skills + Revisions + Proposals work. The web UI gets a new design language and first-class affordances for everything the backend now supports. ## Visual direction - Tailwind v4 with custom @theme block (oklch tokens). Dark-mode-only (internal tool — light mode doubles QA surface). - Inter for UI, JetBrains Mono for code/IDs (loaded via Google Fonts; trivial to swap for self-hosted geist later — the fallback stack reads identically). - Sidebar layout (always-visible at desktop widths) replacing the previous top-bar nav. Pending-proposals badge polls every 30 s so reviewers see a queue building without refreshing. - Lucide icons throughout. - Spacing and radii on Tailwind defaults. Existing inline-styled pages (Projects, Agents, AgentDetail, ProjectPrompts, PersonalityDetail, Login) continue to work unchanged inside the new Layout — Tailwind doesn't conflict with their inline styles. A follow-up can migrate them incrementally. ## What's added ### Build infra (src/web/) - package.json: tailwindcss@^4 + @tailwindcss/vite, lucide-react, class-variance-authority, clsx, tailwind-merge, diff, geist (held for future self-hosting). - vite.config.ts: registers the @tailwindcss/vite plugin. - src/index.css: Tailwind import + @theme tokens + @layer base. - src/main.tsx: imports index.css. - src/lib/utils.ts: shadcn-style cn() helper. ### shadcn-style primitives (src/components/ui/) Hand-written rather than generated via `npx shadcn` so the repo doesn't depend on a CLI tool that needs interactive runtime: - button.tsx — variants: primary / secondary / ghost / danger / link; sizes: sm / md / lg / icon. - card.tsx — Card + Header/Title/Description/Content/Footer subparts. - badge.tsx — variants: default / info / success / warning / danger / outline. - input.tsx — Input + Textarea + Label. - tabs.tsx — no-dep accessible Tabs (no Radix needed for our use). - separator.tsx — h/v separator with role=separator. ### Diff component (src/components/Diff.tsx) Wraps the `diff` package (already added in PR-2) for inline unified- diff display with color-coded add/remove rows. Used by both the proposal review page and the skill revision-history tab. ### New pages (src/pages/) - Dashboard.tsx — at-a-glance home. Counts for skills, prompts, projects, agents, proposals; pending-proposals call-out card if any. - Skills.tsx — list view, separated into Global vs Project/Agent- scoped sections. - SkillDetail.tsx — name + semver + description; tabs for SKILL.md / Files / Metadata / History. History tab shows revisions with click-to-diff against the live body. - Proposals.tsx — queue with Pending/Approved/Rejected tabs. Pending count is highlighted in amber. - ProposalDetail.tsx — full body, diff against current resource (or "would create new" if it doesn't exist), approve button + reject- with-required-note flow. ### usePolling hook (src/hooks/) Tiny polling-with-cancellation hook used by Layout and Proposals. ### Layout rewrite (src/components/Layout.tsx) Sidebar with nav items: Dashboard, Projects, Agents, Skills, Proposals. Lucide icons. Active-route highlighting via NavLink. Pending-proposals warning badge on the Proposals item. ### Routes (src/App.tsx) New routes: /dashboard, /skills, /skills/:name, /proposals, /proposals/:id. Default redirects to /dashboard. ### API types (src/api.ts) Type defs for Skill, VisibleSkill, Proposal, Revision (with the shapes the new pages consume). ## Tests Existing 7 web tests still pass (Login + api). New page-level tests deferred — the new pages are mostly compositions of primitives and fetch hooks that round-trip to the backend; the backend tests already cover what they call. PR-7 polish can add render-and-click tests if coverage drift surfaces. ## Verification - `pnpm --filter @mcpctl/web build` clean, no warnings. - `pnpm test:run` whole monorepo: 162 test files / 2157 tests green. - Visual smoke deferred — needs a running mcpd to populate the fixtures. Manual smoke tested locally is the next step. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-07 17:54:55 +01:00
client-only@0.0.1:
resolution: {integrity: sha512-IV3Ou0jSMzZrd3pZ48nLkT9DA7Ag1pnPzaiQhpW7c3RbcqqzvzzVu+L8gfqMp/8IM2MQtSiqaCxrrcfu8I8rMA==}
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
cliui@8.0.1:
resolution: {integrity: sha512-BSeNnyus75C4//NQ9gQt1/csTXyo/8Sb+afLAkzAptFuMsod9HFokGNudZpi/oQV73hnVK+sR+5PVRMd+Dr7YQ==}
engines: {node: '>=12'}
feat(web): bold redesign — Tailwind v4 + shadcn-style primitives + Skills/Proposals/Revisions UI Phase 6 of the Skills + Revisions + Proposals work. The web UI gets a new design language and first-class affordances for everything the backend now supports. ## Visual direction - Tailwind v4 with custom @theme block (oklch tokens). Dark-mode-only (internal tool — light mode doubles QA surface). - Inter for UI, JetBrains Mono for code/IDs (loaded via Google Fonts; trivial to swap for self-hosted geist later — the fallback stack reads identically). - Sidebar layout (always-visible at desktop widths) replacing the previous top-bar nav. Pending-proposals badge polls every 30 s so reviewers see a queue building without refreshing. - Lucide icons throughout. - Spacing and radii on Tailwind defaults. Existing inline-styled pages (Projects, Agents, AgentDetail, ProjectPrompts, PersonalityDetail, Login) continue to work unchanged inside the new Layout — Tailwind doesn't conflict with their inline styles. A follow-up can migrate them incrementally. ## What's added ### Build infra (src/web/) - package.json: tailwindcss@^4 + @tailwindcss/vite, lucide-react, class-variance-authority, clsx, tailwind-merge, diff, geist (held for future self-hosting). - vite.config.ts: registers the @tailwindcss/vite plugin. - src/index.css: Tailwind import + @theme tokens + @layer base. - src/main.tsx: imports index.css. - src/lib/utils.ts: shadcn-style cn() helper. ### shadcn-style primitives (src/components/ui/) Hand-written rather than generated via `npx shadcn` so the repo doesn't depend on a CLI tool that needs interactive runtime: - button.tsx — variants: primary / secondary / ghost / danger / link; sizes: sm / md / lg / icon. - card.tsx — Card + Header/Title/Description/Content/Footer subparts. - badge.tsx — variants: default / info / success / warning / danger / outline. - input.tsx — Input + Textarea + Label. - tabs.tsx — no-dep accessible Tabs (no Radix needed for our use). - separator.tsx — h/v separator with role=separator. ### Diff component (src/components/Diff.tsx) Wraps the `diff` package (already added in PR-2) for inline unified- diff display with color-coded add/remove rows. Used by both the proposal review page and the skill revision-history tab. ### New pages (src/pages/) - Dashboard.tsx — at-a-glance home. Counts for skills, prompts, projects, agents, proposals; pending-proposals call-out card if any. - Skills.tsx — list view, separated into Global vs Project/Agent- scoped sections. - SkillDetail.tsx — name + semver + description; tabs for SKILL.md / Files / Metadata / History. History tab shows revisions with click-to-diff against the live body. - Proposals.tsx — queue with Pending/Approved/Rejected tabs. Pending count is highlighted in amber. - ProposalDetail.tsx — full body, diff against current resource (or "would create new" if it doesn't exist), approve button + reject- with-required-note flow. ### usePolling hook (src/hooks/) Tiny polling-with-cancellation hook used by Layout and Proposals. ### Layout rewrite (src/components/Layout.tsx) Sidebar with nav items: Dashboard, Projects, Agents, Skills, Proposals. Lucide icons. Active-route highlighting via NavLink. Pending-proposals warning badge on the Proposals item. ### Routes (src/App.tsx) New routes: /dashboard, /skills, /skills/:name, /proposals, /proposals/:id. Default redirects to /dashboard. ### API types (src/api.ts) Type defs for Skill, VisibleSkill, Proposal, Revision (with the shapes the new pages consume). ## Tests Existing 7 web tests still pass (Login + api). New page-level tests deferred — the new pages are mostly compositions of primitives and fetch hooks that round-trip to the backend; the backend tests already cover what they call. PR-7 polish can add render-and-click tests if coverage drift surfaces. ## Verification - `pnpm --filter @mcpctl/web build` clean, no warnings. - `pnpm test:run` whole monorepo: 162 test files / 2157 tests green. - Visual smoke deferred — needs a running mcpd to populate the fixtures. Manual smoke tested locally is the next step. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-07 17:54:55 +01:00
clsx@2.1.1:
resolution: {integrity: sha512-eYm0QWBtUrBWZWG0d386OGAw16Z995PiOVo2B7bjWSbHedGl5e0ZWaq65kOGgUSNesEIDkB9ISbTg/JK9dhCZA==}
engines: {node: '>=6'}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
cmd-shim@8.0.0:
resolution: {integrity: sha512-Jk/BK6NCapZ58BKUxlSI+ouKRbjH1NLZCgJkYoab+vEHUY3f6OzpNBN9u7HFSv9J6TRDGs4PLOHezoKGaFRSCA==}
engines: {node: ^20.17.0 || >=22.9.0}
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
code-excerpt@4.0.0:
resolution: {integrity: sha512-xxodCmBen3iy2i0WtAK8FlFNrRzjUqjRsMfho58xT/wvZU1YTM3fCnRjcy1gJPMepaRlgm/0e6w8SpWHpn3/cA==}
engines: {node: ^12.20.0 || ^14.13.1 || >=16.0.0}
first commit
2026-02-21 03:10:39 +00:00
color-convert@2.0.1:
resolution: {integrity: sha512-RRECPsj7iu/xb5oKYcsFHSppFNnsj/52OVTRKb4zP5onXwVF3zVmmToNcOfGC+CRDpfK/U584fMg38ZHCaElKQ==}
engines: {node: '>=7.0.0'}
color-name@1.1.4:
resolution: {integrity: sha512-dOy+3AuW3a2wNbZHIuMZpTcgjGuLU/uBL/ubcZF9OXbDo8ff4O8yVp5Bf0efS8uEoYo5q4Fx7dY9OgQGXgAsQA==}
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
color-support@1.1.3:
resolution: {integrity: sha512-qiBjkpbMLO/HL68y+lh4q0/O1MZFj2RX6X/KmMa3+gJD3z+WwI1ZzDHysvqHGS3mP6mznPckpXmw1nI9cJjyRg==}
hasBin: true
feat: add Kubernetes orchestrator for MCP server pod management mcpd can now deploy MCP server instances as Kubernetes pods instead of Docker containers. Set MCPD_ORCHESTRATOR=kubernetes to enable. - Add @kubernetes/client-node with thin wrapper (context enforcement via MCPD_K8S_CONTEXT to prevent multi-cluster mishaps) - Rewrite KubernetesOrchestrator: pod CRUD, pod IP extraction, exec via SPDY (one-shot + interactive), log streaming - Manifest generator: stdin:true for STDIO servers, args (not command) to preserve runner image entrypoint, security hardening - Orchestrator selection in main.ts via MCPD_ORCHESTRATOR env var - 25 unit tests for k8s orchestrator, all 624 tests pass Tested end-to-end on local k3s: - mcpd deployed via Pulumi, creates pods in mcpctl-servers namespace - NetworkPolicy verified: only mcpd can reach MCP server pods - Python runner (uvx) successfully runs aws-documentation-mcp-server Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-08 01:55:13 +01:00
combined-stream@1.0.8:
resolution: {integrity: sha512-FQN4MRfuJeHf7cBbBMJFXhKSDq+2kAArBlmRBvcvFE5BB1HZKXtSFASDhdlz9zOYwxh8lDdnvmMOe/+5cdoEdg==}
engines: {node: '>= 0.8'}
first commit
2026-02-21 03:10:39 +00:00
commander@13.1.0:
resolution: {integrity: sha512-/rFeCpNJQbhSZjGVwO9RFV3xPqbnERS8MmIQzCtD/zl6gpJuV/bMLuN92oG3F7d8oDEHHRrujSXNUr8fpjntKw==}
engines: {node: '>=18'}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
common-ancestor-path@2.0.0:
resolution: {integrity: sha512-dnN3ibLeoRf2HNC+OlCiNc5d2zxbLJXOtiZUudNFSXZrNSydxcCsSpRzXwfu7BBWCIfHPw+xTayeBvJCP/D8Ng==}
engines: {node: '>= 18'}
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
concat-map@0.0.1:
resolution: {integrity: sha512-/Srv4dswyQNBfohGpz9o6Yb3Gz3SrUDqBH5rTuhGR7ahtlbYKnVxw2bCFMRljaA7EXHaXZ8wsHdodFvbkhKmqg==}
first commit
2026-02-21 03:10:39 +00:00
confbox@0.2.4:
resolution: {integrity: sha512-ysOGlgTFbN2/Y6Cg3Iye8YKulHw+R2fNXHrgSmXISQdMnomY6eNDprVdW9R5xBguEqI954+S6709UyiO7B+6OQ==}
consola@3.4.2:
resolution: {integrity: sha512-5IKcdX0nnYavi6G7TtOhwkYzyjfJlatbjMjuLSfE2kYT5pMDOilZ4OvMhi637CcDICTmz3wARPoyhqyX1Y+XvA==}
engines: {node: ^14.18.0 || >=16.10.0}
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
console-control-strings@1.1.0:
resolution: {integrity: sha512-ty/fTekppD2fIwRvnZAVdeOiGd1c7YXEixbgJTNzqcxJWKQnjJ/V1bNEEE6hygpM3WjwHFUVK6HTjWSzV4a8sQ==}
feat(mcpd+deploy): serve web UI at /ui + smoke tests + docs (Stage 6) The closing stage. mcpd now hosts the Stage 5 SPA, the Docker image bundles the build artifact, a smoke test exercises the personality HTTP surface end-to-end, and the user-facing docs spell out the mental model. mcpd: - Add @fastify/static dep. - New routes/web-ui.ts: registers /ui/* against a static bundle. Looks for the bundle at $MCPD_WEB_ROOT, then /usr/share/mcpd/web (the Docker image path), then a dev-tree fallback. Logs and skips cleanly if missing — API-only deploys keep working. - SPA fallback: any /ui/<path> that doesn't match a file falls through to index.html so direct hits to react-router URLs work. - /ui/* falls through to `kind: skip` in mapUrlToPermission, so the static assets are served unauthenticated. Each API call from the SPA still carries the bearer token. Deploy: - Dockerfile.mcpd builds the @mcpctl/web bundle in the same builder stage and copies dist/ to /usr/share/mcpd/web in the runtime image. Smoke (personality.smoke.test.ts): - Live mcpd flow: create secret/llm/agent/personality, attach an agent-direct prompt, verify the binding listing, reject double- attach (409) + foreign-agent prompt (400), set defaultPersonality by name, detach + delete cleanup. Docs: - New docs/personalities.md: VLAN-on-ethernet model, system-block ordering table, three prompt scopes, CLI walkthrough, web UI walkthrough, full API surface, RBAC notes. - agents.md and chat.md cross-link. - README's Agents section gains a Personalities subsection. Test count after Stage 6: mcpd: 801/801 cli: 430/430 web: 7/7 db: 58/62 (4 pre-existing) Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:48:43 +01:00
content-disposition@0.5.4:
resolution: {integrity: sha512-FveZTNuGw04cxlAiWbzi6zTAL/lhehaWbTtgluJh4/E95DqMwTmha3KZN1aAWA8cFIhHzMZUvLevkw5Rqk+tSQ==}
engines: {node: '>= 0.6'}
first commit
2026-02-21 03:10:39 +00:00
content-disposition@1.0.1:
resolution: {integrity: sha512-oIXISMynqSqm241k6kcQ5UwttDILMK4BiurCfGEREw6+X9jkkpEe5T9FZaApyLGGOnFuyMWZpdolTXMtvEJ08Q==}
engines: {node: '>=18'}
content-type@1.0.5:
resolution: {integrity: sha512-nTjqfcBFEipKdXCv4YDQWCfmcLZKm81ldF0pAopTvyrFGVbcR6P/VAAd5G7N+0tTr8QqiU0tFadD6FK4NtJwOA==}
engines: {node: '>= 0.6'}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
convert-source-map@2.0.0:
resolution: {integrity: sha512-Kvp459HrV2FEJ1CAsi1Ku+MY3kasH19TFykTz2xWmMeq6bk2NU3XXvfJ+Q61m0xktWwt+1HSYf3JZsTms3aRJg==}
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
convert-to-spaces@2.0.1:
resolution: {integrity: sha512-rcQ1bsQO9799wq24uE5AM2tAILy4gXGIK/njFWcVQkGNZ96edlpY+A7bjwvzjYvLDyzmG1MmMLZhpcsb+klNMQ==}
engines: {node: ^12.20.0 || ^14.13.1 || >=16.0.0}
first commit
2026-02-21 03:10:39 +00:00
cookie-signature@1.2.2:
resolution: {integrity: sha512-D76uU73ulSXrD1UXF4KE2TMxVVwhsnCgfAyTg9k8P6KGZjlXKrOLe4dJQKI3Bxi5wjesZoFXJWElNWBjPZMbhg==}
engines: {node: '>=6.6.0'}
cookie@0.7.2:
resolution: {integrity: sha512-yki5XnKuf750l50uGTllt6kKILY4nQ1eNIQatoXEByZ5dWgnKqbnqmTrBE5B4N7lrMJKQ2ytWMiTO2o0v6Ew/w==}
engines: {node: '>= 0.6'}
cookie@1.1.1:
resolution: {integrity: sha512-ei8Aos7ja0weRpFzJnEA9UHJ/7XQmqglbRwnf2ATjcB9Wq874VKH9kfjjirM6UhU2/E5fFYadylyhFldcqSidQ==}
engines: {node: '>=18'}
cors@2.8.6:
resolution: {integrity: sha512-tJtZBBHA6vjIAaF6EnIaq6laBBP9aq/Y3ouVJjEfoHbRBcHBAHYcMh/w8LDrk2PvIMMq8gmopa5D4V8RmbrxGw==}
engines: {node: '>= 0.10'}
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
cpu-features@0.0.10:
resolution: {integrity: sha512-9IkYqtX3YHPCzoVg1Py+o9057a3i0fp7S530UWokCSaFVTc7CwXPRiOjRjBQQ18ZCNafx78YfnG+HALxtVmOGA==}
engines: {node: '>=10.0.0'}
first commit
2026-02-21 03:10:39 +00:00
cross-spawn@7.0.6:
resolution: {integrity: sha512-uV2QOWP2nWzsy2aMp8aRibhi9dlzF5Hgh5SHaB9OiTGEyDTiJJyx0uy51QXdyWbtAHNua4XJzUKca3OzKUd3vA==}
engines: {node: '>= 8'}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
css-tree@3.2.1:
resolution: {integrity: sha512-X7sjQzceUhu1u7Y/ylrRZFU2FS6LRiFVp6rKLPg23y3x3c3DOKAwuXGDp+PAGjh6CSnCjYeAul8pcT8bAl+lSA==}
engines: {node: ^10 || ^12.20.0 || ^14.13.0 || >=15.0.0}
css.escape@1.5.1:
resolution: {integrity: sha512-YUifsXXuknHlUsmlgyY0PKzgPOr7/FjCePfHNt0jxm83wHZi44VDMQ7/fGNkjY3/jV1MC+1CmZbaHzugyeRtpg==}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
cssesc@3.0.0:
resolution: {integrity: sha512-/Tb/JcjK111nNScGob5MNtsntNM1aCNUDipB/TkwZFhyDrrE47SOx/18wF2bbjgc3ZzCSKW1T5nt5EbFoAz/Vg==}
engines: {node: '>=4'}
hasBin: true
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
cssstyle@6.2.0:
resolution: {integrity: sha512-Fm5NvhYathRnXNVndkUsCCuR63DCLVVwGOOwQw782coXFi5HhkXdu289l59HlXZBawsyNccXfWRYvLzcDCdDig==}
engines: {node: '>=20'}
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
csstype@3.2.3:
resolution: {integrity: sha512-z1HGKcYy2xA8AGQfwrn0PAy+PB7X/GSj3UVJW9qKyn43xWa+gl5nXmU4qqLMRzWVLFC8KusUX8T/0kCiOYpAIQ==}
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
data-uri-to-buffer@4.0.1:
resolution: {integrity: sha512-0R9ikRb668HB7QDxT1vkpuUBtqc53YyAwMwGeUFKRojY/NWKvdZ+9UYtRfGmhqNbRkTSVpMbmyhXipFFv2cb/A==}
engines: {node: '>= 12'}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
data-urls@7.0.0:
resolution: {integrity: sha512-23XHcCF+coGYevirZceTVD7NdJOqVn+49IHyxgszm+JIiHLoB2TkmPtsYkNWT1pvRSGkc35L6NHs0yHkN2SumA==}
engines: {node: ^20.19.0 || ^22.12.0 || >=24.0.0}
first commit
2026-02-21 03:10:39 +00:00
debug@4.4.3:
resolution: {integrity: sha512-RGwwWnwQvkVfavKVt22FGLw+xYSdzARwm0ru6DhTVA3umU5hZc28V3kO4stgYryrTlLpuvgI9GiijltAjNbcqA==}
engines: {node: '>=6.0'}
peerDependencies:
supports-color: '*'
peerDependenciesMeta:
supports-color:
optional: true
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
decimal.js@10.6.0:
resolution: {integrity: sha512-YpgQiITW3JXGntzdUmyUR1V812Hn8T1YVXhCu+wO3OpS4eU9l4YdD3qjyiKdV6mvV29zapkMeD390UVEf2lkUg==}
first commit
2026-02-21 03:10:39 +00:00
deep-is@0.1.4:
resolution: {integrity: sha512-oIPzksmTg4/MriiaYGO+okXDT7ztn/w3Eptv/+gSIdMdKsJo0u4CfYNFJPy+4SKMuCqGw2wxnA+URMg3t8a/bQ==}
deepmerge-ts@7.1.5:
resolution: {integrity: sha512-HOJkrhaYsweh+W+e74Yn7YStZOilkoPb6fycpwNLKzSPtruFs48nYis0zy5yJz1+ktUhHxoRDJ27RQAWLIJVJw==}
engines: {node: '>=16.0.0'}
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
deepmerge@4.3.1:
resolution: {integrity: sha512-3sUqbMEc77XqpdNO7FRyRog+eW3ph+GYCbj+rK+uYyRMuwsVy0rMiVtPn+QJlKFvWP/1PYpapqYn0Me2knFn+A==}
engines: {node: '>=0.10.0'}
first commit
2026-02-21 03:10:39 +00:00
defu@6.1.4:
resolution: {integrity: sha512-mEQCMmwJu317oSz8CwdIOdwf3xMif1ttiM8LTufzc3g6kR+9Pe236twL8j3IYT1F7GfRgGcW6MWxzZjLIkuHIg==}
feat: add Kubernetes orchestrator for MCP server pod management mcpd can now deploy MCP server instances as Kubernetes pods instead of Docker containers. Set MCPD_ORCHESTRATOR=kubernetes to enable. - Add @kubernetes/client-node with thin wrapper (context enforcement via MCPD_K8S_CONTEXT to prevent multi-cluster mishaps) - Rewrite KubernetesOrchestrator: pod CRUD, pod IP extraction, exec via SPDY (one-shot + interactive), log streaming - Manifest generator: stdin:true for STDIO servers, args (not command) to preserve runner image entrypoint, security hardening - Orchestrator selection in main.ts via MCPD_ORCHESTRATOR env var - 25 unit tests for k8s orchestrator, all 624 tests pass Tested end-to-end on local k3s: - mcpd deployed via Pulumi, creates pods in mcpctl-servers namespace - NetworkPolicy verified: only mcpd can reach MCP server pods - Python runner (uvx) successfully runs aws-documentation-mcp-server Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-08 01:55:13 +01:00
delayed-stream@1.0.0:
resolution: {integrity: sha512-ZySD7Nf91aLB0RxL4KGrKHBXl7Eds1DAmEdcoVawXnLD7SDhpNgtuII2aAkg7a7QS41jxPSZ17p4VdGnMHk3MQ==}
engines: {node: '>=0.4.0'}
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
delegates@1.0.0:
resolution: {integrity: sha512-bd2L678uiWATM6m5Z1VzNCErI3jiGzt6HGY8OVICs40JQq/HALfbyNJmp0UDakEY4pMMaN0Ly5om/B1VI/+xfQ==}
first commit
2026-02-21 03:10:39 +00:00
depd@2.0.0:
resolution: {integrity: sha512-g7nH6P6dyDioJogAAGprGpCtVImJhpPk/roCzdb3fIh61/s/nPsfR6onyMwkCAR/OlC3yBC0lESvUoQEAssIrw==}
engines: {node: '>= 0.8'}
dequal@2.0.3:
resolution: {integrity: sha512-0je+qPKHEMohvfRTCEo3CrPG6cAzAYgmzKyxRiYSSDkS6eGJdyVJm7WaYA5ECaAD9wLB2T4EEeymA5aFVcYXCA==}
engines: {node: '>=6'}
destr@2.0.5:
resolution: {integrity: sha512-ugFTXCtDZunbzasqBxrK93Ik/DRYsO6S/fedkWEMKqt04xZ4csmnmwGDBAb07QWNaGMAmnTIemsYZCksjATwsA==}
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
detect-libc@2.1.2:
resolution: {integrity: sha512-Btj2BOOO83o3WyH59e8MgXsxEQVcarkUOpEYrubB0urwnN10yQ364rsiByU11nZlqWYZm05i/of7io4mzihBtQ==}
engines: {node: '>=8'}
feat(mcpd): ResourceRevision + ResourceProposal services + Prompt revision integration Phase 2 of the Skills + Revisions + Proposals work. Stands up the generic revision/proposal layer and wires Prompt into it. Skills will plug into the same infrastructure in PR-3 with no further service changes required. This PR is intentionally additive: PromptRequest table and routes are unchanged. The /api/v1/proposals API runs side-by-side with the legacy /api/v1/promptrequests API. The PromptRequest cutover (rename + backfill + mcplocal rewire) is deferred to a later PR so this one stays reviewable. ## What's added ### Repositories (src/mcpd/src/repositories/) - resource-revision.repository.ts — append-only revision log keyed by (resourceType, resourceId). Soft FK; no relations declared. Supports history listing, semver lookup, and contentHash cross-resource search. - resource-proposal.repository.ts — generic propose queue. Status lifecycle pending → approved | rejected. Mirrors Prompt's `?? ''` workaround for nullable-FK compound lookups. ### Services (src/mcpd/src/services/) - resource-revision.service.ts — record() inserts a revision with a stable sha256 contentHash computed from canonicalised JSON (key-sorted at every level so reordered objects produce the same hash). Caller passes a pre-computed semver; service does NOT decide bump policy. - resource-proposal.service.ts — propose / approve / reject / list, with a per-resourceType handler registry. PromptService registers the 'prompt' handler at construction; the SkillService will register 'skill' in PR-3. approve() runs in a Prisma $transaction so the resource update + revision insert + proposal status flip are atomic. ### Pure utility (src/mcpd/src/utils/semver.ts) - bumpSemver(current, kind) for major / minor / patch - compareSemver(a, b) — numeric, not lex (10 > 9) - isValidSemver(s) - Invalid input falls back to '0.1.0' rather than throwing — keeps the audit-write path from blowing up the prompt update if a row's semver ever drifts out of MAJOR.MINOR.PATCH shape. ### Routes (src/mcpd/src/routes/) - revisions.ts — GET /api/v1/revisions?resourceType=&resourceId=, GET /api/v1/revisions/:id, GET /api/v1/revisions/:id/diff?against=<id|live> (unified-format diff via the `diff` package), and POST /api/v1/prompts/:id/restore-revision { revisionId, note? }. - proposals.ts — GET / POST /api/v1/proposals, GET /api/v1/proposals/:id, PUT for body updates, POST .../approve and POST .../reject, plus DELETE. ## What's changed - PromptService.create / update now record a ResourceRevision when the revision service is wired. Update auto-bumps patch on content change; authors can override via `--bump major|minor|patch` or `--semver X.Y.Z` on the CLI (forwarded into the PUT body). Best-effort: revision write failures are swallowed so the prompt save still succeeds (revision is audit, not source of truth). - PromptService.setProposalService registers a 'prompt' approval handler with the proposal service. Approval runs in a Prisma transaction: upsert prompt → record revision → update currentRevisionId → flip proposal status. semver bumps to 0.1.0 on first approval, patch thereafter. - New CLI flags on `mcpctl edit prompt`: --bump, --semver, --note. They're prompt-only (validated client-side); other resources reject them. - Aliases in shared.ts: `proposal`/`prop` → proposals, `revision`/`rev` → revisions. - diff dependency added to mcpd. ## Tests - src/mcpd/tests/utils/semver.test.ts — covers bump/compare/validate including numeric (not lex) semver compare and invalid-input fallback. - prompt-service.test.ts updated: makePrompt fixture now sets semver + agentId + currentRevisionId; updatePrompt assertion expects the auto-bumped patch in the same update call. - prompt-routes.test.ts updated symmetrically. ## RBAC `proposals` and `revisions` URL segments map to the existing `prompts` permission for now. PR-7 may split if a "reviewer" role becomes useful. ## Verification Full suite: 158 test files / 2127 tests green. `pnpm build` clean across all 6 workspace packages. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-07 00:38:35 +01:00
diff@5.2.2:
resolution: {integrity: sha512-vtcDfH3TOjP8UekytvnHH1o1P4FcUdt4eQ1Y+Abap1tk/OB2MWQvcwS2ClCd1zuIhc3JKOx6p3kod8Vfys3E+A==}
engines: {node: '>=0.3.1'}
feat: eager vLLM warmup and smart page titles in paginate stage - Add warmup() to LlmProvider interface for eager subprocess startup - ManagedVllmProvider.warmup() starts vLLM in background on project load - ProviderRegistry.warmupAll() triggers all managed providers - NamedProvider proxies warmup() to inner provider - paginate stage generates LLM-powered descriptive page titles when available, cached by content hash, falls back to generic "Page N" - project-mcp-endpoint calls warmupAll() on router creation so vLLM is loading while the session initializes Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-03-03 19:07:39 +00:00
diff@8.0.3:
resolution: {integrity: sha512-qejHi7bcSD4hQAZE0tNAawRK1ZtafHDmMTMkrrIGgSLl7hTnQHmKCeB45xAcbfTqK2zowkM3j3bHt/4b/ARbYQ==}
engines: {node: '>=0.3.1'}
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
diff@8.0.4:
resolution: {integrity: sha512-DPi0FmjiSU5EvQV0++GFDOJ9ASQUVFh5kD+OzOnYdi7n3Wpm9hWWGfB/O2blfHcMVTL5WkQXSnRiK9makhrcnw==}
engines: {node: '>=0.3.1'}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
diff@9.0.0:
resolution: {integrity: sha512-svtcdpS8CgJyqAjEQIXdb3OjhFVVYjzGAPO8WGCmRbrml64SPw/jJD4GoE98aR7r25A0XcgrK3F02yw9R/vhQw==}
engines: {node: '>=0.3.1'}
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
docker-modem@5.0.6:
resolution: {integrity: sha512-ens7BiayssQz/uAxGzH8zGXCtiV24rRWXdjNha5V4zSOcxmAZsfGVm/PPFbwQdqEkDnhG+SyR9E3zSHUbOKXBQ==}
engines: {node: '>= 8.0'}
dockerode@4.0.9:
resolution: {integrity: sha512-iND4mcOWhPaCNh54WmK/KoSb35AFqPAUWFMffTQcp52uQt36b5uNwEJTSXntJZBbeGad72Crbi/hvDIv6us/6Q==}
engines: {node: '>= 8.0'}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
dom-accessibility-api@0.5.16:
resolution: {integrity: sha512-X7BJ2yElsnOJ30pZF4uIIDfBEVgF4XEBxL9Bxhy6dnrm5hkzqmsWHGTiHqRiITNhMyFLyAiWndIJP7Z1NTteDg==}
dom-accessibility-api@0.6.3:
resolution: {integrity: sha512-7ZgogeTnjuHbo+ct10G9Ffp0mif17idi0IyWNVA/wcwcm7NPOD/WEHVP3n7n3MhXqxoIYm8d6MuZohYWIZ4T3w==}
dompurify@3.2.7:
resolution: {integrity: sha512-WhL/YuveyGXJaerVlMYGWhvQswa7myDG17P7Vu65EWC05o8vfeNbvNf4d/BOvH99+ZW+LlQsc1GDKMa1vNK6dw==}
first commit
2026-02-21 03:10:39 +00:00
dotenv@16.6.1:
resolution: {integrity: sha512-uBq4egWHTcTt33a72vpSG0z3HnPuIl6NqYcTrKEg2azoEyl2hpW0zqlxysq2pK9HlDIHyHyakeYaYnSAwd8bow==}
engines: {node: '>=12'}
dunder-proto@1.0.1:
resolution: {integrity: sha512-KIN/nDJBQRcXw0MLVhZE9iQHmG68qAVIBg9CqmUYjmQIhgij9U5MFvrqkUL5FbtyyzZuOeOt0zdeRe4UY7ct+A==}
engines: {node: '>= 0.4'}
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
ecdsa-sig-formatter@1.0.11:
resolution: {integrity: sha512-nagl3RYrbNv6kQkeJIpt6NJZy8twLB/2vtz6yN9Z4vRKHN4/QZJIEbqohALSgwKdnksuY3k5Addp5lg8sVoVcQ==}
first commit
2026-02-21 03:10:39 +00:00
ee-first@1.1.1:
resolution: {integrity: sha512-WMwm9LhRUo+WUaRN+vRuETqG89IgZphVSNkdFgeb6sS/E4OrDIN7t48CAewSHXc6C8lefD8KKfr5vY61brQlow==}
effect@3.18.4:
resolution: {integrity: sha512-b1LXQJLe9D11wfnOKAk3PKxuqYshQ0Heez+y5pnkd3jLj1yx9QhM72zZ9uUrOQyNvrs2GZZd/3maL0ZV18YuDA==}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
effect@4.0.0-beta.83:
resolution: {integrity: sha512-0wsak8RtgGAr9UWSbVDgJHZcUqMSvicHcvaZv1MbMM7MCGgW4Rn/137J1MHQbwYPcwYGxT/IqehFd+UbYuj78w==}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
electron-to-chromium@1.5.344:
resolution: {integrity: sha512-4MxfbmNDm+KPh066EZy+eUnkcDPcZ35wNmOWzFuh/ijvHsve6kbLTLURy88uCNK5FbpN+yk2nQY6BYh1GEt+wg==}
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
emoji-regex@10.6.0:
resolution: {integrity: sha512-toUI84YS5YmxW219erniWD0CIVOo46xGKColeNQRgOzDorgBi1v4D71/OFzgD9GO2UGKIv1C3Sp8DAn0+j5w7A==}
first commit
2026-02-21 03:10:39 +00:00
emoji-regex@8.0.0:
resolution: {integrity: sha512-MSjYzcWNOA0ewAHpz0MxpYFvwg6yjy1NG3xteoqz644VCo/RPgnr1/GGt+ic3iJTzQ8Eu3TdM14SawnVUmGE6A==}
empathic@2.0.0:
resolution: {integrity: sha512-i6UzDscO/XfAcNYD75CfICkmfLedpyPDdozrLMmQc5ORaQcdMoc21OnlEylMIqI7U8eniKrPMxxtj8k0vhmJhA==}
engines: {node: '>=14'}
encodeurl@2.0.0:
resolution: {integrity: sha512-Q0n9HRi4m6JuGIV1eFlmvJB7ZEVxu93IrMyiMsGC0lrMJMWzRgx6WGquyfQgZVb31vhGgXnfmPNNXmxnOkRBrg==}
engines: {node: '>= 0.8'}
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
end-of-stream@1.4.5:
resolution: {integrity: sha512-ooEGc6HP26xXq/N+GCGOT0JKCLDGrq2bQUZrQ7gyrJiZANJ/8YDTxTpQBXGMn+WbIQXNVpyWymm7KYVICQnyOg==}
feat(web): bold redesign — Tailwind v4 + shadcn-style primitives + Skills/Proposals/Revisions UI Phase 6 of the Skills + Revisions + Proposals work. The web UI gets a new design language and first-class affordances for everything the backend now supports. ## Visual direction - Tailwind v4 with custom @theme block (oklch tokens). Dark-mode-only (internal tool — light mode doubles QA surface). - Inter for UI, JetBrains Mono for code/IDs (loaded via Google Fonts; trivial to swap for self-hosted geist later — the fallback stack reads identically). - Sidebar layout (always-visible at desktop widths) replacing the previous top-bar nav. Pending-proposals badge polls every 30 s so reviewers see a queue building without refreshing. - Lucide icons throughout. - Spacing and radii on Tailwind defaults. Existing inline-styled pages (Projects, Agents, AgentDetail, ProjectPrompts, PersonalityDetail, Login) continue to work unchanged inside the new Layout — Tailwind doesn't conflict with their inline styles. A follow-up can migrate them incrementally. ## What's added ### Build infra (src/web/) - package.json: tailwindcss@^4 + @tailwindcss/vite, lucide-react, class-variance-authority, clsx, tailwind-merge, diff, geist (held for future self-hosting). - vite.config.ts: registers the @tailwindcss/vite plugin. - src/index.css: Tailwind import + @theme tokens + @layer base. - src/main.tsx: imports index.css. - src/lib/utils.ts: shadcn-style cn() helper. ### shadcn-style primitives (src/components/ui/) Hand-written rather than generated via `npx shadcn` so the repo doesn't depend on a CLI tool that needs interactive runtime: - button.tsx — variants: primary / secondary / ghost / danger / link; sizes: sm / md / lg / icon. - card.tsx — Card + Header/Title/Description/Content/Footer subparts. - badge.tsx — variants: default / info / success / warning / danger / outline. - input.tsx — Input + Textarea + Label. - tabs.tsx — no-dep accessible Tabs (no Radix needed for our use). - separator.tsx — h/v separator with role=separator. ### Diff component (src/components/Diff.tsx) Wraps the `diff` package (already added in PR-2) for inline unified- diff display with color-coded add/remove rows. Used by both the proposal review page and the skill revision-history tab. ### New pages (src/pages/) - Dashboard.tsx — at-a-glance home. Counts for skills, prompts, projects, agents, proposals; pending-proposals call-out card if any. - Skills.tsx — list view, separated into Global vs Project/Agent- scoped sections. - SkillDetail.tsx — name + semver + description; tabs for SKILL.md / Files / Metadata / History. History tab shows revisions with click-to-diff against the live body. - Proposals.tsx — queue with Pending/Approved/Rejected tabs. Pending count is highlighted in amber. - ProposalDetail.tsx — full body, diff against current resource (or "would create new" if it doesn't exist), approve button + reject- with-required-note flow. ### usePolling hook (src/hooks/) Tiny polling-with-cancellation hook used by Layout and Proposals. ### Layout rewrite (src/components/Layout.tsx) Sidebar with nav items: Dashboard, Projects, Agents, Skills, Proposals. Lucide icons. Active-route highlighting via NavLink. Pending-proposals warning badge on the Proposals item. ### Routes (src/App.tsx) New routes: /dashboard, /skills, /skills/:name, /proposals, /proposals/:id. Default redirects to /dashboard. ### API types (src/api.ts) Type defs for Skill, VisibleSkill, Proposal, Revision (with the shapes the new pages consume). ## Tests Existing 7 web tests still pass (Login + api). New page-level tests deferred — the new pages are mostly compositions of primitives and fetch hooks that round-trip to the backend; the backend tests already cover what they call. PR-7 polish can add render-and-click tests if coverage drift surfaces. ## Verification - `pnpm --filter @mcpctl/web build` clean, no warnings. - `pnpm test:run` whole monorepo: 162 test files / 2157 tests green. - Visual smoke deferred — needs a running mcpd to populate the fixtures. Manual smoke tested locally is the next step. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-07 17:54:55 +01:00
enhanced-resolve@5.21.0:
resolution: {integrity: sha512-otxSQPw4lkOZWkHpB3zaEQs6gWYEsmX4xQF68ElXC/TWvGxGMSGOvoNbaLXm6/cS/fSfHtsEdw90y20PCd+sCA==}
engines: {node: '>=10.13.0'}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
entities@6.0.1:
resolution: {integrity: sha512-aN97NXWF6AWBTahfVOIrB/NShkzi5H7F9r1s9mD3cDj4Ko5f2qhhVoYMibXF7GlLveb/D2ioWay8lxI97Ven3g==}
engines: {node: '>=0.12'}
entities@7.0.1:
resolution: {integrity: sha512-TWrgLOFUQTH994YUyl1yT4uyavY5nNB5muff+RtWaqNVCAK408b5ZnnbNAUEWLTCpum9w6arT70i1XdQ4UeOPA==}
engines: {node: '>=0.12'}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
entities@8.0.0:
resolution: {integrity: sha512-zwfzJecQ/Uej6tusMqwAqU/6KL2XaB2VZ2Jg54Je6ahNBGNH6Ek6g3jjNCF0fG9EWQKGZNddNjU5F1ZQn/sBnA==}
engines: {node: '>=20.19.0'}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
env-paths@2.2.1:
resolution: {integrity: sha512-+h1lkLKhZMTYjog1VEpJNG7NZJWcuc2DDk/qsqSTRRCOXiLjeQ1d1/udrUGhqMxUgAlwKNZ0cf2uqan5GLuS2A==}
engines: {node: '>=6'}
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
environment@1.1.0:
resolution: {integrity: sha512-xUtoPkMggbz0MPyPiIWr1Kp4aeWJjDZ6SMvURhimjdZgsRuDplF5/s9hcgGhyXMhs+6vpnuoiZ2kFiu3FMnS8Q==}
engines: {node: '>=18'}
first commit
2026-02-21 03:10:39 +00:00
es-define-property@1.0.1:
resolution: {integrity: sha512-e3nRfgfUZ4rNGL232gUgX06QNyyez04KdjFrF+LTRoOXmrOgFKDg4BCdsjW8EnT69eqdYGmRpJwiPVYNrCaW3g==}
engines: {node: '>= 0.4'}
es-errors@1.3.0:
resolution: {integrity: sha512-Zf5H2Kxt2xjTvbJvP2ZWLEICxA6j+hAmMzIlypy4xcBg1vKVnx89Wy0GbS+kf5cwCVFFzdCFh2XSCFNULS6csw==}
engines: {node: '>= 0.4'}
es-module-lexer@1.7.0:
resolution: {integrity: sha512-jEQoCwk8hyb2AZziIOLhDqpm5+2ww5uIE6lkO/6jcOCusfk6LhMHpXXfBLXTZ7Ydyt0j4VoUQv6uGNYbdW+kBA==}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
es-module-lexer@2.3.1:
resolution: {integrity: sha512-shc1dbU90Yl/xq1QrC7QRtfcwURZuVRfPhZbDoldJ1cn1gzDvBaBWlv0eFolj5+0znnPJz5TXLxsN77X/12KTA==}
first commit
2026-02-21 03:10:39 +00:00
es-object-atoms@1.1.1:
resolution: {integrity: sha512-FGgH2h8zKNim9ljj7dankFPcICIK9Cp5bm+c2gQSYePhpaG5+esrLODihIorn+Pe6FGJzWhXQotPv73jTaldXA==}
engines: {node: '>= 0.4'}
feat: add Kubernetes orchestrator for MCP server pod management mcpd can now deploy MCP server instances as Kubernetes pods instead of Docker containers. Set MCPD_ORCHESTRATOR=kubernetes to enable. - Add @kubernetes/client-node with thin wrapper (context enforcement via MCPD_K8S_CONTEXT to prevent multi-cluster mishaps) - Rewrite KubernetesOrchestrator: pod CRUD, pod IP extraction, exec via SPDY (one-shot + interactive), log streaming - Manifest generator: stdin:true for STDIO servers, args (not command) to preserve runner image entrypoint, security hardening - Orchestrator selection in main.ts via MCPD_ORCHESTRATOR env var - 25 unit tests for k8s orchestrator, all 624 tests pass Tested end-to-end on local k3s: - mcpd deployed via Pulumi, creates pods in mcpctl-servers namespace - NetworkPolicy verified: only mcpd can reach MCP server pods - Python runner (uvx) successfully runs aws-documentation-mcp-server Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-08 01:55:13 +01:00
es-set-tostringtag@2.1.0:
resolution: {integrity: sha512-j6vWzfrGVfyXxge+O0x5sh6cvxAog0a/4Rdd2K36zCMV5eJ+/+tOAngRO8cODMNWbVRdVlmGZQL2YS3yR8bIUA==}
engines: {node: '>= 0.4'}
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
es-toolkit@1.44.0:
resolution: {integrity: sha512-6penXeZalaV88MM3cGkFZZfOoLGWshWWfdy0tWw/RlVVyhvMaWSBTOvXNeiW3e5FwdS5ePW0LGEu17zT139ktg==}
first commit
2026-02-21 03:10:39 +00:00
esbuild@0.27.3:
resolution: {integrity: sha512-8VwMnyGCONIs6cWue2IdpHxHnAjzxnw2Zr7MkVxB2vjmQ2ivqGFb4LEG3SMnv0Gb2F/G/2yA8zUaiL1gywDCCg==}
engines: {node: '>=18'}
hasBin: true
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
escalade@3.2.0:
resolution: {integrity: sha512-WUj2qlxaQtO4g6Pq5c29GTcWGDyd8itL8zTlipgECz3JesAiiOKotd8JU6otB3PACgG6xkJUyVhboMS+bje/jA==}
engines: {node: '>=6'}
first commit
2026-02-21 03:10:39 +00:00
escape-html@1.0.3:
resolution: {integrity: sha512-NiSupZ4OeuGwr68lGIeym/ksIZMJodUGOSCZ/FSnTxcrekbvqrgdUxlJOMpijaKZVjAJrWrGs/6Jy8OMuyj9ow==}
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
escape-string-regexp@2.0.0:
resolution: {integrity: sha512-UpzcLCXolUWcNu5HtVMHYdXJjArjsF9C0aNnquZYY4uW/Vu0miy5YoWvbV345HauVvcAUnpRuhMMcqTcGOY2+w==}
engines: {node: '>=8'}
first commit
2026-02-21 03:10:39 +00:00
escape-string-regexp@4.0.0:
resolution: {integrity: sha512-TtpcNJ3XAzx3Gq8sWRzJaVajRs0uVxA2YAkdb1jm2YkPz4G6egUFAyA3n5vtEIZefPk5Wa4UXbKuS5fKkJWdgA==}
engines: {node: '>=10'}
eslint-config-prettier@10.1.8:
resolution: {integrity: sha512-82GZUjRS0p/jganf6q1rEO25VSoHH0hKPCTrgillPjdI/3bgBhAE1QzHrHTizjpRvy6pGAvKjDJtk2pF9NDq8w==}
hasBin: true
peerDependencies:
eslint: '>=7.0.0'
eslint-scope@9.1.1:
resolution: {integrity: sha512-GaUN0sWim5qc8KVErfPBWmc31LEsOkrUJbvJZV+xuL3u2phMUK4HIvXlWAakfC8W4nzlK+chPEAkYOYb5ZScIw==}
engines: {node: ^20.19.0 || ^22.13.0 || >=24}
eslint-visitor-keys@3.4.3:
resolution: {integrity: sha512-wpc+LXeiyiisxPlEkUzU6svyS1frIO3Mgxj1fdy7Pm8Ygzguax2N3Fa/D/ag1WqbOprdI+uY6wMUl8/a2G+iag==}
engines: {node: ^12.22.0 || ^14.17.0 || >=16.0.0}
eslint-visitor-keys@5.0.1:
resolution: {integrity: sha512-tD40eHxA35h0PEIZNeIjkHoDR4YjjJp34biM0mDvplBe//mB+IHCqHDGV7pxF+7MklTvighcCPPZC7ynWyjdTA==}
engines: {node: ^20.19.0 || ^22.13.0 || >=24}
eslint@10.0.1:
resolution: {integrity: sha512-20MV9SUdeN6Jd84xESsKhRly+/vxI+hwvpBMA93s+9dAcjdCuCojn4IqUGS3lvVaqjVYGYHSRMCpeFtF2rQYxQ==}
engines: {node: ^20.19.0 || ^22.13.0 || >=24}
hasBin: true
peerDependencies:
jiti: '*'
peerDependenciesMeta:
jiti:
optional: true
espree@11.1.1:
resolution: {integrity: sha512-AVHPqQoZYc+RUM4/3Ly5udlZY/U4LS8pIG05jEjWM2lQMU/oaZ7qshzAl2YP1tfNmXfftH3ohurfwNAug+MnsQ==}
engines: {node: ^20.19.0 || ^22.13.0 || >=24}
esquery@1.7.0:
resolution: {integrity: sha512-Ap6G0WQwcU/LHsvLwON1fAQX9Zp0A2Y6Y/cJBl9r/JbW90Zyg4/zbG6zzKa2OTALELarYHmKu0GhpM5EO+7T0g==}
engines: {node: '>=0.10'}
esrecurse@4.3.0:
resolution: {integrity: sha512-KmfKL3b6G+RXvP8N1vr3Tq1kL/oCFgn2NYXEtqP8/L3pKapUA4G8cFVaoF3SU323CD4XypR/ffioHmkti6/Tag==}
engines: {node: '>=4.0'}
estraverse@5.3.0:
resolution: {integrity: sha512-MMdARuVEQziNTeJD8DgMqmhwR11BRQ/cBP+pLtYdSTnf3MIO8fFeiINEbX36ZdNlfU/7A9f3gUw49B3oQsvwBA==}
engines: {node: '>=4.0'}
estree-walker@3.0.3:
resolution: {integrity: sha512-7RUKfXgSMMkzt6ZuXmqapOurLGPPfgj6l9uRZ7lRGolvk0y2yocc35LdcxKC5PQZdn2DMqioAQ2NoWcrTKmm6g==}
esutils@2.0.3:
resolution: {integrity: sha512-kVscqXk4OCp68SZ0dkgEKVi6/8ij300KBWTJq32P/dYeWTSwK41WyTxalN1eRmA5Z9UU/LX9D7FWSmV9SAYx6g==}
engines: {node: '>=0.10.0'}
etag@1.8.1:
resolution: {integrity: sha512-aIL5Fx7mawVa300al2BnEE4iNvo1qETxLrPI/o05L7z6go7fCw1J6EQmbK4FmJ2AS7kgVF/KEZWufBfdClMcPg==}
engines: {node: '>= 0.6'}
feat: add Kubernetes orchestrator for MCP server pod management mcpd can now deploy MCP server instances as Kubernetes pods instead of Docker containers. Set MCPD_ORCHESTRATOR=kubernetes to enable. - Add @kubernetes/client-node with thin wrapper (context enforcement via MCPD_K8S_CONTEXT to prevent multi-cluster mishaps) - Rewrite KubernetesOrchestrator: pod CRUD, pod IP extraction, exec via SPDY (one-shot + interactive), log streaming - Manifest generator: stdin:true for STDIO servers, args (not command) to preserve runner image entrypoint, security hardening - Orchestrator selection in main.ts via MCPD_ORCHESTRATOR env var - 25 unit tests for k8s orchestrator, all 624 tests pass Tested end-to-end on local k3s: - mcpd deployed via Pulumi, creates pods in mcpctl-servers namespace - NetworkPolicy verified: only mcpd can reach MCP server pods - Python runner (uvx) successfully runs aws-documentation-mcp-server Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-08 01:55:13 +01:00
events-universal@1.0.1:
resolution: {integrity: sha512-LUd5euvbMLpwOF8m6ivPCbhQeSiYVNb8Vs0fQ8QjXo0JTkEHpz8pxdQf0gStltaPpw0Cca8b39KxvK9cfKRiAw==}
first commit
2026-02-21 03:10:39 +00:00
eventsource-parser@3.0.6:
resolution: {integrity: sha512-Vo1ab+QXPzZ4tCa8SwIHJFaSzy4R6SHf7BY79rFBDf0idraZWAkYrDjDj8uWaSm3S2TK+hJ7/t1CEmZ7jXw+pg==}
engines: {node: '>=18.0.0'}
eventsource@3.0.7:
resolution: {integrity: sha512-CRT1WTyuQoD771GW56XEZFQ/ZoSfWid1alKGDYMmkt2yl8UXrVR4pspqWNEcqKvVIzg6PAltWjxcSSPrboA4iA==}
engines: {node: '>=18.0.0'}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
execa@5.1.1:
resolution: {integrity: sha512-8uSpZZocAZRBAPIEINJj3Lo9HyGitllczc27Eh5YYojjMFMn8yHMDMaUHE2Jqfq05D/wucwI4JGURyXt1vchyg==}
engines: {node: '>=10'}
first commit
2026-02-21 03:10:39 +00:00
expect-type@1.3.0:
resolution: {integrity: sha512-knvyeauYhqjOYvQ66MznSMs83wmHrCycNEN6Ao+2AeYEfxUIkuiVxdEa1qlGEPK+We3n0THiDciYSsCcgW/DoA==}
engines: {node: '>=12.0.0'}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
exponential-backoff@3.1.3:
resolution: {integrity: sha512-ZgEeZXj30q+I0EN+CbSSpIyPaJ5HVQD18Z1m+u1FXbAeT94mr1zw50q4q6jiiC447Nl/YTcIYSAftiGqetwXCA==}
first commit
2026-02-21 03:10:39 +00:00
express-rate-limit@8.2.1:
resolution: {integrity: sha512-PCZEIEIxqwhzw4KF0n7QF4QqruVTcF73O5kFKUnGOyjbCCgizBBiFaYpd/fnBLUMPw/BWw9OsiN7GgrNYr7j6g==}
engines: {node: '>= 16'}
peerDependencies:
express: '>= 4.11'
express@5.2.1:
resolution: {integrity: sha512-hIS4idWWai69NezIdRt2xFVofaF4j+6INOpJlVOLDO8zXGpUVEVzIYk12UUi2JzjEzWL3IOAxcTubgz9Po0yXw==}
engines: {node: '>= 18'}
exsolve@1.0.8:
resolution: {integrity: sha512-LmDxfWXwcTArk8fUEnOfSZpHOJ6zOMUJKOtFLFqJLoKJetuQG874Uc7/Kki7zFLzYybmZhp1M7+98pfMqeX8yA==}
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
extend@3.0.2:
resolution: {integrity: sha512-fjquC59cD7CyW6urNXK0FBufkZcoiGG80wTuPujX590cB5Ttln20E2UB4S/WARVqhXffZl2LNgS+gQdPIIim/g==}
first commit
2026-02-21 03:10:39 +00:00
fast-check@3.23.2:
resolution: {integrity: sha512-h5+1OzzfCC3Ef7VbtKdcv7zsstUQwUDlYpUTvjeUsJAssPgLn7QzbboPtL5ro04Mq0rPOsMzl7q5hIbRs2wD1A==}
engines: {node: '>=8.0.0'}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
fast-check@4.9.0:
resolution: {integrity: sha512-7ms6T7SybUev/PQITciI0yLM2pOSFy5zpG8Ty7tQofcVaQUvrMXp6CBwqF6fThLCLOrfBtuHAtwq6Yu4XPCllg==}
engines: {node: '>=12.17.0'}
first commit
2026-02-21 03:10:39 +00:00
fast-decode-uri-component@1.0.1:
resolution: {integrity: sha512-WKgKWg5eUxvRZGwW8FvfbaH7AXSh2cL+3j5fMGzUMCxWBJ3dV3a7Wz8y2f/uQ0e3B6WmodD3oS54jTQ9HVTIIg==}
fast-deep-equal@3.1.3:
resolution: {integrity: sha512-f3qQ9oQy9j2AhBe/H9VC91wLmKBCCU/gDOnKNAYG5hswO7BLKj09Hc5HYNz9cGI++xlpDCIgDaitVs03ATR84Q==}
feat: add Kubernetes orchestrator for MCP server pod management mcpd can now deploy MCP server instances as Kubernetes pods instead of Docker containers. Set MCPD_ORCHESTRATOR=kubernetes to enable. - Add @kubernetes/client-node with thin wrapper (context enforcement via MCPD_K8S_CONTEXT to prevent multi-cluster mishaps) - Rewrite KubernetesOrchestrator: pod CRUD, pod IP extraction, exec via SPDY (one-shot + interactive), log streaming - Manifest generator: stdin:true for STDIO servers, args (not command) to preserve runner image entrypoint, security hardening - Orchestrator selection in main.ts via MCPD_ORCHESTRATOR env var - 25 unit tests for k8s orchestrator, all 624 tests pass Tested end-to-end on local k3s: - mcpd deployed via Pulumi, creates pods in mcpctl-servers namespace - NetworkPolicy verified: only mcpd can reach MCP server pods - Python runner (uvx) successfully runs aws-documentation-mcp-server Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-08 01:55:13 +01:00
fast-fifo@1.3.2:
resolution: {integrity: sha512-/d9sfos4yxzpwkDkuN7k2SqFKtYNmCTzgfEpz82x34IM9/zc8KGxQoXg1liNC/izpRM/MBdt44Nmx41ZWqk+FQ==}
first commit
2026-02-21 03:10:39 +00:00
fast-json-stable-stringify@2.1.0:
resolution: {integrity: sha512-lhd/wF+Lk98HZoTCtlVraHtfh5XYijIjalXck7saUtuanSDyLMxnHhSXEDJqHxD7msR8D0uCmqlkwjCV8xvwHw==}
fast-json-stringify@6.3.0:
resolution: {integrity: sha512-oRCntNDY/329HJPlmdNLIdogNtt6Vyjb1WuT01Soss3slIdyUp8kAcDU3saQTOquEK8KFVfwIIF7FebxUAu+yA==}
fast-levenshtein@2.0.6:
resolution: {integrity: sha512-DCXu6Ifhqcks7TZKY3Hxp3y6qphY5SJZmrWMDrKcERSOXWQdMhU9Ig/PYrzyw/ul9jOIyh0N4M0tbC5hodg8dw==}
fast-querystring@1.1.2:
resolution: {integrity: sha512-g6KuKWmFXc0fID8WWH0jit4g0AGBoJhCkJMb1RmbsSEUNvQ+ZC8D6CUZ+GtF8nMzSPXnhiePyyqqipzNNEnHjg==}
fast-uri@3.1.0:
resolution: {integrity: sha512-iPeeDKJSWf4IEOasVVrknXpaBV0IApz/gp7S2bb7Z4Lljbl2MGJRqInZiUrQwV16cpzw/D3S5j5Julj/gT52AA==}
fastify-plugin@5.1.0:
resolution: {integrity: sha512-FAIDA8eovSt5qcDgcBvDuX/v0Cjz0ohGhENZ/wpc3y+oZCY2afZ9Baqql3g/lC+OHRnciQol4ww7tuthOb9idw==}
fastify@5.7.4:
resolution: {integrity: sha512-e6l5NsRdaEP8rdD8VR0ErJASeyaRbzXYpmkrpr2SuvuMq6Si3lvsaVy5C+7gLanEkvjpMDzBXWE5HPeb/hgTxA==}
fastq@1.20.1:
resolution: {integrity: sha512-GGToxJ/w1x32s/D2EKND7kTil4n8OVk/9mycTc4VDza13lOvpUZTGX3mFSCtV9ksdGBVzvsyAVLM6mHFThxXxw==}
fdir@6.5.0:
resolution: {integrity: sha512-tIbYtZbucOs0BRGqPJkshJUYdL+SDH7dVM8gjy+ERp3WAUjLEFJE+02kanyHtwjWOnwrKYBiwAmM0p4kLJAnXg==}
engines: {node: '>=12.0.0'}
peerDependencies:
picomatch: ^3 || ^4
peerDependenciesMeta:
picomatch:
optional: true
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
fetch-blob@3.2.0:
resolution: {integrity: sha512-7yAQpD2UMJzLi1Dqv7qFYnPbaPx7ZfFK6PiIxQ4PfkGPyNyl2Ugx+a/umUonmKqjhM4DnfbMvdX6otXq83soQQ==}
engines: {node: ^12.20 || >= 14.13}
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
figures@6.1.0:
resolution: {integrity: sha512-d+l3qxjSesT4V7v2fh+QnmFnUWv9lSpjarhShNTgBOfA0ttejbQUAlHLitbjkoRiDulW0OPoQPYIGhIC8ohejg==}
engines: {node: '>=18'}
first commit
2026-02-21 03:10:39 +00:00
file-entry-cache@8.0.0:
resolution: {integrity: sha512-XXTUwCvisa5oacNGRP9SfNtYBNAMi+RPwBFmblZEF7N7swHYQS6/Zfk7SRwx4D5j3CH211YNRco1DEMNVfZCnQ==}
engines: {node: '>=16.0.0'}
finalhandler@2.1.1:
resolution: {integrity: sha512-S8KoZgRZN+a5rNwqTxlZZePjT/4cnm0ROV70LedRHZ0p8u9fRID0hJUZQpkKLzro8LfmC8sx23bY6tVNxv8pQA==}
engines: {node: '>= 18.0.0'}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
find-babel-config@2.1.2:
resolution: {integrity: sha512-ZfZp1rQyp4gyuxqt1ZqjFGVeVBvmpURMqdIWXbPRfB97Bf6BzdK/xSIbylEINzQ0kB5tlDQfn9HkNXXWsqTqLg==}
find-my-way-ts@0.1.6:
resolution: {integrity: sha512-a85L9ZoXtNAey3Y6Z+eBWW658kO/MwR7zIafkIUPUMf3isZG0NCs2pjW2wtjxAKuJPxMAsHUIP4ZPGv0o5gyTA==}
first commit
2026-02-21 03:10:39 +00:00
find-my-way@9.4.0:
resolution: {integrity: sha512-5Ye4vHsypZRYtS01ob/iwHzGRUDELlsoCftI/OZFhcLs1M0tkGPcXldE80TAZC5yYuJMBPJQQ43UHlqbJWiX2w==}
engines: {node: '>=20'}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
find-up@3.0.0:
resolution: {integrity: sha512-1yD6RmLI1XBfxugvORwlck6f75tYL+iR0jqwsOrOxMZyGYqUuDhJ0l4AXdO1iX/FTs9cBAMEk1gWSEx1kSbylg==}
engines: {node: '>=6'}
first commit
2026-02-21 03:10:39 +00:00
find-up@5.0.0:
resolution: {integrity: sha512-78/PXT1wlLLDgTzDs7sjq9hzz0vXD+zn+7wypEe4fXQxCmdmqfGsEPQxmiCSQI3ajFV91bVSsvNtrJRiW6nGng==}
engines: {node: '>=10'}
flat-cache@4.0.1:
resolution: {integrity: sha512-f7ccFPK3SXFHpx15UIGyRJ/FJQctuKZ0zVuN3frBo4HnK3cay9VEW0R6yPYFHC0AgqhukPzKjq22t5DmAyqGyw==}
engines: {node: '>=16'}
flatted@3.3.3:
resolution: {integrity: sha512-GX+ysw4PBCz0PzosHDepZGANEuFCMLrnRTiEy9McGjmkCQYwRq4A/X786G/fjM/+OjsWSU1ZrY5qyARZmO/uwg==}
feat(mcpd+deploy): serve web UI at /ui + smoke tests + docs (Stage 6) The closing stage. mcpd now hosts the Stage 5 SPA, the Docker image bundles the build artifact, a smoke test exercises the personality HTTP surface end-to-end, and the user-facing docs spell out the mental model. mcpd: - Add @fastify/static dep. - New routes/web-ui.ts: registers /ui/* against a static bundle. Looks for the bundle at $MCPD_WEB_ROOT, then /usr/share/mcpd/web (the Docker image path), then a dev-tree fallback. Logs and skips cleanly if missing — API-only deploys keep working. - SPA fallback: any /ui/<path> that doesn't match a file falls through to index.html so direct hits to react-router URLs work. - /ui/* falls through to `kind: skip` in mapUrlToPermission, so the static assets are served unauthenticated. Each API call from the SPA still carries the bearer token. Deploy: - Dockerfile.mcpd builds the @mcpctl/web bundle in the same builder stage and copies dist/ to /usr/share/mcpd/web in the runtime image. Smoke (personality.smoke.test.ts): - Live mcpd flow: create secret/llm/agent/personality, attach an agent-direct prompt, verify the binding listing, reject double- attach (409) + foreign-agent prompt (400), set defaultPersonality by name, detach + delete cleanup. Docs: - New docs/personalities.md: VLAN-on-ethernet model, system-block ordering table, three prompt scopes, CLI walkthrough, web UI walkthrough, full API surface, RBAC notes. - agents.md and chat.md cross-link. - README's Agents section gains a Personalities subsection. Test count after Stage 6: mcpd: 801/801 cli: 430/430 web: 7/7 db: 58/62 (4 pre-existing) Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:48:43 +01:00
foreground-child@3.3.1:
resolution: {integrity: sha512-gIXjKqtFuWEgzFRJA9WCQeSJLZDjgJUOMCMzxtvFq/37KojM1BFGufqsCy0r4qSQmYLsZYMeyRqzIWOMup03sw==}
engines: {node: '>=14'}
feat: add Kubernetes orchestrator for MCP server pod management mcpd can now deploy MCP server instances as Kubernetes pods instead of Docker containers. Set MCPD_ORCHESTRATOR=kubernetes to enable. - Add @kubernetes/client-node with thin wrapper (context enforcement via MCPD_K8S_CONTEXT to prevent multi-cluster mishaps) - Rewrite KubernetesOrchestrator: pod CRUD, pod IP extraction, exec via SPDY (one-shot + interactive), log streaming - Manifest generator: stdin:true for STDIO servers, args (not command) to preserve runner image entrypoint, security hardening - Orchestrator selection in main.ts via MCPD_ORCHESTRATOR env var - 25 unit tests for k8s orchestrator, all 624 tests pass Tested end-to-end on local k3s: - mcpd deployed via Pulumi, creates pods in mcpctl-servers namespace - NetworkPolicy verified: only mcpd can reach MCP server pods - Python runner (uvx) successfully runs aws-documentation-mcp-server Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-08 01:55:13 +01:00
form-data@4.0.5:
resolution: {integrity: sha512-8RipRLol37bNs2bhoV67fiTEvdTrbMUYcFTiy3+wuuOnUog2QBHCZWXDRijWQfAkhBj2Uf5UnVaiWwA5vdd82w==}
engines: {node: '>= 6'}
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
formdata-polyfill@4.0.10:
resolution: {integrity: sha512-buewHzMvYL29jdeQTVILecSaZKnt/RJWjoZCF5OW60Z67/GmSLBkOFM7qh1PI3zFNtJbaZL5eQu1vLfazOwj4g==}
engines: {node: '>=12.20.0'}
first commit
2026-02-21 03:10:39 +00:00
forwarded@0.2.0:
resolution: {integrity: sha512-buRG0fpBtRHSTCOASe6hD258tEubFoRLb4ZNA6NxMVHNw2gOcwHo9wyablzMzOA5z9xA9L1KNjk/Nt6MT9aYow==}
engines: {node: '>= 0.6'}
fresh@2.0.0:
resolution: {integrity: sha512-Rx/WycZ60HOaqLKAi6cHRKKI7zxWbJ31MhntmtwMoaTeF7XFH9hhBp8vITaMidfljRQ6eYWCKkaTK+ykVJHP2A==}
engines: {node: '>= 0.8'}
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
fs-constants@1.0.0:
resolution: {integrity: sha512-y6OAwoSIf7FyjMIv94u+b5rdheZEjzR63GTyZJm5qh4Bi+2YgwLCcI/fPFZkL5PSixOt6ZNKm+w+Hfp/Bciwow==}
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
fs-minipass@2.1.0:
resolution: {integrity: sha512-V/JgOLFCS+R6Vcq0slCuaeWEdNC3ouDlJMNIsacH2VtALiu9mV4LPrHc5cDl8k5aw6J8jwgWWpiTo5RYhmIzvg==}
engines: {node: '>= 8'}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
fs-minipass@3.0.3:
resolution: {integrity: sha512-XUBA9XClHbnJWSfBzjkm6RvPsyg3sryZt06BEQoXcF7EK/xpGaQYJgQKDJSUH5SGZ76Y7pFx1QBnXz09rU5Fbw==}
engines: {node: ^14.17.0 || ^16.13.0 || >=18.0.0}
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
fs.realpath@1.0.0:
resolution: {integrity: sha512-OO0pH2lK6a0hZnAdau5ItzHPI6pUlvI7jMVnxUQRtw4owF2wk8lOSabtGDCTP4Ggrg2MbGnWO9X8K1t4+fGMDw==}
first commit
2026-02-21 03:10:39 +00:00
fsevents@2.3.3:
resolution: {integrity: sha512-5xoDfX+fL7faATnagmWPpbFtwh/R77WmMMqqHGS65C3vvB0YHrgF+B1YmZ3441tMj5n63k0212XNoJwzlhffQw==}
engines: {node: ^8.16.0 || ^10.6.0 || >=11.0.0}
os: [darwin]
function-bind@1.1.2:
resolution: {integrity: sha512-7XHNxH7qX9xG5mIwxkhumTox/MIRNcOgDrxWsMt2pAr23WHp6MrRlN7FBSFpCpr+oVO0F744iUgR82nJMfG2SA==}
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
gauge@3.0.2:
resolution: {integrity: sha512-+5J6MS/5XksCuXq++uFRsnUd7Ovu1XenbeuIuNRJxYWjgQbPuFhT14lAvsWfqfAmnwluf1OwMjz39HjfLPci0Q==}
engines: {node: '>=10'}
deprecated: This package is no longer supported.
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
gaxios@7.3.0:
resolution: {integrity: sha512-RB5vLV+vvQeoFPCX4QMK6/hjVkbIamPp1QSUD0CiZcnj12qbpiL+pLbYtgD+oZkWl0tl9z+o2Utp+MpM3QRhBA==}
engines: {node: '>=18'}
gcp-metadata@8.1.2:
resolution: {integrity: sha512-zV/5HKTfCeKWnxG0Dmrw51hEWFGfcF2xiXqcA3+J90WDuP0SvoiSO5ORvcBsifmx/FoIjgQN3oNOGaQ5PhLFkg==}
engines: {node: '>=18'}
feat(web): bold redesign — Tailwind v4 + shadcn-style primitives + Skills/Proposals/Revisions UI Phase 6 of the Skills + Revisions + Proposals work. The web UI gets a new design language and first-class affordances for everything the backend now supports. ## Visual direction - Tailwind v4 with custom @theme block (oklch tokens). Dark-mode-only (internal tool — light mode doubles QA surface). - Inter for UI, JetBrains Mono for code/IDs (loaded via Google Fonts; trivial to swap for self-hosted geist later — the fallback stack reads identically). - Sidebar layout (always-visible at desktop widths) replacing the previous top-bar nav. Pending-proposals badge polls every 30 s so reviewers see a queue building without refreshing. - Lucide icons throughout. - Spacing and radii on Tailwind defaults. Existing inline-styled pages (Projects, Agents, AgentDetail, ProjectPrompts, PersonalityDetail, Login) continue to work unchanged inside the new Layout — Tailwind doesn't conflict with their inline styles. A follow-up can migrate them incrementally. ## What's added ### Build infra (src/web/) - package.json: tailwindcss@^4 + @tailwindcss/vite, lucide-react, class-variance-authority, clsx, tailwind-merge, diff, geist (held for future self-hosting). - vite.config.ts: registers the @tailwindcss/vite plugin. - src/index.css: Tailwind import + @theme tokens + @layer base. - src/main.tsx: imports index.css. - src/lib/utils.ts: shadcn-style cn() helper. ### shadcn-style primitives (src/components/ui/) Hand-written rather than generated via `npx shadcn` so the repo doesn't depend on a CLI tool that needs interactive runtime: - button.tsx — variants: primary / secondary / ghost / danger / link; sizes: sm / md / lg / icon. - card.tsx — Card + Header/Title/Description/Content/Footer subparts. - badge.tsx — variants: default / info / success / warning / danger / outline. - input.tsx — Input + Textarea + Label. - tabs.tsx — no-dep accessible Tabs (no Radix needed for our use). - separator.tsx — h/v separator with role=separator. ### Diff component (src/components/Diff.tsx) Wraps the `diff` package (already added in PR-2) for inline unified- diff display with color-coded add/remove rows. Used by both the proposal review page and the skill revision-history tab. ### New pages (src/pages/) - Dashboard.tsx — at-a-glance home. Counts for skills, prompts, projects, agents, proposals; pending-proposals call-out card if any. - Skills.tsx — list view, separated into Global vs Project/Agent- scoped sections. - SkillDetail.tsx — name + semver + description; tabs for SKILL.md / Files / Metadata / History. History tab shows revisions with click-to-diff against the live body. - Proposals.tsx — queue with Pending/Approved/Rejected tabs. Pending count is highlighted in amber. - ProposalDetail.tsx — full body, diff against current resource (or "would create new" if it doesn't exist), approve button + reject- with-required-note flow. ### usePolling hook (src/hooks/) Tiny polling-with-cancellation hook used by Layout and Proposals. ### Layout rewrite (src/components/Layout.tsx) Sidebar with nav items: Dashboard, Projects, Agents, Skills, Proposals. Lucide icons. Active-route highlighting via NavLink. Pending-proposals warning badge on the Proposals item. ### Routes (src/App.tsx) New routes: /dashboard, /skills, /skills/:name, /proposals, /proposals/:id. Default redirects to /dashboard. ### API types (src/api.ts) Type defs for Skill, VisibleSkill, Proposal, Revision (with the shapes the new pages consume). ## Tests Existing 7 web tests still pass (Login + api). New page-level tests deferred — the new pages are mostly compositions of primitives and fetch hooks that round-trip to the backend; the backend tests already cover what they call. PR-7 polish can add render-and-click tests if coverage drift surfaces. ## Verification - `pnpm --filter @mcpctl/web build` clean, no warnings. - `pnpm test:run` whole monorepo: 162 test files / 2157 tests green. - Visual smoke deferred — needs a running mcpd to populate the fixtures. Manual smoke tested locally is the next step. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-07 17:54:55 +01:00
geist@1.7.0:
resolution: {integrity: sha512-ZaoiZwkSf0DwwB1ncdLKp+ggAldqxl5L1+SXaNIBGkPAqcu+xjVJLxlf3/S8vLt9UHx1xu5fz3lbzKCj5iOVdQ==}
peerDependencies:
next: '>=13.2.0'
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
gensync@1.0.0-beta.2:
resolution: {integrity: sha512-3hN7NaskYvMDLQY55gnW3NQ+mesEAepTqlg+VEbj7zzqEMBVNhzcGYYeqFo/TlYz6eQiFcp1HcsCZO+nGgS8zg==}
engines: {node: '>=6.9.0'}
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
get-caller-file@2.0.5:
resolution: {integrity: sha512-DyFP3BM/3YHTQOCUL/w0OZHR0lpKeGrxotcHWcqNEdnltqFwXVfhEBQ94eIo34AfQpo0rGki4cyIiftY06h2Fg==}
engines: {node: 6.* || 8.* || >= 10.*}
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
get-east-asian-width@1.5.0:
resolution: {integrity: sha512-CQ+bEO+Tva/qlmw24dCejulK5pMzVnUOFOijVogd3KQs07HnRIgp8TGipvCCRT06xeYEbpbgwaCxglFyiuIcmA==}
engines: {node: '>=18'}
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
get-east-asian-width@1.6.0:
resolution: {integrity: sha512-QRbvDIbx6YklUe6RxeTeleMR0yv3cYH6PsPZHcnVn7xv7zO1BHN8r0XETu8n6Ye3Q+ahtSarc3WgtNWmehIBfA==}
engines: {node: '>=18'}
first commit
2026-02-21 03:10:39 +00:00
get-intrinsic@1.3.0:
resolution: {integrity: sha512-9fSjSaos/fRIVIp+xSJlE6lfwhES7LNtKaCBIamHsjr2na1BiABJPo0mOjjz8GJDURarmCPGqaiVg5mfjb98CQ==}
engines: {node: '>= 0.4'}
get-proto@1.0.1:
resolution: {integrity: sha512-sTSfBjoXBp89JvIKIefqw7U2CCebsc74kiY6awiGogKtoSGbgjYE/G/+l9sF3MWFPNc9IcoOC4ODfKHfxFmp0g==}
engines: {node: '>= 0.4'}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
get-stream@6.0.1:
resolution: {integrity: sha512-ts6Wi+2j3jQjqi70w5AlN8DFnkSwC+MqmxEzdEALB2qXZYV3X/b1CTfgPLGJNMeAWxdPfU8FO1ms3NUfaHCPYg==}
engines: {node: '>=10'}
first commit
2026-02-21 03:10:39 +00:00
get-tsconfig@4.13.6:
resolution: {integrity: sha512-shZT/QMiSHc/YBLxxOkMtgSid5HFoauqCE3/exfsEcwg1WkeqjG+V40yBbBrsD+jW2HDXcs28xOfcbm2jI8Ddw==}
giget@2.0.0:
resolution: {integrity: sha512-L5bGsVkxJbJgdnwyuheIunkGatUF/zssUoxxjACCseZYAVbaqdh9Tsmmlkl8vYan09H7sbvKt4pS8GqKLBrEzA==}
hasBin: true
glob-parent@6.0.2:
resolution: {integrity: sha512-XxwI8EOhVQgWp6iDL+3b0r86f4d6AX6zSU55HfB4ydCEuXLXc5FcYeOu+nnGftS4TEju/11rt4KJPTMgbfmv4A==}
engines: {node: '>=10.13.0'}
feat(mcpd+deploy): serve web UI at /ui + smoke tests + docs (Stage 6) The closing stage. mcpd now hosts the Stage 5 SPA, the Docker image bundles the build artifact, a smoke test exercises the personality HTTP surface end-to-end, and the user-facing docs spell out the mental model. mcpd: - Add @fastify/static dep. - New routes/web-ui.ts: registers /ui/* against a static bundle. Looks for the bundle at $MCPD_WEB_ROOT, then /usr/share/mcpd/web (the Docker image path), then a dev-tree fallback. Logs and skips cleanly if missing — API-only deploys keep working. - SPA fallback: any /ui/<path> that doesn't match a file falls through to index.html so direct hits to react-router URLs work. - /ui/* falls through to `kind: skip` in mapUrlToPermission, so the static assets are served unauthenticated. Each API call from the SPA still carries the bearer token. Deploy: - Dockerfile.mcpd builds the @mcpctl/web bundle in the same builder stage and copies dist/ to /usr/share/mcpd/web in the runtime image. Smoke (personality.smoke.test.ts): - Live mcpd flow: create secret/llm/agent/personality, attach an agent-direct prompt, verify the binding listing, reject double- attach (409) + foreign-agent prompt (400), set defaultPersonality by name, detach + delete cleanup. Docs: - New docs/personalities.md: VLAN-on-ethernet model, system-block ordering table, three prompt scopes, CLI walkthrough, web UI walkthrough, full API surface, RBAC notes. - agents.md and chat.md cross-link. - README's Agents section gains a Personalities subsection. Test count after Stage 6: mcpd: 801/801 cli: 430/430 web: 7/7 db: 58/62 (4 pre-existing) Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:48:43 +01:00
glob@11.1.0:
resolution: {integrity: sha512-vuNwKSaKiqm7g0THUBu2x7ckSs3XJLXE+2ssL7/MfTGPLLcrJQ/4Uq1CjPTtO5cCIiRxqvN6Twy1qOwhL0Xjcw==}
engines: {node: 20 || >=22}
deprecated: Old versions of glob are not supported, and contain widely publicized security vulnerabilities, which have been fixed in the current version. Please update. Support for old versions may be purchased (at exorbitant rates) by contacting i@izs.me
hasBin: true
first commit
2026-02-21 03:10:39 +00:00
glob@13.0.6:
resolution: {integrity: sha512-Wjlyrolmm8uDpm/ogGyXZXb1Z+Ca2B8NbJwqBVg0axK9GbBeoS7yGV6vjXnYdGm6X53iehEuxxbyiKp8QmN4Vw==}
engines: {node: 18 || 20 || >=22}
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
glob@7.2.3:
resolution: {integrity: sha512-nFR0zLpU2YCaRxwoCJvL6UvCH2JFyFVIvwTLsIf21AuHlMskA1hhTdk+LlYJtOlYt9v6dvszD2BGRqBL+iQK9Q==}
deprecated: Old versions of glob are not supported, and contain widely publicized security vulnerabilities, which have been fixed in the current version. Please update. Support for old versions may be purchased (at exorbitant rates) by contacting i@izs.me
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
glob@9.3.5:
resolution: {integrity: sha512-e1LleDykUz2Iu+MTYdkSsuWX8lvAjAcs0Xef0lNIu0S2wOAzuTxCJtcd9S3cijlwYF18EsU3rzb8jPVobxDh9Q==}
engines: {node: '>=16 || 14 >=14.17'}
deprecated: Old versions of glob are not supported, and contain widely publicized security vulnerabilities, which have been fixed in the current version. Please update. Support for old versions may be purchased (at exorbitant rates) by contacting i@izs.me
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
google-auth-library@10.9.1:
resolution: {integrity: sha512-i1ydyHrqcIxXkWh/uBmVkzCvIuq5yiK2ATndIe5XxKholrG/MTYP9xGYka4sQhrbIAgGjL2B6NOE7rFaiF3fXw==}
engines: {node: '>=18'}
google-logging-utils@1.1.3:
resolution: {integrity: sha512-eAmLkjDjAFCVXg7A1unxHsLf961m6y17QFqXqAXGj/gVkKFrEICfStRfwUlGNfeCEjNRa32JEWOUTlYXPyyKvA==}
engines: {node: '>=14'}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
google-protobuf@3.21.4:
resolution: {integrity: sha512-MnG7N936zcKTco4Jd2PX2U96Kf9PxygAPKBug+74LHzmHXmceN16MmRcdgZv+DGef/S9YvQAfRsNCn4cjf9yyQ==}
first commit
2026-02-21 03:10:39 +00:00
gopd@1.2.0:
resolution: {integrity: sha512-ZUKRh6/kUFoAiTAtTYPZJ3hw9wNxx+BIBOijnlG9PnrJsCcSjs1wyyD6vJpaYtgnzDrKYRSqf3OO6Rfa93xsRg==}
engines: {node: '>= 0.4'}
feat(web): bold redesign — Tailwind v4 + shadcn-style primitives + Skills/Proposals/Revisions UI Phase 6 of the Skills + Revisions + Proposals work. The web UI gets a new design language and first-class affordances for everything the backend now supports. ## Visual direction - Tailwind v4 with custom @theme block (oklch tokens). Dark-mode-only (internal tool — light mode doubles QA surface). - Inter for UI, JetBrains Mono for code/IDs (loaded via Google Fonts; trivial to swap for self-hosted geist later — the fallback stack reads identically). - Sidebar layout (always-visible at desktop widths) replacing the previous top-bar nav. Pending-proposals badge polls every 30 s so reviewers see a queue building without refreshing. - Lucide icons throughout. - Spacing and radii on Tailwind defaults. Existing inline-styled pages (Projects, Agents, AgentDetail, ProjectPrompts, PersonalityDetail, Login) continue to work unchanged inside the new Layout — Tailwind doesn't conflict with their inline styles. A follow-up can migrate them incrementally. ## What's added ### Build infra (src/web/) - package.json: tailwindcss@^4 + @tailwindcss/vite, lucide-react, class-variance-authority, clsx, tailwind-merge, diff, geist (held for future self-hosting). - vite.config.ts: registers the @tailwindcss/vite plugin. - src/index.css: Tailwind import + @theme tokens + @layer base. - src/main.tsx: imports index.css. - src/lib/utils.ts: shadcn-style cn() helper. ### shadcn-style primitives (src/components/ui/) Hand-written rather than generated via `npx shadcn` so the repo doesn't depend on a CLI tool that needs interactive runtime: - button.tsx — variants: primary / secondary / ghost / danger / link; sizes: sm / md / lg / icon. - card.tsx — Card + Header/Title/Description/Content/Footer subparts. - badge.tsx — variants: default / info / success / warning / danger / outline. - input.tsx — Input + Textarea + Label. - tabs.tsx — no-dep accessible Tabs (no Radix needed for our use). - separator.tsx — h/v separator with role=separator. ### Diff component (src/components/Diff.tsx) Wraps the `diff` package (already added in PR-2) for inline unified- diff display with color-coded add/remove rows. Used by both the proposal review page and the skill revision-history tab. ### New pages (src/pages/) - Dashboard.tsx — at-a-glance home. Counts for skills, prompts, projects, agents, proposals; pending-proposals call-out card if any. - Skills.tsx — list view, separated into Global vs Project/Agent- scoped sections. - SkillDetail.tsx — name + semver + description; tabs for SKILL.md / Files / Metadata / History. History tab shows revisions with click-to-diff against the live body. - Proposals.tsx — queue with Pending/Approved/Rejected tabs. Pending count is highlighted in amber. - ProposalDetail.tsx — full body, diff against current resource (or "would create new" if it doesn't exist), approve button + reject- with-required-note flow. ### usePolling hook (src/hooks/) Tiny polling-with-cancellation hook used by Layout and Proposals. ### Layout rewrite (src/components/Layout.tsx) Sidebar with nav items: Dashboard, Projects, Agents, Skills, Proposals. Lucide icons. Active-route highlighting via NavLink. Pending-proposals warning badge on the Proposals item. ### Routes (src/App.tsx) New routes: /dashboard, /skills, /skills/:name, /proposals, /proposals/:id. Default redirects to /dashboard. ### API types (src/api.ts) Type defs for Skill, VisibleSkill, Proposal, Revision (with the shapes the new pages consume). ## Tests Existing 7 web tests still pass (Login + api). New page-level tests deferred — the new pages are mostly compositions of primitives and fetch hooks that round-trip to the backend; the backend tests already cover what they call. PR-7 polish can add render-and-click tests if coverage drift surfaces. ## Verification - `pnpm --filter @mcpctl/web build` clean, no warnings. - `pnpm test:run` whole monorepo: 162 test files / 2157 tests green. - Visual smoke deferred — needs a running mcpd to populate the fixtures. Manual smoke tested locally is the next step. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-07 17:54:55 +01:00
graceful-fs@4.2.11:
resolution: {integrity: sha512-RbJ5/jmFcNNCcDV5o9eTnBLJ/HszWV0P73bc+Ff4nS/rJj+YaS6IGyiOL0VoBYX+l1Wrl3k63h/KrH+nhJ0XvQ==}
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
grok-mermaid@0.2.2:
resolution: {integrity: sha512-XcJEP5dDC8liHBh52mlLjU18fNvu1ckFsu0QpIG3+APZ270fsj9wxpiA6cOURmbUEuoMVgjbC2+UYgTdCqqgzA==}
engines: {node: '>=18'}
first commit
2026-02-21 03:10:39 +00:00
has-flag@4.0.0:
resolution: {integrity: sha512-EykJT/Q1KjTWctppgIAgfSO0tKVuZUjhgMr17kqTumMl6Afv3EISleU7qZUzoXDFTAHTDC4NOoG/ZxU3EvlMPQ==}
engines: {node: '>=8'}
has-symbols@1.1.0:
resolution: {integrity: sha512-1cDNdwJ2Jaohmb3sg4OmKaMBwuC48sYni5HUw2DvsC8LjGTLK9h+eb1X6RyuOHe4hT0ULCW68iomhjUoKUqlPQ==}
engines: {node: '>= 0.4'}
feat: add Kubernetes orchestrator for MCP server pod management mcpd can now deploy MCP server instances as Kubernetes pods instead of Docker containers. Set MCPD_ORCHESTRATOR=kubernetes to enable. - Add @kubernetes/client-node with thin wrapper (context enforcement via MCPD_K8S_CONTEXT to prevent multi-cluster mishaps) - Rewrite KubernetesOrchestrator: pod CRUD, pod IP extraction, exec via SPDY (one-shot + interactive), log streaming - Manifest generator: stdin:true for STDIO servers, args (not command) to preserve runner image entrypoint, security hardening - Orchestrator selection in main.ts via MCPD_ORCHESTRATOR env var - 25 unit tests for k8s orchestrator, all 624 tests pass Tested end-to-end on local k3s: - mcpd deployed via Pulumi, creates pods in mcpctl-servers namespace - NetworkPolicy verified: only mcpd can reach MCP server pods - Python runner (uvx) successfully runs aws-documentation-mcp-server Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-08 01:55:13 +01:00
has-tostringtag@1.0.2:
resolution: {integrity: sha512-NqADB8VjPFLM2V0VvHUewwwsw0ZWBaIdgo+ieHtK3hasLz4qeCRjYcqfB6AQrBggRKppKF8L52/VqdVsO47Dlw==}
engines: {node: '>= 0.4'}
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
has-unicode@2.0.1:
resolution: {integrity: sha512-8Rf9Y83NBReMnx0gFzA8JImQACstCYWUplepDa9xprwwtmgEZUF0h/i5xSA625zB/I37EtrswSST6OXxwaaIJQ==}
first commit
2026-02-21 03:10:39 +00:00
hasown@2.0.2:
resolution: {integrity: sha512-0hJU9SCPvmMzIBdZFqNPXWa6dqh7WdH0cII9y+CyS8rG3nL48Bclra9HmKhVVUHyPWNH5Y7xDwAB7bfgSjkUMQ==}
engines: {node: '>= 0.4'}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
hasown@2.0.4:
resolution: {integrity: sha512-T2UbfbBEF32wiepXIsMlTW9+dDYC6wMh/t/vYA4tuOMKqWz/n3vr1NFSxQiyP+zk2mXsoMA/i/7qV6LKut1t1A==}
engines: {node: '>= 0.4'}
first commit
2026-02-21 03:10:39 +00:00
helmet@7.2.0:
resolution: {integrity: sha512-ZRiwvN089JfMXokizgqEPXsl2Guk094yExfoDXR0cBYWxtBbaSww/w+vT4WEJsBW2iTUi1GgZ6swmoug3Oy4Xw==}
engines: {node: '>=16.0.0'}
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
highlight.js@10.7.3:
resolution: {integrity: sha512-tzcUFauisWKNHaRkN4Wjl/ZA07gENAjFl3J/c480dprkGTg5EQstgaNFqBfUqCq54kZRIEcreTsAgF/m2quD7A==}
first commit
2026-02-21 03:10:39 +00:00
hono@4.12.0:
resolution: {integrity: sha512-NekXntS5M94pUfiVZ8oXXK/kkri+5WpX2/Ik+LVsl+uvw+soj4roXIsPqO+XsWrAw20mOzaXOZf3Q7PfB9A/IA==}
engines: {node: '>=16.9.0'}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
hosted-git-info@7.0.2:
resolution: {integrity: sha512-puUZAUKT5m8Zzvs72XWy3HtvVbTWljRE66cP60bxJzAqf2DgICo7lYTY2IHUmLnNpjYvw5bvmoHvPc0QO2a62w==}
engines: {node: ^16.14.0 || >=18.0.0}
hosted-git-info@9.0.3:
resolution: {integrity: sha512-Hc+ghLoSt6QaYZUv0WBiIvmMDZuZZ7oaDvdH8MbfOO4lOsxdXLEvuC6ePoGs9H1X9oCLyq6+NVN0MKqD+ydxyg==}
engines: {node: ^20.17.0 || >=22.9.0}
feat: add Kubernetes orchestrator for MCP server pod management mcpd can now deploy MCP server instances as Kubernetes pods instead of Docker containers. Set MCPD_ORCHESTRATOR=kubernetes to enable. - Add @kubernetes/client-node with thin wrapper (context enforcement via MCPD_K8S_CONTEXT to prevent multi-cluster mishaps) - Rewrite KubernetesOrchestrator: pod CRUD, pod IP extraction, exec via SPDY (one-shot + interactive), log streaming - Manifest generator: stdin:true for STDIO servers, args (not command) to preserve runner image entrypoint, security hardening - Orchestrator selection in main.ts via MCPD_ORCHESTRATOR env var - 25 unit tests for k8s orchestrator, all 624 tests pass Tested end-to-end on local k3s: - mcpd deployed via Pulumi, creates pods in mcpctl-servers namespace - NetworkPolicy verified: only mcpd can reach MCP server pods - Python runner (uvx) successfully runs aws-documentation-mcp-server Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-08 01:55:13 +01:00
hpagent@1.2.0:
resolution: {integrity: sha512-A91dYTeIB6NoXG+PxTQpCCDDnfHsW9kc06Lvpu1TEe9gnd6ZFeiBoRO9JvzEv6xK7EX97/dUE8g/vBMTqTS3CA==}
engines: {node: '>=14'}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
html-encoding-sniffer@6.0.0:
resolution: {integrity: sha512-CV9TW3Y3f8/wT0BRFc1/KAVQ3TUHiXmaAb6VW9vtiMFf7SLoMd1PdAc4W3KFOFETBJUb90KatHqlsZMWV+R9Gg==}
engines: {node: ^20.19.0 || ^22.12.0 || >=24.0.0}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
html-entities@2.3.3:
resolution: {integrity: sha512-DV5Ln36z34NNTDgnz0EWGBLZENelNAtkiFA4kyNOG2tDI6Mz1uSWiq1wAKdyjnJwyDiDO7Fa2SO1CTxPXL8VxA==}
first commit
2026-02-21 03:10:39 +00:00
html-escaper@2.0.2:
resolution: {integrity: sha512-H2iMtd0I4Mt5eYiapRdIDjp+XzelXQ0tFE4JS7YFwFevXXMmOp9myNrUvCg0D6ws8iqkRPBfKHgbwig1SmlLfg==}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
http-cache-semantics@4.2.0:
resolution: {integrity: sha512-dTxcvPXqPvXBQpq5dUr6mEMJX4oIEFv6bwom3FDwKRDsuIjjJGANqhBuoAn9c1RQJIdAKav33ED65E2ys+87QQ==}
first commit
2026-02-21 03:10:39 +00:00
http-errors@2.0.1:
resolution: {integrity: sha512-4FbRdAX+bSdmo4AUFuS0WNiPz8NgFt+r8ThgNWmlrjQjt1Q7ZR9+zTlce2859x4KSXrwIsaeTqDoKQmtP8pLmQ==}
engines: {node: '>= 0.8'}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
http-proxy-agent@7.0.2:
resolution: {integrity: sha512-T1gkAiYYDWYx3V5Bmyu7HcfcvL7mUrTWiM6yOfa3PIphViJ/gFPbvidQ+veqSOHci/PxBcDabeUNCzpOODJZig==}
engines: {node: '>= 14'}
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
https-proxy-agent@5.0.1:
resolution: {integrity: sha512-dFcAjpTQFgoLMzC2VwU+C/CbS7uRL0lWmxDITmqm7C+7F0Odmj6s9l6alZc6AELXhrnggM2CeWSXHGOdX2YtwA==}
engines: {node: '>= 6'}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
https-proxy-agent@7.0.6:
resolution: {integrity: sha512-vK9P5/iUfdl95AI+JVyUuIcVtd4ofvtrOr3HNtM2yxC9bnMbEdp3x01OhQNnjb8IJYi38VlTE3mBXwcfvywuSw==}
engines: {node: '>= 14'}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
human-signals@2.1.0:
resolution: {integrity: sha512-B4FFZ6q/T2jhhksgkbEW3HBvWIfDW85snkQgawt07S7J5QXTk6BkNV+0yAeZrM5QpMAdYlocGoljn0sJ/WQkFw==}
engines: {node: '>=10.17.0'}
first commit
2026-02-21 03:10:39 +00:00
iconv-lite@0.7.2:
resolution: {integrity: sha512-im9DjEDQ55s9fL4EYzOAv0yMqmMBSZp6G0VvFyTMPKWxiSBHUj9NW/qqLmXUwXrrM7AvqSlTCfvqRb0cM8yYqw==}
engines: {node: '>=0.10.0'}
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
ieee754@1.2.1:
resolution: {integrity: sha512-dcyqhDvX1C46lXZcVqCpK+FtMRQVdIMN6/Df5js2zouUsqG7I6sFxitIC+7KYK29KdXOLHdu9zL4sFnoVQnqaA==}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
ignore-walk@8.0.0:
resolution: {integrity: sha512-FCeMZT4NiRQGh+YkeKMtWrOmBgWjHjMJ26WQWrRQyoyzqevdaGSakUaJW5xQYmjLlUVk2qUnCjYVBax9EKKg8A==}
engines: {node: ^20.17.0 || >=22.9.0}
first commit
2026-02-21 03:10:39 +00:00
ignore@5.3.2:
resolution: {integrity: sha512-hsBTNUqQTDwkWtcdYI2i06Y/nUBEsNEDJKjWdigLvegy8kDuJAS8uRlpkkcQpyEXL0Z/pjDy5HBmMjRCJ2gq+g==}
engines: {node: '>= 4'}
ignore@7.0.5:
resolution: {integrity: sha512-Hs59xBNfUIunMFgWAbGX5cq6893IbWg4KnrjbYwX3tx0ztorVgTDA6B2sxf8ejHJ4wz8BqGUMYlnzNBer5NvGg==}
engines: {node: '>= 4'}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
import-in-the-middle@3.3.1:
resolution: {integrity: sha512-0rymlHSFLwZ0ixx8DaQkoIyZojJPY2a0K2nEYslhKJ6jIYO/m0IcCb7iQsFPmS7WmKwISZiIrv5Icstrw/CmqA==}
engines: {node: '>=18'}
first commit
2026-02-21 03:10:39 +00:00
imurmurhash@0.1.4:
resolution: {integrity: sha512-JmXMZ6wuvDmLiHEml9ykzqO6lwFbof0GG4IkcGaENdCRDDmMVnny7s5HsIgHCbaq0w2MyPhDqkhTUgS2LU2PHA==}
engines: {node: '>=0.8.19'}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
indent-string@4.0.0:
resolution: {integrity: sha512-EdDDZu4A2OyIK7Lr/2zG+w5jmbuk1DVBnEwREQvBzspBJkCEbRa8GxU1lghYcaGJCnRWibjDXlq779X1/y5xwg==}
engines: {node: '>=8'}
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
indent-string@5.0.0:
resolution: {integrity: sha512-m6FAo/spmsW2Ab2fU35JTYwtOKa2yAwXSwgjSv1TJzh4Mh7mC3lzAOVLBprb72XsTrgkEIsl7YrFNAiDiRhIGg==}
engines: {node: '>=12'}
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
inflight@1.0.6:
resolution: {integrity: sha512-k92I/b08q4wvFscXCLvqfsHCrjrF7yiXsQuIVvVE7N82W3+aqpzuUdBbfhWcy/FZR3/4IgflMgKLOsvPDrGCJA==}
deprecated: This module is not supported, and leaks memory. Do not use it. Check out lru-cache if you want a good and tested way to coalesce async requests by a key value, which is much more comprehensive and powerful.
first commit
2026-02-21 03:10:39 +00:00
inherits@2.0.4:
resolution: {integrity: sha512-k/vGaX4/Yla3WzyMCvTQOXYeIHvqOKtnqBduzTHpzpQZzAskKMhZ2K+EnBiSM9zGSoIFeMpXKxa4dYeZIQqewQ==}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
ini@2.0.0:
resolution: {integrity: sha512-7PnF4oN3CvZF23ADhA5wRaYEQpJ8qygSkbtTXWBeXWXmEVRXK+1ITciHWwHhsjv1TmW0MgacIv6hEi5pX5NQdA==}
engines: {node: '>=10'}
ini@6.0.0:
resolution: {integrity: sha512-IBTdIkzZNOpqm7q3dRqJvMaldXjDHWkEDfrwGEQTs5eaQMWV+djAhR+wahyNNMAa+qpbDUhBMVt4ZKNwpPm7xQ==}
engines: {node: ^20.17.0 || >=22.9.0}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
ini@7.0.0:
resolution: {integrity: sha512-ifK0CgjALofS5bkrcTy4RaQ9Vx2Knf/eLeIO+NaswQEpH1UblrtTSCIvN71qQDMq0PeQ/SSPojvEJp9vvvfr+w==}
engines: {node: ^22.22.2 || ^24.15.0 || >=26.0.0}
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
ink@6.8.0:
resolution: {integrity: sha512-sbl1RdLOgkO9isK42WCZlJCFN9hb++sX9dsklOvfd1YQ3bQ2AiFu12Q6tFlr0HvEUvzraJntQCCpfEoUe9DSzA==}
engines: {node: '>=20'}
peerDependencies:
'@types/react': '>=19.0.0'
react: '>=19.0.0'
react-devtools-core: '>=6.1.2'
peerDependenciesMeta:
'@types/react':
optional: true
react-devtools-core:
optional: true
first commit
2026-02-21 03:10:39 +00:00
inquirer@12.11.1:
resolution: {integrity: sha512-9VF7mrY+3OmsAfjH3yKz/pLbJ5z22E23hENKw3/LNSaA/sAt3v49bDRY+Ygct1xwuKT+U+cBfTzjCPySna69Qw==}
engines: {node: '>=18'}
peerDependencies:
'@types/node': '>=18'
peerDependenciesMeta:
'@types/node':
optional: true
ip-address@10.0.1:
resolution: {integrity: sha512-NWv9YLW4PoW2B7xtzaS3NCot75m6nK7Icdv0o3lfMceJVRfSoQwqD4wEH5rLwoKJwUiZ/rfpiVBhnaF0FK4HoA==}
engines: {node: '>= 12'}
ipaddr.js@1.9.1:
resolution: {integrity: sha512-0KI/607xoxSToH7GjN1FfSbLoU0+btTicjsQSWQlh/hZykN8KpmMf7uYwPW3R+akZ6R/w18ZlXSHBYXiYUPO3g==}
engines: {node: '>= 0.10'}
ipaddr.js@2.3.0:
resolution: {integrity: sha512-Zv/pA+ciVFbCSBBjGfaKUya/CcGmUHzTydLMaTwrUUEM2DIEO3iZvueGxmacvmN50fGpGVKeTXpb2LcYQxeVdg==}
engines: {node: '>= 10'}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
is-core-module@2.16.2:
resolution: {integrity: sha512-evOr8xfXKxE6qSR0hSXL2r3sd7ALj8+7jQEUvPYcm5sgZFdJ+AYzT6yNmJenvIYQBgIGwfwz08sL8zoL7yq2BA==}
engines: {node: '>= 0.4'}
first commit
2026-02-21 03:10:39 +00:00
is-extglob@2.1.1:
resolution: {integrity: sha512-SbKbANkN603Vi4jEZv49LeVJMn4yGwsbzZworEoyEiutsN3nJYdbO36zfhGJ6QEDpOZIFkDtnq5JRxmvl3jsoQ==}
engines: {node: '>=0.10.0'}
is-fullwidth-code-point@3.0.0:
resolution: {integrity: sha512-zymm5+u+sCsSWyD9qNaejV3DFvhCKclKdizYaJUuHA83RLjb7nSuGnddCHGv0hk+KY7BMAlsWeK4Ueg6EV6XQg==}
engines: {node: '>=8'}
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
is-fullwidth-code-point@5.1.0:
resolution: {integrity: sha512-5XHYaSyiqADb4RnZ1Bdad6cPp8Toise4TzEjcOYDHZkTCbKgiUl7WTUCpNWHuxmDt91wnsZBc9xinNzopv3JMQ==}
engines: {node: '>=18'}
first commit
2026-02-21 03:10:39 +00:00
is-glob@4.0.3:
resolution: {integrity: sha512-xelSayHH36ZgE7ZWhli7pW34hNbNl8Ojv5KVmkJD4hBdD3th8Tfk9vYasLM+mXWOZhFkgZfxhLSnrwRr4elSSg==}
engines: {node: '>=0.10.0'}
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
is-in-ci@2.0.0:
resolution: {integrity: sha512-cFeerHriAnhrQSbpAxL37W1wcJKUUX07HyLWZCW1URJT/ra3GyUTzBgUnh24TMVfNTV2Hij2HLxkPHFZfOZy5w==}
engines: {node: '>=20'}
hasBin: true
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
is-potential-custom-element-name@1.0.1:
resolution: {integrity: sha512-bCYeRA2rVibKZd+s2625gGnGF/t7DSqDs4dP7CrLA1m7jKWz6pps0LpYLJN8Q64HtmPKJ1hrN3nzPNKFEKOUiQ==}
first commit
2026-02-21 03:10:39 +00:00
is-promise@4.0.0:
resolution: {integrity: sha512-hvpoI6korhJMnej285dSg6nu1+e6uxs7zG3BYAm5byqDsgJNWwxzM6z6iZiAgQR4TJ30JmBTOwqZUw3WlyH3AQ==}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
is-stream@2.0.1:
resolution: {integrity: sha512-hFoiJiTl63nn+kstHGBtewWSKnQLpyb155KHheA1l39uvtO9nWIop1p3udqPcUd/xbF1VLMO4n7OI6p7RbngDg==}
engines: {node: '>=8'}
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
is-unicode-supported@2.1.0:
resolution: {integrity: sha512-mE00Gnza5EEB3Ds0HfMyllZzbBrmLOX3vfWoj9A9PEnTfratQ/BcaJOuMhnkhjXvb2+FkY3VuHqtAGpTPmglFQ==}
engines: {node: '>=18'}
first commit
2026-02-21 03:10:39 +00:00
isexe@2.0.0:
resolution: {integrity: sha512-RHxMLp9lnKHGHRng9QFhRCMbYAcVpn69smSGcq3f36xjgVVWThj4qqLbTLlq7Ssj8B+fIQ1EuCEGI2lKsyQeIw==}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
isexe@4.0.0:
resolution: {integrity: sha512-FFUtZMpoZ8RqHS3XeXEmHWLA4thH+ZxCv2lOiPIn1Xc7CxrqhWzNSDzD+/chS/zbYezmiwWLdQC09JdQKmthOw==}
engines: {node: '>=20'}
feat: add Kubernetes orchestrator for MCP server pod management mcpd can now deploy MCP server instances as Kubernetes pods instead of Docker containers. Set MCPD_ORCHESTRATOR=kubernetes to enable. - Add @kubernetes/client-node with thin wrapper (context enforcement via MCPD_K8S_CONTEXT to prevent multi-cluster mishaps) - Rewrite KubernetesOrchestrator: pod CRUD, pod IP extraction, exec via SPDY (one-shot + interactive), log streaming - Manifest generator: stdin:true for STDIO servers, args (not command) to preserve runner image entrypoint, security hardening - Orchestrator selection in main.ts via MCPD_ORCHESTRATOR env var - 25 unit tests for k8s orchestrator, all 624 tests pass Tested end-to-end on local k3s: - mcpd deployed via Pulumi, creates pods in mcpctl-servers namespace - NetworkPolicy verified: only mcpd can reach MCP server pods - Python runner (uvx) successfully runs aws-documentation-mcp-server Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-08 01:55:13 +01:00
isomorphic-ws@5.0.0:
resolution: {integrity: sha512-muId7Zzn9ywDsyXgTIafTry2sV3nySZeUDe6YedVd1Hvuuep5AsIlqK+XefWpYTyJG5e503F2xIuT2lcU6rCSw==}
peerDependencies:
ws: '*'
first commit
2026-02-21 03:10:39 +00:00
istanbul-lib-coverage@3.2.2:
resolution: {integrity: sha512-O8dpsF+r0WV/8MNRKfnmrtCWhuKjxrq2w+jpzBL5UZKTi2LeVWnWOmWRxFlesJONmc+wLAGvKQZEOanko0LFTg==}
engines: {node: '>=8'}
istanbul-lib-report@3.0.1:
resolution: {integrity: sha512-GCfE1mtsHGOELCU8e/Z7YWzpmybrx/+dSTfLrvY8qRmaY6zXTKWn6WQIjaAFw069icm6GVMNkgu0NzI4iPZUNw==}
engines: {node: '>=10'}
istanbul-reports@3.2.0:
resolution: {integrity: sha512-HGYWWS/ehqTV3xN10i23tkPkpH46MLCIMFNCaaKNavAXTF1RkqxawEPtnjnGZ6XKSInBKkiOA5BKS+aZiY3AvA==}
engines: {node: '>=8'}
feat(mcpd+deploy): serve web UI at /ui + smoke tests + docs (Stage 6) The closing stage. mcpd now hosts the Stage 5 SPA, the Docker image bundles the build artifact, a smoke test exercises the personality HTTP surface end-to-end, and the user-facing docs spell out the mental model. mcpd: - Add @fastify/static dep. - New routes/web-ui.ts: registers /ui/* against a static bundle. Looks for the bundle at $MCPD_WEB_ROOT, then /usr/share/mcpd/web (the Docker image path), then a dev-tree fallback. Logs and skips cleanly if missing — API-only deploys keep working. - SPA fallback: any /ui/<path> that doesn't match a file falls through to index.html so direct hits to react-router URLs work. - /ui/* falls through to `kind: skip` in mapUrlToPermission, so the static assets are served unauthenticated. Each API call from the SPA still carries the bearer token. Deploy: - Dockerfile.mcpd builds the @mcpctl/web bundle in the same builder stage and copies dist/ to /usr/share/mcpd/web in the runtime image. Smoke (personality.smoke.test.ts): - Live mcpd flow: create secret/llm/agent/personality, attach an agent-direct prompt, verify the binding listing, reject double- attach (409) + foreign-agent prompt (400), set defaultPersonality by name, detach + delete cleanup. Docs: - New docs/personalities.md: VLAN-on-ethernet model, system-block ordering table, three prompt scopes, CLI walkthrough, web UI walkthrough, full API surface, RBAC notes. - agents.md and chat.md cross-link. - README's Agents section gains a Personalities subsection. Test count after Stage 6: mcpd: 801/801 cli: 430/430 web: 7/7 db: 58/62 (4 pre-existing) Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:48:43 +01:00
jackspeak@4.2.3:
resolution: {integrity: sha512-ykkVRwrYvFm1nb2AJfKKYPr0emF6IiXDYUaFx4Zn9ZuIH7MrzEZ3sD5RlqGXNRpHtvUHJyOnCEFxOlNDtGo7wg==}
engines: {node: 20 || >=22}
first commit
2026-02-21 03:10:39 +00:00
jiti@2.6.1:
resolution: {integrity: sha512-ekilCSN1jwRvIbgeg/57YFh8qQDNbwDb9xT/qu2DAHbFFZUicIl4ygVaAvzveMhMVr3LnpSKTNnwt8PoOfmKhQ==}
hasBin: true
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
jiti@2.7.0:
resolution: {integrity: sha512-AC/7JofJvZGrrneWNaEnJeOLUx+JlGt7tNa0wZiRPT4MY1wmfKjt2+6O2p2uz2+skll8OZZmJMNqeke7kKbNgQ==}
hasBin: true
first commit
2026-02-21 03:10:39 +00:00
jose@6.1.3:
resolution: {integrity: sha512-0TpaTfihd4QMNwrz/ob2Bp7X04yuxJkjRGi4aKmOqwhov54i6u79oCv7T+C7lo70MKH6BesI3vscD1yb/yzKXQ==}
js-tokens@10.0.0:
resolution: {integrity: sha512-lM/UBzQmfJRo9ABXbPWemivdCW8V2G8FHaHdypQaIy523snUjog0W71ayWXTjiR+ixeMyVHN2XcpnTd/liPg/Q==}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
js-tokens@4.0.0:
resolution: {integrity: sha512-RdJUflcE3cUzKiMqQgsCu06FPu9UdIJO0beYbPhHN4k6apgJtifcoCtT9bcxOpYBtpD2kCM6Sbzg4CausW/PKQ==}
first commit
2026-02-21 03:10:39 +00:00
js-yaml@4.1.1:
resolution: {integrity: sha512-qQKT4zQxXl8lLwBtHMWwaTcGfFOZviOJet3Oy/xmGk2gZH677CJM9EvtfdSkgWcATZhj/55JZ0rmy3myCT5lsA==}
hasBin: true
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
jsdom@28.1.0:
resolution: {integrity: sha512-0+MoQNYyr2rBHqO1xilltfDjV9G7ymYGlAUazgcDLQaUf8JDHbuGwsxN6U9qWaElZ4w1B2r7yEGIL3GdeW3Rug==}
engines: {node: ^20.19.0 || ^22.12.0 || >=24.0.0}
peerDependencies:
canvas: ^3.0.0
peerDependenciesMeta:
canvas:
optional: true
feat: add Kubernetes orchestrator for MCP server pod management mcpd can now deploy MCP server instances as Kubernetes pods instead of Docker containers. Set MCPD_ORCHESTRATOR=kubernetes to enable. - Add @kubernetes/client-node with thin wrapper (context enforcement via MCPD_K8S_CONTEXT to prevent multi-cluster mishaps) - Rewrite KubernetesOrchestrator: pod CRUD, pod IP extraction, exec via SPDY (one-shot + interactive), log streaming - Manifest generator: stdin:true for STDIO servers, args (not command) to preserve runner image entrypoint, security hardening - Orchestrator selection in main.ts via MCPD_ORCHESTRATOR env var - 25 unit tests for k8s orchestrator, all 624 tests pass Tested end-to-end on local k3s: - mcpd deployed via Pulumi, creates pods in mcpctl-servers namespace - NetworkPolicy verified: only mcpd can reach MCP server pods - Python runner (uvx) successfully runs aws-documentation-mcp-server Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-08 01:55:13 +01:00
jsep@1.4.0:
resolution: {integrity: sha512-B7qPcEVE3NVkmSJbaYxvv4cHkVW7DQsZz13pUMrfS8z8Q/BuShN+gcTXrUlPiGqM2/t/EEaI030bpxMqY8gMlw==}
engines: {node: '>= 10.16.0'}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
jsesc@3.1.0:
resolution: {integrity: sha512-/sM3dO2FOzXjKQhJuo0Q173wf2KOo8t4I8vHy6lF9poUp7bKT0/NHE8fPX23PwfhnykfqnC2xRxOnVw5XuGIaA==}
engines: {node: '>=6'}
hasBin: true
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
json-bigint@1.0.0:
resolution: {integrity: sha512-SiPv/8VpZuWbvLSMtTDU8hEfrZWg/mH/nV/b4o0CYbSxu1UIQPLdwKOCIyLQX+VIPO5vrLX3i8qtqFyhdPSUSQ==}
first commit
2026-02-21 03:10:39 +00:00
json-buffer@3.0.1:
resolution: {integrity: sha512-4bV5BfR2mqfQTJm+V5tPPdf+ZpuhiIvTuAB5g8kcrXOZpTT/QwwVRWBywX1ozr6lEuPdbHxwaJlm9G6mI2sfSQ==}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
json-parse-even-better-errors@5.0.0:
resolution: {integrity: sha512-ZF1nxZ28VhQouRWhUcVlUIN3qwSgPuswK05s/HIaoetAoE/9tngVmCHjSxmSQPav1nd+lPtTL0YZ/2AFdR/iYQ==}
engines: {node: ^20.17.0 || >=22.9.0}
first commit
2026-02-21 03:10:39 +00:00
json-schema-ref-resolver@3.0.0:
resolution: {integrity: sha512-hOrZIVL5jyYFjzk7+y7n5JDzGlU8rfWDuYyHwGa2WA8/pcmMHezp2xsVwxrebD/Q9t8Nc5DboieySDpCp4WG4A==}
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
json-schema-to-ts@3.1.1:
resolution: {integrity: sha512-+DWg8jCJG2TEnpy7kOm/7/AxaYoaRbjVB4LFZLySZlWn8exGs3A4OLJR966cVvU26N7X9TWxl+Jsw7dzAqKT6g==}
engines: {node: '>=16'}
first commit
2026-02-21 03:10:39 +00:00
json-schema-traverse@0.4.1:
resolution: {integrity: sha512-xbbCH5dCYU5T8LcEhhuh7HJ88HXuW3qsI3Y0zOZFKfZEHcpWiHU/Jxzk629Brsab/mMiHQti9wMP+845RPe3Vg==}
json-schema-traverse@1.0.0:
resolution: {integrity: sha512-NM8/P9n3XjXhIZn1lLhkFaACTOURQXjWhV4BA/RnOv8xvgqtqpAX9IO4mRQxSx1Rlo4tqzeqb0sOlruaOy3dug==}
json-schema-typed@8.0.2:
resolution: {integrity: sha512-fQhoXdcvc3V28x7C7BMs4P5+kNlgUURe2jmUT1T//oBRMDrqy1QPelJimwZGo7Hg9VPV3EQV5Bnq4hbFy2vetA==}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
json-schema@0.4.0:
resolution: {integrity: sha512-es94M3nTIfsEPisRafak+HDLfHXnKBhV3vU5eqPcS3flIWqcxJWgXHXiey3YrpaNsanY5ei1VoYEbOzijuq9BA==}
first commit
2026-02-21 03:10:39 +00:00
json-stable-stringify-without-jsonify@1.0.1:
resolution: {integrity: sha512-Bdboy+l7tA3OGW6FjyFHWkP5LuByj1Tk33Ljyq0axyzdk9//JSi2u3fP1QSmd1KNwq6VOKYGlAu87CisVir6Pw==}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
json-stringify-nice@1.1.4:
resolution: {integrity: sha512-5Z5RFW63yxReJ7vANgW6eZFGWaQvnPE3WNmZoOJrSkGju2etKA2L5rrOa1sm877TVTFt57A80BH1bArcmlLfPw==}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
json5@2.2.3:
resolution: {integrity: sha512-XmOWe7eyHYH14cLdVPoyg+GOH3rYX++KpzrylJwSW98t3Nk+U8XOl8FWKOgwtzdb8lXGf6zYwDUzeHMWfxasyg==}
engines: {node: '>=6'}
hasBin: true
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
jsonparse@1.3.1:
resolution: {integrity: sha512-POQXvpdL69+CluYsillJ7SUhKvytYjW9vG/GKpnf+xP8UWgYEM/RaMzHHofbALDiKbbP1W8UEYmgGl39WkPZsg==}
engines: {'0': node >= 0.2.0}
feat: add Kubernetes orchestrator for MCP server pod management mcpd can now deploy MCP server instances as Kubernetes pods instead of Docker containers. Set MCPD_ORCHESTRATOR=kubernetes to enable. - Add @kubernetes/client-node with thin wrapper (context enforcement via MCPD_K8S_CONTEXT to prevent multi-cluster mishaps) - Rewrite KubernetesOrchestrator: pod CRUD, pod IP extraction, exec via SPDY (one-shot + interactive), log streaming - Manifest generator: stdin:true for STDIO servers, args (not command) to preserve runner image entrypoint, security hardening - Orchestrator selection in main.ts via MCPD_ORCHESTRATOR env var - 25 unit tests for k8s orchestrator, all 624 tests pass Tested end-to-end on local k3s: - mcpd deployed via Pulumi, creates pods in mcpctl-servers namespace - NetworkPolicy verified: only mcpd can reach MCP server pods - Python runner (uvx) successfully runs aws-documentation-mcp-server Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-08 01:55:13 +01:00
jsonpath-plus@10.4.0:
resolution: {integrity: sha512-T92WWatJXmhBbKsgH/0hl+jxjdXrifi5IKeMY02DWggRxX0UElcbVzPlmgLTbvsPeW1PasQ6xE2Q75stkhGbsA==}
engines: {node: '>=18.0.0'}
hasBin: true
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
just-diff-apply@5.5.0:
resolution: {integrity: sha512-OYTthRfSh55WOItVqwpefPtNt2VdKsq5AnAK6apdtR6yCH8pr0CmSr710J0Mf+WdQy7K/OzMy7K2MgAfdQURDw==}
just-diff@6.0.2:
resolution: {integrity: sha512-S59eriX5u3/QhMNq3v/gm8Kd0w8OS6Tz2FS1NG4blv+z0MuQcBRJyFWjdovM0Rad4/P4aUPFtnkNjMjyMlMSYA==}
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
jwa@2.0.1:
resolution: {integrity: sha512-hRF04fqJIP8Abbkq5NKGN0Bbr3JxlQ+qhZufXVr0DvujKy93ZCbXZMHDL4EOtodSbCWxOqR8MS1tXA5hwqCXDg==}
jws@4.0.1:
resolution: {integrity: sha512-EKI/M/yqPncGUUh44xz0PxSidXFr/+r0pA70+gIYhjv+et7yxM+s29Y+VGDkovRofQem0fs7Uvf4+YmAdyRduA==}
first commit
2026-02-21 03:10:39 +00:00
keyv@4.5.4:
resolution: {integrity: sha512-oxVHkHR/EJf2CNXnWxRLW6mg7JyCCUcG0DtEGmL2ctUo1PNTin1PUil+r/+4r5MpVgC/fn1kjsx7mjSujKqIpw==}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
kubernetes-types@1.30.0:
resolution: {integrity: sha512-Dew1okvhM/SQcIa2rcgujNndZwU8VnSapDgdxlYoB84ZlpAD43U6KLAFqYo17ykSFGHNPrg0qry0bP+GJd9v7Q==}
first commit
2026-02-21 03:10:39 +00:00
levn@0.4.1:
resolution: {integrity: sha512-+bT2uH4E5LGE7h/n3evcS/sQlJXCpIp6ym8OWJ5eV6+67Dsql/LaaT7qJBAt2rzfoa/5QBGBhxDix1dMt2kQKQ==}
engines: {node: '>= 0.8.0'}
light-my-request@6.6.0:
resolution: {integrity: sha512-CHYbu8RtboSIoVsHZ6Ye4cj4Aw/yg2oAFimlF7mNvfDV192LR7nDiKtSIfCuLT7KokPSTn/9kfVLm5OGN0A28A==}
feat(web): bold redesign — Tailwind v4 + shadcn-style primitives + Skills/Proposals/Revisions UI Phase 6 of the Skills + Revisions + Proposals work. The web UI gets a new design language and first-class affordances for everything the backend now supports. ## Visual direction - Tailwind v4 with custom @theme block (oklch tokens). Dark-mode-only (internal tool — light mode doubles QA surface). - Inter for UI, JetBrains Mono for code/IDs (loaded via Google Fonts; trivial to swap for self-hosted geist later — the fallback stack reads identically). - Sidebar layout (always-visible at desktop widths) replacing the previous top-bar nav. Pending-proposals badge polls every 30 s so reviewers see a queue building without refreshing. - Lucide icons throughout. - Spacing and radii on Tailwind defaults. Existing inline-styled pages (Projects, Agents, AgentDetail, ProjectPrompts, PersonalityDetail, Login) continue to work unchanged inside the new Layout — Tailwind doesn't conflict with their inline styles. A follow-up can migrate them incrementally. ## What's added ### Build infra (src/web/) - package.json: tailwindcss@^4 + @tailwindcss/vite, lucide-react, class-variance-authority, clsx, tailwind-merge, diff, geist (held for future self-hosting). - vite.config.ts: registers the @tailwindcss/vite plugin. - src/index.css: Tailwind import + @theme tokens + @layer base. - src/main.tsx: imports index.css. - src/lib/utils.ts: shadcn-style cn() helper. ### shadcn-style primitives (src/components/ui/) Hand-written rather than generated via `npx shadcn` so the repo doesn't depend on a CLI tool that needs interactive runtime: - button.tsx — variants: primary / secondary / ghost / danger / link; sizes: sm / md / lg / icon. - card.tsx — Card + Header/Title/Description/Content/Footer subparts. - badge.tsx — variants: default / info / success / warning / danger / outline. - input.tsx — Input + Textarea + Label. - tabs.tsx — no-dep accessible Tabs (no Radix needed for our use). - separator.tsx — h/v separator with role=separator. ### Diff component (src/components/Diff.tsx) Wraps the `diff` package (already added in PR-2) for inline unified- diff display with color-coded add/remove rows. Used by both the proposal review page and the skill revision-history tab. ### New pages (src/pages/) - Dashboard.tsx — at-a-glance home. Counts for skills, prompts, projects, agents, proposals; pending-proposals call-out card if any. - Skills.tsx — list view, separated into Global vs Project/Agent- scoped sections. - SkillDetail.tsx — name + semver + description; tabs for SKILL.md / Files / Metadata / History. History tab shows revisions with click-to-diff against the live body. - Proposals.tsx — queue with Pending/Approved/Rejected tabs. Pending count is highlighted in amber. - ProposalDetail.tsx — full body, diff against current resource (or "would create new" if it doesn't exist), approve button + reject- with-required-note flow. ### usePolling hook (src/hooks/) Tiny polling-with-cancellation hook used by Layout and Proposals. ### Layout rewrite (src/components/Layout.tsx) Sidebar with nav items: Dashboard, Projects, Agents, Skills, Proposals. Lucide icons. Active-route highlighting via NavLink. Pending-proposals warning badge on the Proposals item. ### Routes (src/App.tsx) New routes: /dashboard, /skills, /skills/:name, /proposals, /proposals/:id. Default redirects to /dashboard. ### API types (src/api.ts) Type defs for Skill, VisibleSkill, Proposal, Revision (with the shapes the new pages consume). ## Tests Existing 7 web tests still pass (Login + api). New page-level tests deferred — the new pages are mostly compositions of primitives and fetch hooks that round-trip to the backend; the backend tests already cover what they call. PR-7 polish can add render-and-click tests if coverage drift surfaces. ## Verification - `pnpm --filter @mcpctl/web build` clean, no warnings. - `pnpm test:run` whole monorepo: 162 test files / 2157 tests green. - Visual smoke deferred — needs a running mcpd to populate the fixtures. Manual smoke tested locally is the next step. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-07 17:54:55 +01:00
lightningcss-android-arm64@1.32.0:
resolution: {integrity: sha512-YK7/ClTt4kAK0vo6w3X+Pnm0D2cf2vPHbhOXdoNti1Ga0al1P4TBZhwjATvjNwLEBCnKvjJc2jQgHXH0NEwlAg==}
engines: {node: '>= 12.0.0'}
cpu: [arm64]
os: [android]
lightningcss-darwin-arm64@1.32.0:
resolution: {integrity: sha512-RzeG9Ju5bag2Bv1/lwlVJvBE3q6TtXskdZLLCyfg5pt+HLz9BqlICO7LZM7VHNTTn/5PRhHFBSjk5lc4cmscPQ==}
engines: {node: '>= 12.0.0'}
cpu: [arm64]
os: [darwin]
lightningcss-darwin-x64@1.32.0:
resolution: {integrity: sha512-U+QsBp2m/s2wqpUYT/6wnlagdZbtZdndSmut/NJqlCcMLTWp5muCrID+K5UJ6jqD2BFshejCYXniPDbNh73V8w==}
engines: {node: '>= 12.0.0'}
cpu: [x64]
os: [darwin]
lightningcss-freebsd-x64@1.32.0:
resolution: {integrity: sha512-JCTigedEksZk3tHTTthnMdVfGf61Fky8Ji2E4YjUTEQX14xiy/lTzXnu1vwiZe3bYe0q+SpsSH/CTeDXK6WHig==}
engines: {node: '>= 12.0.0'}
cpu: [x64]
os: [freebsd]
lightningcss-linux-arm-gnueabihf@1.32.0:
resolution: {integrity: sha512-x6rnnpRa2GL0zQOkt6rts3YDPzduLpWvwAF6EMhXFVZXD4tPrBkEFqzGowzCsIWsPjqSK+tyNEODUBXeeVHSkw==}
engines: {node: '>= 12.0.0'}
cpu: [arm]
os: [linux]
lightningcss-linux-arm64-gnu@1.32.0:
resolution: {integrity: sha512-0nnMyoyOLRJXfbMOilaSRcLH3Jw5z9HDNGfT/gwCPgaDjnx0i8w7vBzFLFR1f6CMLKF8gVbebmkUN3fa/kQJpQ==}
engines: {node: '>= 12.0.0'}
cpu: [arm64]
os: [linux]
lightningcss-linux-arm64-musl@1.32.0:
resolution: {integrity: sha512-UpQkoenr4UJEzgVIYpI80lDFvRmPVg6oqboNHfoH4CQIfNA+HOrZ7Mo7KZP02dC6LjghPQJeBsvXhJod/wnIBg==}
engines: {node: '>= 12.0.0'}
cpu: [arm64]
os: [linux]
lightningcss-linux-x64-gnu@1.32.0:
resolution: {integrity: sha512-V7Qr52IhZmdKPVr+Vtw8o+WLsQJYCTd8loIfpDaMRWGUZfBOYEJeyJIkqGIDMZPwPx24pUMfwSxxI8phr/MbOA==}
engines: {node: '>= 12.0.0'}
cpu: [x64]
os: [linux]
lightningcss-linux-x64-musl@1.32.0:
resolution: {integrity: sha512-bYcLp+Vb0awsiXg/80uCRezCYHNg1/l3mt0gzHnWV9XP1W5sKa5/TCdGWaR/zBM2PeF/HbsQv/j2URNOiVuxWg==}
engines: {node: '>= 12.0.0'}
cpu: [x64]
os: [linux]
lightningcss-win32-arm64-msvc@1.32.0:
resolution: {integrity: sha512-8SbC8BR40pS6baCM8sbtYDSwEVQd4JlFTOlaD3gWGHfThTcABnNDBda6eTZeqbofalIJhFx0qKzgHJmcPTnGdw==}
engines: {node: '>= 12.0.0'}
cpu: [arm64]
os: [win32]
lightningcss-win32-x64-msvc@1.32.0:
resolution: {integrity: sha512-Amq9B/SoZYdDi1kFrojnoqPLxYhQ4Wo5XiL8EVJrVsB8ARoC1PWW6VGtT0WKCemjy8aC+louJnjS7U18x3b06Q==}
engines: {node: '>= 12.0.0'}
cpu: [x64]
os: [win32]
lightningcss@1.32.0:
resolution: {integrity: sha512-NXYBzinNrblfraPGyrbPoD19C1h9lfI/1mzgWYvXUTe414Gz/X1FD2XBZSZM7rRTrMA8JL3OtAaGifrIKhQ5yQ==}
engines: {node: '>= 12.0.0'}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
locate-path@3.0.0:
resolution: {integrity: sha512-7AO748wWnIhNqAuaty2ZWHkQHRSNfPVIsPIfwEOWO22AmaoVrWavlOcMR5nzTLNYvp36X220/maaRsrec1G65A==}
engines: {node: '>=6'}
first commit
2026-02-21 03:10:39 +00:00
locate-path@6.0.0:
resolution: {integrity: sha512-iPZK6eYjbxRu3uB4/WZ3EsEIMJFMqAoopl3R+zuq0UjcAm/MO6KCweDgPfP3elTztoKP3KtnVHxTn2NHBSDVUw==}
engines: {node: '>=10'}
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
lodash.camelcase@4.3.0:
resolution: {integrity: sha512-TwuEnCnxbc3rAvhf/LbG7tJUDzhqXyFnv3dtzLOPgCG/hODL7WFnsbwktkD7yUV0RrreP/l1PALq/YSg6VvjlA==}
long@5.3.2:
resolution: {integrity: sha512-mNAgZ1GmyNhD7AuqnTG3/VQ26o760+ZYBPKjPvugO8+nLbYfX6TVpJPseBvopbdY+qpZ/lKUnmEc1LeZYS3QAA==}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
lru-cache@10.4.3:
resolution: {integrity: sha512-JNAzZcXrCt42VGLuYz0zfAzDfAvJWW6AfYlDBQyDV5DClI2m5sAmK+OIO7s59XfsRsWHp02jAJrRadPRGTt6SQ==}
first commit
2026-02-21 03:10:39 +00:00
lru-cache@11.2.6:
resolution: {integrity: sha512-ESL2CrkS/2wTPfuend7Zhkzo2u0daGJ/A2VucJOgQ/C48S/zB8MMeMHSGKYpXhIjbPxfuezITkaBH1wqv00DDQ==}
engines: {node: 20 || >=22}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
lru-cache@5.1.1:
resolution: {integrity: sha512-KpNARQA3Iwv+jTA0utUVVbrh+Jlrr1Fv0e56GGzAFOXN7dk/FviaDW8LHmK52DlcH4WP2n6gI8vN1aesBFgo9w==}
feat(web): bold redesign — Tailwind v4 + shadcn-style primitives + Skills/Proposals/Revisions UI Phase 6 of the Skills + Revisions + Proposals work. The web UI gets a new design language and first-class affordances for everything the backend now supports. ## Visual direction - Tailwind v4 with custom @theme block (oklch tokens). Dark-mode-only (internal tool — light mode doubles QA surface). - Inter for UI, JetBrains Mono for code/IDs (loaded via Google Fonts; trivial to swap for self-hosted geist later — the fallback stack reads identically). - Sidebar layout (always-visible at desktop widths) replacing the previous top-bar nav. Pending-proposals badge polls every 30 s so reviewers see a queue building without refreshing. - Lucide icons throughout. - Spacing and radii on Tailwind defaults. Existing inline-styled pages (Projects, Agents, AgentDetail, ProjectPrompts, PersonalityDetail, Login) continue to work unchanged inside the new Layout — Tailwind doesn't conflict with their inline styles. A follow-up can migrate them incrementally. ## What's added ### Build infra (src/web/) - package.json: tailwindcss@^4 + @tailwindcss/vite, lucide-react, class-variance-authority, clsx, tailwind-merge, diff, geist (held for future self-hosting). - vite.config.ts: registers the @tailwindcss/vite plugin. - src/index.css: Tailwind import + @theme tokens + @layer base. - src/main.tsx: imports index.css. - src/lib/utils.ts: shadcn-style cn() helper. ### shadcn-style primitives (src/components/ui/) Hand-written rather than generated via `npx shadcn` so the repo doesn't depend on a CLI tool that needs interactive runtime: - button.tsx — variants: primary / secondary / ghost / danger / link; sizes: sm / md / lg / icon. - card.tsx — Card + Header/Title/Description/Content/Footer subparts. - badge.tsx — variants: default / info / success / warning / danger / outline. - input.tsx — Input + Textarea + Label. - tabs.tsx — no-dep accessible Tabs (no Radix needed for our use). - separator.tsx — h/v separator with role=separator. ### Diff component (src/components/Diff.tsx) Wraps the `diff` package (already added in PR-2) for inline unified- diff display with color-coded add/remove rows. Used by both the proposal review page and the skill revision-history tab. ### New pages (src/pages/) - Dashboard.tsx — at-a-glance home. Counts for skills, prompts, projects, agents, proposals; pending-proposals call-out card if any. - Skills.tsx — list view, separated into Global vs Project/Agent- scoped sections. - SkillDetail.tsx — name + semver + description; tabs for SKILL.md / Files / Metadata / History. History tab shows revisions with click-to-diff against the live body. - Proposals.tsx — queue with Pending/Approved/Rejected tabs. Pending count is highlighted in amber. - ProposalDetail.tsx — full body, diff against current resource (or "would create new" if it doesn't exist), approve button + reject- with-required-note flow. ### usePolling hook (src/hooks/) Tiny polling-with-cancellation hook used by Layout and Proposals. ### Layout rewrite (src/components/Layout.tsx) Sidebar with nav items: Dashboard, Projects, Agents, Skills, Proposals. Lucide icons. Active-route highlighting via NavLink. Pending-proposals warning badge on the Proposals item. ### Routes (src/App.tsx) New routes: /dashboard, /skills, /skills/:name, /proposals, /proposals/:id. Default redirects to /dashboard. ### API types (src/api.ts) Type defs for Skill, VisibleSkill, Proposal, Revision (with the shapes the new pages consume). ## Tests Existing 7 web tests still pass (Login + api). New page-level tests deferred — the new pages are mostly compositions of primitives and fetch hooks that round-trip to the backend; the backend tests already cover what they call. PR-7 polish can add render-and-click tests if coverage drift surfaces. ## Verification - `pnpm --filter @mcpctl/web build` clean, no warnings. - `pnpm test:run` whole monorepo: 162 test files / 2157 tests green. - Visual smoke deferred — needs a running mcpd to populate the fixtures. Manual smoke tested locally is the next step. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-07 17:54:55 +01:00
lucide-react@0.487.0:
resolution: {integrity: sha512-aKqhOQ+YmFnwq8dWgGjOuLc8V1R9/c/yOd+zDY4+ohsR2Jo05lSGc3WsstYPIzcTpeosN7LoCkLReUUITvaIvw==}
peerDependencies:
react: ^16.5.1 || ^17.0.0 || ^18.0.0 || ^19.0.0
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
lz-string@1.5.0:
resolution: {integrity: sha512-h5bgJWpxJNswbU7qCrV0tIKQCaS3blPDrqKWx+QxzuzL1zGUzij9XCWLrSLsJPu5t+eWA/ycetzYAO5IOMcWAQ==}
hasBin: true
first commit
2026-02-21 03:10:39 +00:00
magic-string@0.30.21:
resolution: {integrity: sha512-vd2F4YUyEXKGcLHoq+TEyCjxueSeHnFxyyjNp80yg0XV4vUhnDer/lvvlqM/arB5bXQN5K2/3oinyCRyx8T2CQ==}
magicast@0.5.2:
resolution: {integrity: sha512-E3ZJh4J3S9KfwdjZhe2afj6R9lGIN5Pher1pF39UGrXRqq/VDaGVIGN13BjHd2u8B61hArAGOnso7nBOouW3TQ==}
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
make-dir@3.1.0:
resolution: {integrity: sha512-g3FeP20LNwhALb/6Cz6Dd4F2ngze0jz7tbzrD2wAV+o9FeNHe4rL+yK2md0J/fiSf1sa1ADhXqi5+oVwOM/eGw==}
engines: {node: '>=8'}
first commit
2026-02-21 03:10:39 +00:00
make-dir@4.0.0:
resolution: {integrity: sha512-hXdUTZYIVOt1Ex//jAQi+wTZZpUpwBj/0QsOzqegb3rGMMeJiSEu5xLHnYfBrRV4RH2+OCSOO95Is/7x1WJ4bw==}
engines: {node: '>=10'}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
make-fetch-happen@15.0.6:
resolution: {integrity: sha512-Je0fLJ0F5atA7F+eIlLzk+Wkcl57JDf4kf+EW8xiP5E31xOQxkIxTbgf1Oi1Lw9tRI9UEMRdI5Vz2xTzoNU1Jw==}
engines: {node: ^20.17.0 || >=22.9.0}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
marked@14.0.0:
resolution: {integrity: sha512-uIj4+faQ+MgHgwUW1l2PsPglZLOLOT1uErt06dAPtx2kjteLAkbsd/0FiYg/MGS+i7ZKLb7w2WClxHkzOOuryQ==}
engines: {node: '>= 18'}
hasBin: true
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
marked@17.0.1:
resolution: {integrity: sha512-boeBdiS0ghpWcSwoNm/jJBwdpFaMnZWRzjA6SkUMYb40SVaN1x7mmfGKp0jvexGcx+7y2La5zRZsYFZI6Qpypg==}
engines: {node: '>= 20'}
hasBin: true
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
marked@18.0.5:
resolution: {integrity: sha512-S6GcvALHg6K4ohtu4E7x0a1AqhAjp6cV8KhLSyN9qVapnzJkusVBxZRcIU9AeYsbe6P1hKDusSbEOzGyyuce6w==}
engines: {node: '>= 20'}
hasBin: true
first commit
2026-02-21 03:10:39 +00:00
math-intrinsics@1.1.0:
resolution: {integrity: sha512-/IXtbwEk5HTPyEwyKX6hGkYXxM9nbj64B+ilVJnC/R6B0pH5G4V3b0pVbL7DBj4tkhBAppbQUlf6F6Xl9LHu1g==}
engines: {node: '>= 0.4'}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
mdn-data@2.27.1:
resolution: {integrity: sha512-9Yubnt3e8A0OKwxYSXyhLymGW4sCufcLG6VdiDdUGVkPhpqLxlvP5vl1983gQjJl3tqbrM731mjaZaP68AgosQ==}
first commit
2026-02-21 03:10:39 +00:00
media-typer@1.1.0:
resolution: {integrity: sha512-aisnrDP4GNe06UcKFnV5bfMNPBUw4jsLGaWwWfnH3v02GnBuXX2MCVn5RbrWo0j3pczUilYblq7fQ7Nw2t5XKw==}
engines: {node: '>= 0.8'}
merge-descriptors@2.0.0:
resolution: {integrity: sha512-Snk314V5ayFLhp3fkUREub6WtjBfPdCPY1Ln8/8munuLuiYhsABgBVWsozAG+MWMbVEvcdcpbi9R7ww22l9Q3g==}
engines: {node: '>=18'}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
merge-stream@2.0.0:
resolution: {integrity: sha512-abv/qOcuPfk3URPfDzmZU1LKmuw8kT+0nIHvKrKgFrwifol/doWcdA4ZqsWQ8ENrFKkd67Mfpo/LovbIUsbt3w==}
feat: add Kubernetes orchestrator for MCP server pod management mcpd can now deploy MCP server instances as Kubernetes pods instead of Docker containers. Set MCPD_ORCHESTRATOR=kubernetes to enable. - Add @kubernetes/client-node with thin wrapper (context enforcement via MCPD_K8S_CONTEXT to prevent multi-cluster mishaps) - Rewrite KubernetesOrchestrator: pod CRUD, pod IP extraction, exec via SPDY (one-shot + interactive), log streaming - Manifest generator: stdin:true for STDIO servers, args (not command) to preserve runner image entrypoint, security hardening - Orchestrator selection in main.ts via MCPD_ORCHESTRATOR env var - 25 unit tests for k8s orchestrator, all 624 tests pass Tested end-to-end on local k3s: - mcpd deployed via Pulumi, creates pods in mcpctl-servers namespace - NetworkPolicy verified: only mcpd can reach MCP server pods - Python runner (uvx) successfully runs aws-documentation-mcp-server Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-08 01:55:13 +01:00
mime-db@1.52.0:
resolution: {integrity: sha512-sPU4uV7dYlvtWJxwwxHD0PuihVNiE7TyAbQ5SWxDCB9mUYvOgroQOwYQQOKPJ8CIbE+1ETVlOoK1UC2nU3gYvg==}
engines: {node: '>= 0.6'}
first commit
2026-02-21 03:10:39 +00:00
mime-db@1.54.0:
resolution: {integrity: sha512-aU5EJuIN2WDemCcAp2vFBfp/m4EAhWJnUNSSw0ixs7/kXbd6Pg64EmwJkNdFhB8aWt1sH2CTXrLxo/iAGV3oPQ==}
engines: {node: '>= 0.6'}
feat: add Kubernetes orchestrator for MCP server pod management mcpd can now deploy MCP server instances as Kubernetes pods instead of Docker containers. Set MCPD_ORCHESTRATOR=kubernetes to enable. - Add @kubernetes/client-node with thin wrapper (context enforcement via MCPD_K8S_CONTEXT to prevent multi-cluster mishaps) - Rewrite KubernetesOrchestrator: pod CRUD, pod IP extraction, exec via SPDY (one-shot + interactive), log streaming - Manifest generator: stdin:true for STDIO servers, args (not command) to preserve runner image entrypoint, security hardening - Orchestrator selection in main.ts via MCPD_ORCHESTRATOR env var - 25 unit tests for k8s orchestrator, all 624 tests pass Tested end-to-end on local k3s: - mcpd deployed via Pulumi, creates pods in mcpctl-servers namespace - NetworkPolicy verified: only mcpd can reach MCP server pods - Python runner (uvx) successfully runs aws-documentation-mcp-server Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-08 01:55:13 +01:00
mime-types@2.1.35:
resolution: {integrity: sha512-ZDY+bPm5zTTF+YpCrAU9nK0UgICYPT0QtT1NZWFv4s++TNkcgVaT0g6+4R2uI4MjQjzysHB1zxuWL50hzaeXiw==}
engines: {node: '>= 0.6'}
first commit
2026-02-21 03:10:39 +00:00
mime-types@3.0.2:
resolution: {integrity: sha512-Lbgzdk0h4juoQ9fCKXW4by0UJqj+nOOrI9MJ1sSj4nI8aI2eo1qmvQEie4VD1glsS250n15LsWsYtCugiStS5A==}
engines: {node: '>=18'}
feat(mcpd+deploy): serve web UI at /ui + smoke tests + docs (Stage 6) The closing stage. mcpd now hosts the Stage 5 SPA, the Docker image bundles the build artifact, a smoke test exercises the personality HTTP surface end-to-end, and the user-facing docs spell out the mental model. mcpd: - Add @fastify/static dep. - New routes/web-ui.ts: registers /ui/* against a static bundle. Looks for the bundle at $MCPD_WEB_ROOT, then /usr/share/mcpd/web (the Docker image path), then a dev-tree fallback. Logs and skips cleanly if missing — API-only deploys keep working. - SPA fallback: any /ui/<path> that doesn't match a file falls through to index.html so direct hits to react-router URLs work. - /ui/* falls through to `kind: skip` in mapUrlToPermission, so the static assets are served unauthenticated. Each API call from the SPA still carries the bearer token. Deploy: - Dockerfile.mcpd builds the @mcpctl/web bundle in the same builder stage and copies dist/ to /usr/share/mcpd/web in the runtime image. Smoke (personality.smoke.test.ts): - Live mcpd flow: create secret/llm/agent/personality, attach an agent-direct prompt, verify the binding listing, reject double- attach (409) + foreign-agent prompt (400), set defaultPersonality by name, detach + delete cleanup. Docs: - New docs/personalities.md: VLAN-on-ethernet model, system-block ordering table, three prompt scopes, CLI walkthrough, web UI walkthrough, full API surface, RBAC notes. - agents.md and chat.md cross-link. - README's Agents section gains a Personalities subsection. Test count after Stage 6: mcpd: 801/801 cli: 430/430 web: 7/7 db: 58/62 (4 pre-existing) Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:48:43 +01:00
mime@3.0.0:
resolution: {integrity: sha512-jSCU7/VB1loIWBZe14aEYHU/+1UMEHoaO7qxCOVJOw9GgH72VAWppxNcjU+x9a2k3GSIBXNKxXQFqRvvZ7vr3A==}
engines: {node: '>=10.0.0'}
hasBin: true
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
mimic-fn@2.1.0:
resolution: {integrity: sha512-OqbOk5oEQeAZ8WXWydlu9HJjz9WVdEIvamMCcXmuqUYjTknH/sqsWvhQ3vgwKFRR1HpjvNBKQ37nbJgYzGqGcg==}
engines: {node: '>=6'}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
min-indent@1.0.1:
resolution: {integrity: sha512-I9jwMn07Sy/IwOj3zVkVik2JTvgpaykDZEigL6Rx6N9LbMywwUSMtxET+7lVoDLLd3O3IXwJwvuuns8UB/HeAg==}
engines: {node: '>=4'}
first commit
2026-02-21 03:10:39 +00:00
minimatch@10.2.2:
resolution: {integrity: sha512-+G4CpNBxa5MprY+04MbgOw1v7So6n5JY166pFi9KfYwT78fxScCeSNQSNzp6dpPSW2rONOps6Ocam1wFhCgoVw==}
engines: {node: 18 || 20 || >=22}
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
minimatch@10.2.5:
resolution: {integrity: sha512-MULkVLfKGYDFYejP07QOurDLLQpcjk7Fw+7jXS2R2czRQzR56yHRveU5NDJEOviH+hETZKSkIk5c+T23GjFUMg==}
engines: {node: 18 || 20 || >=22}
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
minimatch@3.1.2:
resolution: {integrity: sha512-J7p63hRiAjw1NDEww1W7i37+ByIrOWO5XQQAzZ3VOcL0PNybwpfmV/N05zFAzwQ9USyEcX6t3UO+K5aqBQOIHw==}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
minimatch@8.0.7:
resolution: {integrity: sha512-V+1uQNdzybxa14e/p00HZnQNNcTjnRJjDxg2V8wtkjFctq4M7hXFws4oekyTP0Jebeq7QYtpFyOeBAjc88zvYg==}
engines: {node: '>=16 || 14 >=14.17'}
first commit
2026-02-21 03:10:39 +00:00
minimatch@9.0.5:
resolution: {integrity: sha512-G6T0ZX48xgozx7587koeX9Ys2NYy6Gmv//P89sEte9V9whIapMNF4idKxnW2QtCcLiTWlb/wfCabAtAFWhhBow==}
engines: {node: '>=16 || 14 >=14.17'}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
minimist@1.2.8:
resolution: {integrity: sha512-2yyAR8qBkN3YuheJanUpWC5U3bb5osDywNB8RzDVlDwDHbocAJveqqj1u8+SVD7jkWT4yvsHCpWqqWqAxb0zCA==}
minipass-collect@2.0.1:
resolution: {integrity: sha512-D7V8PO9oaz7PWGLbCACuI1qEOsq7UKfLotx/C0Aet43fCUB/wfQ7DYeq2oR/svFJGYDHPr38SHATeaj/ZoKHKw==}
engines: {node: '>=16 || 14 >=14.17'}
minipass-fetch@5.0.2:
resolution: {integrity: sha512-2d0q2a8eCi2IRg/IGubCNRJoYbA1+YPXAzQVRFmB45gdGZafyivnZ5YSEfo3JikbjGxOdntGFvBQGqaSMXlAFQ==}
engines: {node: ^20.17.0 || >=22.9.0}
minipass-flush@1.0.7:
resolution: {integrity: sha512-TbqTz9cUwWyHS2Dy89P3ocAGUGxKjjLuR9z8w4WUTGAVgEj17/4nhgo2Du56i0Fm3Pm30g4iA8Lcqctc76jCzA==}
engines: {node: '>= 8'}
minipass-pipeline@1.2.4:
resolution: {integrity: sha512-xuIq7cIOt09RPRJ19gdi4b+RiNvDFYe5JH+ggNvBqGqpQXcru3PcRmOZuHBKWK1Txf9+cQ+HMVN4d6z46LZP7A==}
engines: {node: '>=8'}
minipass-sized@2.0.0:
resolution: {integrity: sha512-zSsHhto5BcUVM2m1LurnXY6M//cGhVaegT71OfOXoprxT6o780GZd792ea6FfrQkuU4usHZIUczAQMRUE2plzA==}
engines: {node: '>=8'}
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
minipass@3.3.6:
resolution: {integrity: sha512-DxiNidxSEK+tHG6zOIklvNOwm3hvCrbUrdtzY74U6HKTJxvIDfOUL5W5P2Ghd3DTkhhKPYGqeNUIh5qcM4YBfw==}
engines: {node: '>=8'}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
minipass@4.2.8:
resolution: {integrity: sha512-fNzuVyifolSLFL4NzpF+wEF4qrgqaaKX0haXPQEdQ7NKAN+WecoKMHV09YcuL/DHxrUsYQOK3MiuDf7Ip2OXfQ==}
engines: {node: '>=8'}
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
minipass@5.0.0:
resolution: {integrity: sha512-3FnjYuehv9k6ovOEbyOswadCDPX1piCfhV8ncmYtHOjuPwylVWsghTLo7rabjC3Rx5xD4HDx8Wm1xnMF7S5qFQ==}
engines: {node: '>=8'}
first commit
2026-02-21 03:10:39 +00:00
minipass@7.1.3:
resolution: {integrity: sha512-tEBHqDnIoM/1rXME1zgka9g6Q2lcoCkxHLuc7ODJ5BxbP5d4c2Z5cGgtXAku59200Cx7diuHTOYfSBD8n6mm8A==}
engines: {node: '>=16 || 14 >=14.17'}
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
minizlib@2.1.2:
resolution: {integrity: sha512-bAxsR8BVfj60DWXHE3u30oHzfl4G7khkSuPW+qvpd7jFRHm7dLxOjUk1EHACJ/hxLY8phGJ0YhYHZo7jil7Qdg==}
engines: {node: '>= 8'}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
minizlib@3.1.0:
resolution: {integrity: sha512-KZxYo1BUkWD2TVFLr0MQoM8vUUigWD3LlD83a/75BqC+4qE0Hb1Vo5v1FgcfaNXvfXzr+5EhQ6ing/CaBijTlw==}
engines: {node: '>= 18'}
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
mkdirp-classic@0.5.3:
resolution: {integrity: sha512-gKLcREMhtuZRwRAfqP3RFW+TK4JqApVBtOIftVgjuABpAtpxhPGaDcfvbhNvD0B8iD1oUr/txX35NjcaY6Ns/A==}
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
mkdirp@1.0.4:
resolution: {integrity: sha512-vVqVZQyf3WLx2Shd0qJ9xuvqgAyKPLAiqITEtqW0oIUjzo3PePDd6fW9iFz30ef7Ysp/oiWqbhszeGWW2T6Gzw==}
engines: {node: '>=10'}
hasBin: true
first commit
2026-02-21 03:10:39 +00:00
mnemonist@0.40.0:
resolution: {integrity: sha512-kdd8AFNig2AD5Rkih7EPCXhu/iMvwevQFX/uEiGhZyPZi7fHqOoF4V4kHLpCfysxXMgQ4B52kdPMCwARshKvEg==}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
module-details-from-path@1.0.4:
resolution: {integrity: sha512-EGWKgxALGMgzvxYF1UyGTy0HXX/2vHLkw6+NvDKW2jypWbHpjQuj4UMcqQWXHERJhVGKikolT06G3bcKe4fi7w==}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
monaco-editor@0.55.1:
resolution: {integrity: sha512-jz4x+TJNFHwHtwuV9vA9rMujcZRb0CEilTEwG2rRSpe/A7Jdkuj8xPKttCgOh+v/lkHy7HsZ64oj+q3xoAFl9A==}
first commit
2026-02-21 03:10:39 +00:00
ms@2.1.3:
resolution: {integrity: sha512-6FlzubTLZG3J2a/NVCAleEhjzq5oxgHyaCU9yYXvcLsvoVaHJq/s5xXI6/XXP6tz7R9xAOtHnSO/tXtF3WRTlA==}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
msgpackr-extract@3.0.4:
resolution: {integrity: sha512-4kmO/MdyUIkLIvTPr8VHLil4AtoKIoniWPIEk5+CDy0xnWC84azhSFmuJ7PxZdsYtiP5kEeQsORAVIeMgxT+Hw==}
hasBin: true
msgpackr@2.0.5:
resolution: {integrity: sha512-cef05H/dSYpLpqp3sj/qyZh5vhUYCalnaLO7j1yOmpsR0y/XwLVtK7r5gn+U/F7CTEfMowcGhlUQJDLcLf7jcA==}
multipasta@0.2.8:
resolution: {integrity: sha512-ZPWuMKyv0cSO29f7hozp+k6+crZbQijV8ipMvxNxRf2SwtYGTX1ZX89Kd20VV4H9Znonx+EQn+iy1wGQsJ+b+Q==}
first commit
2026-02-21 03:10:39 +00:00
mute-stream@2.0.0:
resolution: {integrity: sha512-WWdIxpyjEn+FhQJQQv9aQAYlHoNVdzIzUySNV1gHUPDSdZJ3yZn7pAAbQcV7B56Mvu881q9FZV+0Vx2xC44VWA==}
engines: {node: ^18.17.0 || >=20.5.0}
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
nan@2.25.0:
resolution: {integrity: sha512-0M90Ag7Xn5KMLLZ7zliPWP3rT90P6PN+IzVFS0VqmnPktBk3700xUVv8Ikm9EUaUE5SDWdp/BIxdENzVznpm1g==}
first commit
2026-02-21 03:10:39 +00:00
nanoid@3.3.11:
resolution: {integrity: sha512-N8SpfPUnUp1bK+PMYW8qSWdl9U+wwNWI4QKxOYDy9JAro3WMX7p2OeVRF9v+347pnakNevPmiHhNmZ2HbFA76w==}
engines: {node: ^10 || ^12 || ^13.7 || ^14 || >=15.0.1}
hasBin: true
natural-compare@1.4.0:
resolution: {integrity: sha512-OWND8ei3VtNC9h7V60qff3SVobHr996CTwgxubgyQYEpg290h9J0buyECNNJexkFm5sOajh5G116RYA1c8ZMSw==}
negotiator@1.0.0:
resolution: {integrity: sha512-8Ofs/AUQh8MaEcrlq5xOX0CQ9ypTF5dl78mjlMNfOK08fzpgTHQRQPBxcPlEtIw0yRpws+Zo/3r+5WRby7u3Gg==}
engines: {node: '>= 0.6'}
feat(web): bold redesign — Tailwind v4 + shadcn-style primitives + Skills/Proposals/Revisions UI Phase 6 of the Skills + Revisions + Proposals work. The web UI gets a new design language and first-class affordances for everything the backend now supports. ## Visual direction - Tailwind v4 with custom @theme block (oklch tokens). Dark-mode-only (internal tool — light mode doubles QA surface). - Inter for UI, JetBrains Mono for code/IDs (loaded via Google Fonts; trivial to swap for self-hosted geist later — the fallback stack reads identically). - Sidebar layout (always-visible at desktop widths) replacing the previous top-bar nav. Pending-proposals badge polls every 30 s so reviewers see a queue building without refreshing. - Lucide icons throughout. - Spacing and radii on Tailwind defaults. Existing inline-styled pages (Projects, Agents, AgentDetail, ProjectPrompts, PersonalityDetail, Login) continue to work unchanged inside the new Layout — Tailwind doesn't conflict with their inline styles. A follow-up can migrate them incrementally. ## What's added ### Build infra (src/web/) - package.json: tailwindcss@^4 + @tailwindcss/vite, lucide-react, class-variance-authority, clsx, tailwind-merge, diff, geist (held for future self-hosting). - vite.config.ts: registers the @tailwindcss/vite plugin. - src/index.css: Tailwind import + @theme tokens + @layer base. - src/main.tsx: imports index.css. - src/lib/utils.ts: shadcn-style cn() helper. ### shadcn-style primitives (src/components/ui/) Hand-written rather than generated via `npx shadcn` so the repo doesn't depend on a CLI tool that needs interactive runtime: - button.tsx — variants: primary / secondary / ghost / danger / link; sizes: sm / md / lg / icon. - card.tsx — Card + Header/Title/Description/Content/Footer subparts. - badge.tsx — variants: default / info / success / warning / danger / outline. - input.tsx — Input + Textarea + Label. - tabs.tsx — no-dep accessible Tabs (no Radix needed for our use). - separator.tsx — h/v separator with role=separator. ### Diff component (src/components/Diff.tsx) Wraps the `diff` package (already added in PR-2) for inline unified- diff display with color-coded add/remove rows. Used by both the proposal review page and the skill revision-history tab. ### New pages (src/pages/) - Dashboard.tsx — at-a-glance home. Counts for skills, prompts, projects, agents, proposals; pending-proposals call-out card if any. - Skills.tsx — list view, separated into Global vs Project/Agent- scoped sections. - SkillDetail.tsx — name + semver + description; tabs for SKILL.md / Files / Metadata / History. History tab shows revisions with click-to-diff against the live body. - Proposals.tsx — queue with Pending/Approved/Rejected tabs. Pending count is highlighted in amber. - ProposalDetail.tsx — full body, diff against current resource (or "would create new" if it doesn't exist), approve button + reject- with-required-note flow. ### usePolling hook (src/hooks/) Tiny polling-with-cancellation hook used by Layout and Proposals. ### Layout rewrite (src/components/Layout.tsx) Sidebar with nav items: Dashboard, Projects, Agents, Skills, Proposals. Lucide icons. Active-route highlighting via NavLink. Pending-proposals warning badge on the Proposals item. ### Routes (src/App.tsx) New routes: /dashboard, /skills, /skills/:name, /proposals, /proposals/:id. Default redirects to /dashboard. ### API types (src/api.ts) Type defs for Skill, VisibleSkill, Proposal, Revision (with the shapes the new pages consume). ## Tests Existing 7 web tests still pass (Login + api). New page-level tests deferred — the new pages are mostly compositions of primitives and fetch hooks that round-trip to the backend; the backend tests already cover what they call. PR-7 polish can add render-and-click tests if coverage drift surfaces. ## Verification - `pnpm --filter @mcpctl/web build` clean, no warnings. - `pnpm test:run` whole monorepo: 162 test files / 2157 tests green. - Visual smoke deferred — needs a running mcpd to populate the fixtures. Manual smoke tested locally is the next step. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-07 17:54:55 +01:00
next@16.2.5:
resolution: {integrity: sha512-TkVTm9F2WEulkgGljm4wPwNgvCCWCVw6StUHsZb8WZpHFRjepoUWg3d7L4IMg7IyjcJ4Co9eVhpro8e8O+KarQ==}
engines: {node: '>=20.9.0'}
hasBin: true
peerDependencies:
'@opentelemetry/api': ^1.1.0
'@playwright/test': ^1.51.1
babel-plugin-react-compiler: '*'
react: ^18.2.0 || 19.0.0-rc-de68d2f4-20241204 || ^19.0.0
react-dom: ^18.2.0 || 19.0.0-rc-de68d2f4-20241204 || ^19.0.0
sass: ^1.3.0
peerDependenciesMeta:
'@opentelemetry/api':
optional: true
'@playwright/test':
optional: true
babel-plugin-react-compiler:
optional: true
sass:
optional: true
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
node-addon-api@5.1.0:
resolution: {integrity: sha512-eh0GgfEkpnoWDq+VY8OyvYhFEzBk6jIYbRKdIlyTiAXIVJ8PyBaKb0rp7oDtoddbdoHWhq8wwr+XZ81F1rpNdA==}
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
node-domexception@1.0.0:
resolution: {integrity: sha512-/jKZoMpw0F8GRwl4/eLROPA3cfcXtLApP0QzLmUT/HuPCZWyB7IY9ZrMeKw2O/nFIqPQB3PVM9aYm0F312AXDQ==}
engines: {node: '>=10.5.0'}
deprecated: Use your platform's native DOMException instead
first commit
2026-02-21 03:10:39 +00:00
node-fetch-native@1.6.7:
resolution: {integrity: sha512-g9yhqoedzIUm0nTnTqAQvueMPVOuIY16bqgAJJC8XOOubYFNwz6IER9qs0Gq2Xd0+CecCKFjtdDTMA4u4xG06Q==}
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
node-fetch@2.7.0:
resolution: {integrity: sha512-c4FRfUm/dbcWZ7U+1Wq0AwCyFL+3nt2bEw05wfxSz+DWpWsitgmSgYmy2dQdWyKC1694ELPqMs/YzUSNozLt8A==}
engines: {node: 4.x || >=6.0.0}
peerDependencies:
encoding: ^0.1.0
peerDependenciesMeta:
encoding:
optional: true
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
node-fetch@3.3.2:
resolution: {integrity: sha512-dRB78srN/l6gqWulah9SrxeYnxeddIG30+GOqK/9OlLVyLg3HPnr6SqOWTWOXKRwC2eGYCkZ59NNuSgvSrpgOA==}
engines: {node: ^12.20.0 || ^14.13.1 || >=16.0.0}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
node-gyp-build-optional-packages@5.2.2:
resolution: {integrity: sha512-s+w+rBWnpTMwSFbaE0UXsRlg7hU4FjekKU4eyAih5T8nJuNZT1nNsskXpxmeqSK9UzkBl6UgRlnKc8hz8IEqOw==}
hasBin: true
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
node-gyp@12.4.0:
resolution: {integrity: sha512-OMcPNvqTCFUnNaBlmdgq+lfNqY7gTiSmNRDjY3uAXRyudeKZEZxu3CLtjMQrx4zZxCX2b/mpNqTtwuCJgXhHkw==}
engines: {node: ^20.17.0 || >=22.9.0}
hasBin: true
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
node-releases@2.0.38:
resolution: {integrity: sha512-3qT/88Y3FbH/Kx4szpQQ4HzUbVrHPKTLVpVocKiLfoYvw9XSGOX2FmD2d6DrXbVYyAQTF2HeF6My8jmzx7/CRw==}
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
nopt@5.0.0:
resolution: {integrity: sha512-Tbj67rffqceeLpcRXrT7vKAN8CwfPeIBgM7E6iBkmKLV7bEMwpGgYLGv0jACUsECaa/vuxP0IjEont6umdMgtQ==}
engines: {node: '>=6'}
hasBin: true
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
nopt@9.0.0:
resolution: {integrity: sha512-Zhq3a+yFKrYwSBluL4H9XP3m3y5uvQkB/09CwDruCiRmR/UJYnn9W4R48ry0uGC70aeTPKLynBtscP9efFFcPw==}
engines: {node: ^20.17.0 || >=22.9.0}
hasBin: true
normalize-package-data@6.0.2:
resolution: {integrity: sha512-V6gygoYb/5EmNI+MEGrWkC+e6+Rr7mTmfHrxDbLzxQogBkgzo76rkok0Am6thgSF7Mv2nLOajAJj5vDJZEFn7g==}
engines: {node: ^16.14.0 || >=18.0.0}
npm-bundled@5.0.0:
resolution: {integrity: sha512-JLSpbzh6UUXIEoqPsYBvVNVmyrjVZ1fzEFbqxKkTJQkWBO3xFzFT+KDnSKQWwOQNbuWRwt5LSD6HOTLGIWzfrw==}
engines: {node: ^20.17.0 || >=22.9.0}
npm-install-checks@8.0.0:
resolution: {integrity: sha512-ScAUdMpyzkbpxoNekQ3tNRdFI8SJ86wgKZSQZdUxT+bj0wVFpsEMWnkXP0twVe1gJyNF5apBWDJhhIbgrIViRA==}
engines: {node: ^20.17.0 || >=22.9.0}
npm-normalize-package-bin@5.0.0:
resolution: {integrity: sha512-CJi3OS4JLsNMmr2u07OJlhcrPxCeOeP/4xq67aWNai6TNWWbTrlNDgl8NcFKVlcBKp18GPj+EzbNIgrBfZhsag==}
engines: {node: ^20.17.0 || >=22.9.0}
npm-package-arg@13.0.2:
resolution: {integrity: sha512-IciCE3SY3uE84Ld8WZU23gAPPV9rIYod4F+rc+vJ7h7cwAJt9Vk6TVsK60ry7Uj3SRS3bqRRIGuTp9YVlk6WNA==}
engines: {node: ^20.17.0 || >=22.9.0}
npm-packlist@10.0.4:
resolution: {integrity: sha512-uMW73iajD8hiH4ZBxEV3HC+eTnppIqwakjOYuvgddnalIw2lJguKviK1pcUJDlIWm1wSJkchpDZDSVVsZEYRng==}
engines: {node: ^20.17.0 || >=22.9.0}
npm-pick-manifest@11.0.3:
resolution: {integrity: sha512-buzyCfeoGY/PxKqmBqn1IUJrZnUi1VVJTdSSRPGI60tJdUhUoSQFhs0zycJokDdOznQentgrpf8LayEHyyYlqQ==}
engines: {node: ^20.17.0 || >=22.9.0}
npm-registry-fetch@19.1.1:
resolution: {integrity: sha512-TakBap6OM1w0H73VZVDf44iFXsOS3h+L4wVMXmbWOQroZgFhMch0juN6XSzBNlD965yIKvWg2dfu7NSiaYLxtw==}
engines: {node: ^20.17.0 || >=22.9.0}
npm-run-path@4.0.1:
resolution: {integrity: sha512-S48WzZW777zhNIrn7gxOlISNAqi9ZC/uQFnRdbeIHhZhCA6UqpkOT8T1G7BvfdgP4Er8gF4sUbaS0i7QvIfCWw==}
engines: {node: '>=8'}
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
npmlog@5.0.1:
resolution: {integrity: sha512-AqZtDUWOMKs1G/8lwylVjrdYgqA4d9nu8hc+0gzRxlDb1I10+FHBGMXs6aiQHFdCUUlqH99MUMuLfzWDNDtfxw==}
deprecated: This package is no longer supported.
first commit
2026-02-21 03:10:39 +00:00
nypm@0.6.5:
resolution: {integrity: sha512-K6AJy1GMVyfyMXRVB88700BJqNUkByijGJM8kEHpLdcAt+vSQAVfkWWHYzuRXHSY6xA2sNc5RjTj0p9rE2izVQ==}
engines: {node: '>=18'}
hasBin: true
feat: add Kubernetes orchestrator for MCP server pod management mcpd can now deploy MCP server instances as Kubernetes pods instead of Docker containers. Set MCPD_ORCHESTRATOR=kubernetes to enable. - Add @kubernetes/client-node with thin wrapper (context enforcement via MCPD_K8S_CONTEXT to prevent multi-cluster mishaps) - Rewrite KubernetesOrchestrator: pod CRUD, pod IP extraction, exec via SPDY (one-shot + interactive), log streaming - Manifest generator: stdin:true for STDIO servers, args (not command) to preserve runner image entrypoint, security hardening - Orchestrator selection in main.ts via MCPD_ORCHESTRATOR env var - 25 unit tests for k8s orchestrator, all 624 tests pass Tested end-to-end on local k3s: - mcpd deployed via Pulumi, creates pods in mcpctl-servers namespace - NetworkPolicy verified: only mcpd can reach MCP server pods - Python runner (uvx) successfully runs aws-documentation-mcp-server Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-08 01:55:13 +01:00
oauth4webapi@3.8.5:
resolution: {integrity: sha512-A8jmyUckVhRJj5lspguklcl90Ydqk61H3dcU0oLhH3Yv13KpAliKTt5hknpGGPZSSfOwGyraNEFmofDYH+1kSg==}
first commit
2026-02-21 03:10:39 +00:00
object-assign@4.1.1:
resolution: {integrity: sha512-rJgTQnkUnH1sFw8yT6VSU3zD3sWmu6sZhIseY8VX+GRu3P6F7Fu+JNDoXfklElbLJSnc3FUQHVe4cU5hj+BcUg==}
engines: {node: '>=0.10.0'}
object-inspect@1.13.4:
resolution: {integrity: sha512-W67iLl4J2EXEGTbfeHCffrjDfitvLANg0UlX3wFUUSTx92KXRFegMHUVgSqE+wvhAbi4WqjGg9czysTV2Epbew==}
engines: {node: '>= 0.4'}
obliterator@2.0.5:
resolution: {integrity: sha512-42CPE9AhahZRsMNslczq0ctAEtqk8Eka26QofnqC346BZdHDySk3LWka23LI7ULIw11NmltpiLagIq8gBozxTw==}
obug@2.1.1:
resolution: {integrity: sha512-uTqF9MuPraAQ+IsnPf366RG4cP9RtUi7MLO1N3KEc+wb0a6yKpeL0lmk2IB1jY5KHPAlTc6T/JRdC/YqxHNwkQ==}
ohash@2.0.11:
resolution: {integrity: sha512-RdR9FQrFwNBNXAr4GixM8YaRZRJ5PUWbKYbE5eOsrwAjJW0q2REGcf79oYPsLyskQCZG1PLN+S/K1V00joZAoQ==}
on-exit-leak-free@2.1.2:
resolution: {integrity: sha512-0eJJY6hXLGf1udHwfNftBqH+g73EU4B504nZeKpz1sYRKafAghwxEJunB2O7rDZkL4PGfsMVnTXZ2EjibbqcsA==}
engines: {node: '>=14.0.0'}
on-finished@2.4.1:
resolution: {integrity: sha512-oVlzkg3ENAhCk2zdv7IJwd/QUD4z2RxRwpkcGY8psCVcCYZNq4wYnVWALHM+brtuJjePWiYF/ClmuDr8Ch5+kg==}
engines: {node: '>= 0.8'}
once@1.4.0:
resolution: {integrity: sha512-lNaJgI+2Q5URQBkccEKHTQOPaXdUxnZZElQTZY0MFUAuaEqe1E+Nyvgdz/aIyNi6Z9MzO5dv1H8n58/GELp3+w==}
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
onetime@5.1.2:
resolution: {integrity: sha512-kbpaSSGJTWdAY5KPVeMOKXSrPtr8C8C7wodJbcsd51jRnmD+GZu8Y0VoU6Dm5Z4vWr0Ig/1NKuWRKf7j5aaYSg==}
engines: {node: '>=6'}
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
openai@6.26.0:
resolution: {integrity: sha512-zd23dbWTjiJ6sSAX6s0HrCZi41JwTA1bQVs0wLQPZ2/5o2gxOJA5wh7yOAUgwYybfhDXyhwlpeQf7Mlgx8EOCA==}
hasBin: true
peerDependencies:
ws: ^8.18.0
zod: ^3.25 || ^4.0
peerDependenciesMeta:
ws:
optional: true
zod:
optional: true
feat: add Kubernetes orchestrator for MCP server pod management mcpd can now deploy MCP server instances as Kubernetes pods instead of Docker containers. Set MCPD_ORCHESTRATOR=kubernetes to enable. - Add @kubernetes/client-node with thin wrapper (context enforcement via MCPD_K8S_CONTEXT to prevent multi-cluster mishaps) - Rewrite KubernetesOrchestrator: pod CRUD, pod IP extraction, exec via SPDY (one-shot + interactive), log streaming - Manifest generator: stdin:true for STDIO servers, args (not command) to preserve runner image entrypoint, security hardening - Orchestrator selection in main.ts via MCPD_ORCHESTRATOR env var - 25 unit tests for k8s orchestrator, all 624 tests pass Tested end-to-end on local k3s: - mcpd deployed via Pulumi, creates pods in mcpctl-servers namespace - NetworkPolicy verified: only mcpd can reach MCP server pods - Python runner (uvx) successfully runs aws-documentation-mcp-server Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-08 01:55:13 +01:00
openid-client@6.8.2:
resolution: {integrity: sha512-uOvTCndr4udZsKihJ68H9bUICrriHdUVJ6Az+4Ns6cW55rwM5h0bjVIzDz2SxgOI84LKjFyjOFvERLzdTUROGA==}
first commit
2026-02-21 03:10:39 +00:00
optionator@0.9.4:
resolution: {integrity: sha512-6IpQ7mKUxRcZNLIObR0hz7lxsapSSIYNZJwXPGeF0mTVqGKFIXj1DQcMoT22S3ROcLyY/rz0PWaWZ9ayWmad9g==}
engines: {node: '>= 0.8.0'}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
p-limit@2.3.0:
resolution: {integrity: sha512-//88mFWSJx8lxCzwdAABTJL2MyWB12+eIY7MDL2SqLmAkeKU9qxRvWuSyTjm3FUmpBEMuFfckAIqEaVGUDxb6w==}
engines: {node: '>=6'}
first commit
2026-02-21 03:10:39 +00:00
p-limit@3.1.0:
resolution: {integrity: sha512-TYOanM3wGwNGsZN2cVTYPArw454xnXj5qmWF1bEoAc4+cU/ol7GVh7odevjp1FNHduHc3KZMcFduxU5Xc6uJRQ==}
engines: {node: '>=10'}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
p-locate@3.0.0:
resolution: {integrity: sha512-x+12w/To+4GFfgJhBEpiDcLozRJGegY+Ei7/z0tSLkMmxGZNybVMSfWj9aJn8Z5Fc7dBUNJOOVgPv2H7IwulSQ==}
engines: {node: '>=6'}
first commit
2026-02-21 03:10:39 +00:00
p-locate@5.0.0:
resolution: {integrity: sha512-LaNjtRWUBY++zB5nE/NwcaoMylSPk+S+ZHNB1TzdbMJMny6dynpAGt7X/tl/QYq3TIeE6nxHppbo2LGymrG5Pw==}
engines: {node: '>=10'}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
p-map@7.0.5:
resolution: {integrity: sha512-e8vJF4XdVkzqqSHguEMz41mQO1wKwxKm5ENrUJQUu9kLDCtn83cxbyHZcszr4QC5zEA7WffRRC4gsTecC7J9oA==}
engines: {node: '>=18'}
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
p-retry@4.6.2:
resolution: {integrity: sha512-312Id396EbJdvRONlngUx0NydfrIQ5lsYu0znKVUzVvArzEIt08V1qhtyESbGVd1FGX7UKtiFp5uwKZdM8wIuQ==}
engines: {node: '>=8'}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
p-try@2.2.0:
resolution: {integrity: sha512-R4nPAVTAU0B9D35/Gk3uJf/7XYbQcyohSKdvAxIRSNghFl4e71hVoGnBNQz9cWaXxO2I10KTC+3jMdvvoKw6dQ==}
engines: {node: '>=6'}
first commit
2026-02-21 03:10:39 +00:00
package-json-from-dist@1.0.1:
resolution: {integrity: sha512-UEZIS3/by4OC8vL3P2dTXRETpebLI2NiI5vIrjaD/5UtrkFX/tNbwjTSRAGC/+7CAo2pIcBaRgWmcBBHcsaCIw==}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
pacote@21.5.1:
resolution: {integrity: sha512-KvcJ9iy3crysCsgqc4+PknH/w6jkrp8JN36mpZBPwNaDRwTfMZD37YzRazNstiZUOhuF5pno9f78n9mEJBavwg==}
engines: {node: ^20.17.0 || >=22.9.0}
hasBin: true
parse-conflict-json@5.0.1:
resolution: {integrity: sha512-ZHEmNKMq1wyJXNwLxyHnluPfRAFSIliBvbK/UiOceROt4Xh9Pz0fq49NytIaeaCUf5VR86hwQ/34FCcNU5/LKQ==}
engines: {node: ^20.17.0 || >=22.9.0}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
parse5@7.3.0:
resolution: {integrity: sha512-IInvU7fabl34qmi9gY8XOVxhYyMyuH2xUNpb2q8/Y+7552KlejkRvqvD19nMoUW/uQGGbqNpA6Tufu5FL5BZgw==}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
parse5@8.0.1:
resolution: {integrity: sha512-z1e/HMG90obSGeidlli3hj7cbocou0/wa5HacvI3ASx34PecNjNQeaHNo5WIZpWofN9kgkqV1q5YvXe3F0FoPw==}
first commit
2026-02-21 03:10:39 +00:00
parseurl@1.3.3:
resolution: {integrity: sha512-CiyeOxFT/JZyN5m0z9PfXw4SCBJ6Sygz1Dpl0wqjlhDEGGBP1GnsUVEL0p63hoG1fcj3fHynXi9NYO4nWOL+qQ==}
engines: {node: '>= 0.8'}
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
partial-json@0.1.7:
resolution: {integrity: sha512-Njv/59hHaokb/hRUjce3Hdv12wd60MtM9Z5Olmn+nehe0QDAsRtRbJPvJ0Z91TusF0SuZRIvnM+S4l6EIP8leA==}
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
patch-console@2.0.0:
resolution: {integrity: sha512-0YNdUceMdaQwoKce1gatDScmMo5pu/tfABfnzEqeG0gtTmd7mh/WcwgUjtAeOU7N8nFFlbQBnFK2gXW5fGvmMA==}
engines: {node: ^12.20.0 || ^14.13.1 || >=16.0.0}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
path-exists@3.0.0:
resolution: {integrity: sha512-bpC7GYwiDYQ4wYLe+FA8lhRjhQCMcQGuSgGGqDkg/QerRWw9CmGRT0iSOVRSZJ29NMLZgIzqaljJ63oaL4NIJQ==}
engines: {node: '>=4'}
first commit
2026-02-21 03:10:39 +00:00
path-exists@4.0.0:
resolution: {integrity: sha512-ak9Qy5Q7jYb2Wwcey5Fpvg2KoAc/ZIhLSLOSBmRmygPsGwkVVt0fZa0qrtMz+m6tJTAHfZQ8FnmB4MG4LWy7/w==}
engines: {node: '>=8'}
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
path-is-absolute@1.0.1:
resolution: {integrity: sha512-AVbw3UJ2e9bq64vSaS9Am0fje1Pa8pbGqTTsmXfaIiMpnr5DlDhfJOuLj9Sf95ZPVDAUerDfEk88MPmPe7UCQg==}
engines: {node: '>=0.10.0'}
first commit
2026-02-21 03:10:39 +00:00
path-key@3.1.1:
resolution: {integrity: sha512-ojmeN0qd+y0jszEtoY48r0Peq5dwMEkIlCOu6Q5f41lfkswXuKtYrhgoTpLnyIcHm24Uhqx+5Tqm2InSwLhE6Q==}
engines: {node: '>=8'}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
path-parse@1.0.7:
resolution: {integrity: sha512-LDJzPVEEEPR+y48z93A0Ed0yXb8pAByGWo/k5YYdYgpY2/2EsOsksJrq7lOHxryrVOn1ejG6oAp8ahvOIQD8sw==}
path-scurry@1.11.1:
resolution: {integrity: sha512-Xa4Nw17FS9ApQFJ9umLiJS4orGjm7ZzwUrwamcGQuHSzDyth9boKDaycYdDcZDuqYATXw4HFXgaqWTctW/v1HA==}
engines: {node: '>=16 || 14 >=14.18'}
first commit
2026-02-21 03:10:39 +00:00
path-scurry@2.0.2:
resolution: {integrity: sha512-3O/iVVsJAPsOnpwWIeD+d6z/7PmqApyQePUtCndjatj/9I5LylHvt5qluFaBT3I5h3r1ejfR056c+FCv+NnNXg==}
engines: {node: 18 || 20 || >=22}
path-to-regexp@8.3.0:
resolution: {integrity: sha512-7jdwVIRtsP8MYpdXSwOS0YdD0Du+qOoF/AEPIt88PcCFrZCzx41oxku1jD88hZBwbNUIEfpqvuhjFaMAqMTWnA==}
pathe@2.0.3:
resolution: {integrity: sha512-WUjGcAqP1gQacoQe+OBJsFA7Ld4DyXuUIjZ5cc75cLHvJ7dtNsTugphxIADwspS+AraAUePCKrSVtPLFj/F88w==}
perfect-debounce@1.0.0:
resolution: {integrity: sha512-xCy9V055GLEqoFaHoC1SoLIaLmWctgCUaBaWxDZ7/Zx4CTyX7cJQLJOok/orfjZAh9kEYpjJa4d0KcJmCbctZA==}
picocolors@1.1.1:
resolution: {integrity: sha512-xceH2snhtb5M9liqDsmEw56le376mTZkEX/jEb/RxNFyegNul7eNslCXP9FDj/Lcu0X8KEyMceP2ntpaHrDEVA==}
picomatch@4.0.3:
resolution: {integrity: sha512-5gTmgEY/sqK6gFXLIsQNH19lWb4ebPDLA4SdLP7dsWkIXHWlG66oPuVvXSGFPppYZz8ZDZq0dYYrbHfBCVUb1Q==}
engines: {node: '>=12'}
pino-abstract-transport@3.0.0:
resolution: {integrity: sha512-wlfUczU+n7Hy/Ha5j9a/gZNy7We5+cXp8YL+X+PG8S0KXxw7n/JXA3c46Y0zQznIJ83URJiwy7Lh56WLokNuxg==}
pino-std-serializers@7.1.0:
resolution: {integrity: sha512-BndPH67/JxGExRgiX1dX0w1FvZck5Wa4aal9198SrRhZjH3GxKQUKIBnYJTdj2HDN3UQAS06HlfcSbQj2OHmaw==}
pino@10.3.1:
resolution: {integrity: sha512-r34yH/GlQpKZbU1BvFFqOjhISRo1MNx1tWYsYvmj6KIRHSPMT2+yHOEb1SG6NMvRoHRF0a07kCOox/9yakl1vg==}
hasBin: true
pkce-challenge@5.0.1:
resolution: {integrity: sha512-wQ0b/W4Fr01qtpHlqSqspcj3EhBvimsdh0KlHhH8HRZnMsEa0ea2fTULOXOS9ccQr3om+GcGRk4e+isrZWV8qQ==}
engines: {node: '>=16.20.0'}
pkg-types@2.3.0:
resolution: {integrity: sha512-SIqCzDRg0s9npO5XQ3tNZioRY1uK06lA41ynBC1YmFTmnY6FjUjVt6s4LoADmwoig1qqD0oK8h1p/8mlMx8Oig==}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
pkg-up@3.1.0:
resolution: {integrity: sha512-nDywThFk1i4BQK4twPQ6TA4RT8bDY96yeuCVBWL3ePARCiEKDRSrNGbFIgUJpLp+XeIR65v8ra7WuJOFUBtkMA==}
engines: {node: '>=8'}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
postcss-selector-parser@7.1.4:
resolution: {integrity: sha512-HeP7D2wyhkR+XaK6v4W8oRF62Dsz4flyuczALJp61GckGm42u1saSSJ/0auvcBqxs3jMRFEcPK34At/0JBKdOg==}
engines: {node: '>=4'}
feat(web): bold redesign — Tailwind v4 + shadcn-style primitives + Skills/Proposals/Revisions UI Phase 6 of the Skills + Revisions + Proposals work. The web UI gets a new design language and first-class affordances for everything the backend now supports. ## Visual direction - Tailwind v4 with custom @theme block (oklch tokens). Dark-mode-only (internal tool — light mode doubles QA surface). - Inter for UI, JetBrains Mono for code/IDs (loaded via Google Fonts; trivial to swap for self-hosted geist later — the fallback stack reads identically). - Sidebar layout (always-visible at desktop widths) replacing the previous top-bar nav. Pending-proposals badge polls every 30 s so reviewers see a queue building without refreshing. - Lucide icons throughout. - Spacing and radii on Tailwind defaults. Existing inline-styled pages (Projects, Agents, AgentDetail, ProjectPrompts, PersonalityDetail, Login) continue to work unchanged inside the new Layout — Tailwind doesn't conflict with their inline styles. A follow-up can migrate them incrementally. ## What's added ### Build infra (src/web/) - package.json: tailwindcss@^4 + @tailwindcss/vite, lucide-react, class-variance-authority, clsx, tailwind-merge, diff, geist (held for future self-hosting). - vite.config.ts: registers the @tailwindcss/vite plugin. - src/index.css: Tailwind import + @theme tokens + @layer base. - src/main.tsx: imports index.css. - src/lib/utils.ts: shadcn-style cn() helper. ### shadcn-style primitives (src/components/ui/) Hand-written rather than generated via `npx shadcn` so the repo doesn't depend on a CLI tool that needs interactive runtime: - button.tsx — variants: primary / secondary / ghost / danger / link; sizes: sm / md / lg / icon. - card.tsx — Card + Header/Title/Description/Content/Footer subparts. - badge.tsx — variants: default / info / success / warning / danger / outline. - input.tsx — Input + Textarea + Label. - tabs.tsx — no-dep accessible Tabs (no Radix needed for our use). - separator.tsx — h/v separator with role=separator. ### Diff component (src/components/Diff.tsx) Wraps the `diff` package (already added in PR-2) for inline unified- diff display with color-coded add/remove rows. Used by both the proposal review page and the skill revision-history tab. ### New pages (src/pages/) - Dashboard.tsx — at-a-glance home. Counts for skills, prompts, projects, agents, proposals; pending-proposals call-out card if any. - Skills.tsx — list view, separated into Global vs Project/Agent- scoped sections. - SkillDetail.tsx — name + semver + description; tabs for SKILL.md / Files / Metadata / History. History tab shows revisions with click-to-diff against the live body. - Proposals.tsx — queue with Pending/Approved/Rejected tabs. Pending count is highlighted in amber. - ProposalDetail.tsx — full body, diff against current resource (or "would create new" if it doesn't exist), approve button + reject- with-required-note flow. ### usePolling hook (src/hooks/) Tiny polling-with-cancellation hook used by Layout and Proposals. ### Layout rewrite (src/components/Layout.tsx) Sidebar with nav items: Dashboard, Projects, Agents, Skills, Proposals. Lucide icons. Active-route highlighting via NavLink. Pending-proposals warning badge on the Proposals item. ### Routes (src/App.tsx) New routes: /dashboard, /skills, /skills/:name, /proposals, /proposals/:id. Default redirects to /dashboard. ### API types (src/api.ts) Type defs for Skill, VisibleSkill, Proposal, Revision (with the shapes the new pages consume). ## Tests Existing 7 web tests still pass (Login + api). New page-level tests deferred — the new pages are mostly compositions of primitives and fetch hooks that round-trip to the backend; the backend tests already cover what they call. PR-7 polish can add render-and-click tests if coverage drift surfaces. ## Verification - `pnpm --filter @mcpctl/web build` clean, no warnings. - `pnpm test:run` whole monorepo: 162 test files / 2157 tests green. - Visual smoke deferred — needs a running mcpd to populate the fixtures. Manual smoke tested locally is the next step. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-07 17:54:55 +01:00
postcss@8.4.31:
resolution: {integrity: sha512-PS08Iboia9mts/2ygV3eLpY5ghnUcfLV/EXTOW1E2qYxJKGGBUtNjN76FYHnMs36RmARn41bC0AZmn+rR0OVpQ==}
engines: {node: ^10 || ^12 || >=14}
first commit
2026-02-21 03:10:39 +00:00
postcss@8.5.6:
resolution: {integrity: sha512-3Ybi1tAuwAP9s0r1UQ2J4n5Y0G05bJkpUIO0/bI9MhwmD70S5aTWbXGBwxHrelT+XM1k6dM0pk+SwNkpTRN7Pg==}
engines: {node: ^10 || ^12 || >=14}
prelude-ls@1.2.1:
resolution: {integrity: sha512-vkcDPrRZo1QZLbn5RLGPpg/WmIQ65qoWWhcGKf/b5eplkkarX0m9z8ppCat4mlOqUsWpyNuYgO3VRyrYHSzX5g==}
engines: {node: '>= 0.8.0'}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
pretty-format@27.5.1:
resolution: {integrity: sha512-Qb1gy5OrP5+zDf2Bvnzdl3jsTf1qXVMazbvCoKhtKqVs4/YK4ozX4gKQJJVyNe+cajNPn0KoC0MC3FUmaHWEmQ==}
engines: {node: ^10.13.0 || ^12.13.0 || ^14.15.0 || >=15.0.0}
first commit
2026-02-21 03:10:39 +00:00
prisma@6.19.2:
resolution: {integrity: sha512-XTKeKxtQElcq3U9/jHyxSPgiRgeYDKxWTPOf6NkXA0dNj5j40MfEsZkMbyNpwDWCUv7YBFUl7I2VK/6ALbmhEg==}
engines: {node: '>=18.18'}
hasBin: true
peerDependencies:
typescript: '>=5.1.0'
peerDependenciesMeta:
typescript:
optional: true
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
proc-log@6.1.0:
resolution: {integrity: sha512-iG+GYldRf2BQ0UDUAd6JQ/RwzaQy6mXmsk/IzlYyal4A4SNFw54MeH4/tLkF4I5WoWG9SQwuqWzS99jaFQHBuQ==}
engines: {node: ^20.17.0 || >=22.9.0}
first commit
2026-02-21 03:10:39 +00:00
process-warning@4.0.1:
resolution: {integrity: sha512-3c2LzQ3rY9d0hc1emcsHhfT9Jwz0cChib/QN89oME2R451w5fy3f0afAhERFZAwrbDU43wk12d0ORBpDVME50Q==}
process-warning@5.0.0:
resolution: {integrity: sha512-a39t9ApHNx2L4+HBnQKqxxHNs1r7KF+Intd8Q/g1bUh6q0WIp9voPXJ/x0j+ZL45KF1pJd9+q2jLIRMfvEshkA==}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
proggy@4.0.0:
resolution: {integrity: sha512-MbA4R+WQT76ZBm/5JUpV9yqcJt92175+Y0Bodg3HgiXzrmKu7Ggq+bpn6y6wHH+gN9NcyKn3yg1+d47VaKwNAQ==}
engines: {node: ^20.17.0 || >=22.9.0}
promise-all-reject-late@1.0.1:
resolution: {integrity: sha512-vuf0Lf0lOxyQREH7GDIOUMLS7kz+gs8i6B+Yi8dC68a2sychGrHTJYghMBD6k7eUcH0H5P73EckCA48xijWqXw==}
promise-call-limit@3.0.2:
resolution: {integrity: sha512-mRPQO2T1QQVw11E7+UdCJu7S61eJVWknzml9sC1heAdj1jxl0fWMBypIt9ZOcLFf8FkG995ZD7RnVk7HH72fZw==}
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
proper-lockfile@4.1.2:
resolution: {integrity: sha512-TjNPblN4BwAWMXU8s9AEz4JmQxnD1NNL7bNOY/AKUzyamc379FWASUhc/K1pL2noVb+XmZKLL68cjzLsiOAMaA==}
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
protobufjs@7.5.4:
resolution: {integrity: sha512-CvexbZtbov6jW2eXAvLukXjXUW1TzFaivC46BpWc/3BpcCysb5Vffu+B3XHMm8lVEuy2Mm4XGex8hBSg1yapPg==}
engines: {node: '>=12.0.0'}
first commit
2026-02-21 03:10:39 +00:00
proxy-addr@2.0.7:
resolution: {integrity: sha512-llQsMLSUDUPT44jdrU/O37qlnifitDP+ZwrmmZcoSKyLKvtZxpyV0n2/bD/N4tBAAZ/gJEdZU7KMraoK1+XYAg==}
engines: {node: '>= 0.10'}
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
pump@3.0.3:
resolution: {integrity: sha512-todwxLMY7/heScKmntwQG8CXVkWUOdYxIvY2s0VWAAMh/nd8SoYiRaKjlr7+iCs984f2P8zvrfWcDDYVb73NfA==}
first commit
2026-02-21 03:10:39 +00:00
punycode@2.3.1:
resolution: {integrity: sha512-vYt7UD1U9Wg6138shLtLOvdAu+8DsC/ilFtEVHcH+wydcSpNE20AfSOduf6MkRFahL5FY7X1oU7nKVZFtfq8Fg==}
engines: {node: '>=6'}
pure-rand@6.1.0:
resolution: {integrity: sha512-bVWawvoZoBYpp6yIoQtQXHZjmz35RSVHnUOTefl8Vcjr8snTPY1wnpSPMWekcFwbxI6gtmT7rSYPFvz71ldiOA==}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
pure-rand@8.4.2:
resolution: {integrity: sha512-vvuOGgcuPJAirlHvuQw1TrOiw7ptaIXXmIbNuiNOY6lNGJJH49PQ1Kj4nd783nPdQhQdicgOjVI2yI/9BD6/Ng==}
first commit
2026-02-21 03:10:39 +00:00
qs@6.15.0:
resolution: {integrity: sha512-mAZTtNCeetKMH+pSjrb76NAM8V9a05I9aBZOHztWy/UqcJdQYNsf59vrRKWnojAT9Y+GbIvoTBC++CPHqpDBhQ==}
engines: {node: '>=0.6'}
quick-format-unescaped@4.0.4:
resolution: {integrity: sha512-tYC1Q1hgyRuHgloV/YXs2w15unPVh8qfu/qCTfhTYamaw7fyhumKa2yGpdSo87vY32rIclj+4fWYQXUMs9EHvg==}
range-parser@1.2.1:
resolution: {integrity: sha512-Hrgsx+orqoygnmhFbKaHE6c296J+HTAQXoxEF6gNupROmmGJRoyzfG3ccAveqCBrwr/2yxQ5BVd/GTl5agOwSg==}
engines: {node: '>= 0.6'}
raw-body@3.0.2:
resolution: {integrity: sha512-K5zQjDllxWkf7Z5xJdV0/B0WTNqx6vxG70zJE4N0kBs4LovmEYWJzQGxC9bS9RAKu3bgM40lrd5zoLJ12MQ5BA==}
engines: {node: '>= 0.10'}
rc9@2.1.2:
resolution: {integrity: sha512-btXCnMmRIBINM2LDZoEmOogIZU7Qe7zn4BpomSKZ/ykbLObuBdvG+mFq11DL6fjH1DRwHhrlgtYWG96bJiC7Cg==}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
react-dom@19.2.5:
resolution: {integrity: sha512-J5bAZz+DXMMwW/wV3xzKke59Af6CHY7G4uYLN1OvBcKEsWOs4pQExj86BBKamxl/Ik5bx9whOrvBlSDfWzgSag==}
peerDependencies:
react: ^19.2.5
react-is@17.0.2:
resolution: {integrity: sha512-w2GsyukL62IJnlaff/nRegPQR94C/XXamvMWmSHRJ4y7Ts/4ocGRmTHvOs8PSE6pB3dWOrD/nueuU5sduBsQ4w==}
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
react-reconciler@0.33.0:
resolution: {integrity: sha512-KetWRytFv1epdpJc3J4G75I4WrplZE5jOL7Yq0p34+OVOKF4Se7WrdIdVC45XsSSmUTlht2FM/fM1FZb1mfQeA==}
engines: {node: '>=0.10.0'}
peerDependencies:
react: ^19.2.0
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
react-refresh@0.18.0:
resolution: {integrity: sha512-QgT5//D3jfjJb6Gsjxv0Slpj23ip+HtOpnNgnb2S5zU3CB26G/IDPGoy4RJB42wzFE46DRsstbW6tKHoKbhAxw==}
engines: {node: '>=0.10.0'}
react-router-dom@7.14.2:
resolution: {integrity: sha512-YZcM5ES8jJSM+KrJ9BdvHHqlnGTg5tH3sC5ChFRj4inosKctdyzBDhOyyHdGk597q2OT6NTrCA1OvB/YDwfekQ==}
engines: {node: '>=20.0.0'}
peerDependencies:
react: '>=18'
react-dom: '>=18'
react-router@7.14.2:
resolution: {integrity: sha512-yCqNne6I8IB6rVCH7XUvlBK7/QKyqypBFGv+8dj4QBFJiiRX+FG7/nkdAvGElyvVZ/HQP5N19wzteuTARXi5Gw==}
engines: {node: '>=20.0.0'}
peerDependencies:
react: '>=18'
react-dom: '>=18'
peerDependenciesMeta:
react-dom:
optional: true
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
react@19.2.4:
resolution: {integrity: sha512-9nfp2hYpCwOjAN+8TZFGhtWEwgvWHXqESH8qT89AT/lWklpLON22Lc8pEtnpsZz7VmawabSU0gCjnj8aC0euHQ==}
engines: {node: '>=0.10.0'}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
react@19.2.5:
resolution: {integrity: sha512-llUJLzz1zTUBrskt2pwZgLq59AemifIftw4aB7JxOqf1HY2FDaGDxgwpAPVzHU1kdWabH7FauP4i1oEeer2WCA==}
engines: {node: '>=0.10.0'}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
read-cmd-shim@6.0.0:
resolution: {integrity: sha512-1zM5HuOfagXCBWMN83fuFI/x+T/UhZ7k+KIzhrHXcQoeX5+7gmaDYjELQHmmzIodumBHeByBJT4QYS7ufAgs7A==}
engines: {node: ^20.17.0 || >=22.9.0}
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
readable-stream@3.6.2:
resolution: {integrity: sha512-9u/sniCrY3D5WdsERHzHE4G2YCXqoG5FTHUiCC4SIbr6XcLZBY05ya9EKjYek9O5xOAwjGq+1JdGBAS7Q9ScoA==}
engines: {node: '>= 6'}
first commit
2026-02-21 03:10:39 +00:00
readdirp@4.1.2:
resolution: {integrity: sha512-GDhwkLfywWL2s6vEjyhri+eXmfH6j1L7JE27WhqLeYzoh/A3DBaYGEj2H/HFZCn/kMfim73FXxEJTw06WtxQwg==}
engines: {node: '>= 14.18.0'}
real-require@0.2.0:
resolution: {integrity: sha512-57frrGM/OCTLqLOAh0mhVA9VBMHd+9U7Zb2THMGdBUoZVOtGbJzjxsYGDJ3A9AYYCP4hn6y1TVbaOfzWtm5GFg==}
engines: {node: '>= 12.13.0'}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
redent@3.0.0:
resolution: {integrity: sha512-6tDA8g98We0zd0GvVeMT9arEOnTw9qM03L9cJXaCjrip1OO764RDBLBfrB4cwzNGDj5OA5ioymC9GkizgWJDUg==}
engines: {node: '>=8'}
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
require-directory@2.1.1:
resolution: {integrity: sha512-fGxEI7+wsG9xrvdjsrlmL22OMTTiHRwAMroiEeMgq8gzoLC/PQr7RsRDSTLUg/bZAZtF+TVIkHc6/4RIKrui+Q==}
engines: {node: '>=0.10.0'}
first commit
2026-02-21 03:10:39 +00:00
require-from-string@2.0.2:
resolution: {integrity: sha512-Xf0nWe6RseziFMu+Ap9biiUbmplq6S9/p+7w7YXP/JBHhrUDDUhwa+vANyubuqfZWTveU//DYVGsDG7RKL/vEw==}
engines: {node: '>=0.10.0'}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
require-in-the-middle@8.0.1:
resolution: {integrity: sha512-QT7FVMXfWOYFbeRBF6nu+I6tr2Tf3u0q8RIEjNob/heKY/nh7drD/k7eeMFmSQgnTtCzLDcCu/XEnpW2wk4xCQ==}
engines: {node: '>=9.3.0 || >=8.10.0 <9.0.0'}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
reselect@4.1.8:
resolution: {integrity: sha512-ab9EmR80F/zQTMNeneUr4cv+jSwPJgIlvEmVwLerwrWVbpLlBuls9XHzIeTFy4cegU2NHBp3va0LKOzU5qFEYQ==}
first commit
2026-02-21 03:10:39 +00:00
resolve-pkg-maps@1.0.0:
resolution: {integrity: sha512-seS2Tj26TBVOC2NIc2rOe2y2ZO7efxITtLZcGSOnHHNOQ7CkiUBfw0Iw2ck6xkIhPwLhKNLS8BO+hEpngQlqzw==}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
resolve@1.22.12:
resolution: {integrity: sha512-TyeJ1zif53BPfHootBGwPRYT1RUt6oGWsaQr8UyZW/eAm9bKoijtvruSDEmZHm92CwS9nj7/fWttqPCgzep8CA==}
engines: {node: '>= 0.4'}
hasBin: true
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
restore-cursor@4.0.0:
resolution: {integrity: sha512-I9fPXU9geO9bHOt9pHHOhOkYerIMsmVaWB0rA2AI9ERh/+x/i7MV5HKBNrg+ljO5eoPVgCcnFuRjJ9uH6I/3eg==}
engines: {node: ^12.20.0 || ^14.13.1 || >=16.0.0}
first commit
2026-02-21 03:10:39 +00:00
ret@0.5.0:
resolution: {integrity: sha512-I1XxrZSQ+oErkRR4jYbAyEEu2I0avBvvMM5JN+6EBprOGRCs63ENqZ3vjavq8fBw2+62G5LF5XelKwuJpcvcxw==}
engines: {node: '>=10'}
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
retry@0.12.0:
resolution: {integrity: sha512-9LkiTwjUh6rT555DtE9rTX+BKByPfrMzEAtnlEtdEwr3Nkffwiihqe2bWADg+OQRjt9gl6ICdmB/ZFDCGAtSow==}
engines: {node: '>= 4'}
retry@0.13.1:
resolution: {integrity: sha512-XQBQ3I8W1Cge0Seh+6gjj03LbmRFWuoszgK9ooCpwYIrhhoO80pfq4cUkU5DkknwfOfFteRwlZ56PYOGYyFWdg==}
engines: {node: '>= 4'}
first commit
2026-02-21 03:10:39 +00:00
reusify@1.1.0:
resolution: {integrity: sha512-g6QUff04oZpHs0eG5p83rFLhHeV00ug/Yf9nZM6fLeUrPguBTkTQOdpAWWspMh55TZfVQDPaN3NQJfbVRAxdIw==}
engines: {iojs: '>=1.0.0', node: '>=0.10.0'}
feat: add Kubernetes orchestrator for MCP server pod management mcpd can now deploy MCP server instances as Kubernetes pods instead of Docker containers. Set MCPD_ORCHESTRATOR=kubernetes to enable. - Add @kubernetes/client-node with thin wrapper (context enforcement via MCPD_K8S_CONTEXT to prevent multi-cluster mishaps) - Rewrite KubernetesOrchestrator: pod CRUD, pod IP extraction, exec via SPDY (one-shot + interactive), log streaming - Manifest generator: stdin:true for STDIO servers, args (not command) to preserve runner image entrypoint, security hardening - Orchestrator selection in main.ts via MCPD_ORCHESTRATOR env var - 25 unit tests for k8s orchestrator, all 624 tests pass Tested end-to-end on local k3s: - mcpd deployed via Pulumi, creates pods in mcpctl-servers namespace - NetworkPolicy verified: only mcpd can reach MCP server pods - Python runner (uvx) successfully runs aws-documentation-mcp-server Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-08 01:55:13 +01:00
rfc4648@1.5.4:
resolution: {integrity: sha512-rRg/6Lb+IGfJqO05HZkN50UtY7K/JhxJag1kP23+zyMfrvoB0B7RWv06MbOzoc79RgCdNTiUaNsTT1AJZ7Z+cg==}
first commit
2026-02-21 03:10:39 +00:00
rfdc@1.4.1:
resolution: {integrity: sha512-q1b3N5QkRUWUl7iyylaaj3kOpIT0N2i9MqIEQXP73GVsN9cw3fdx8X63cEmWhJGi2PPCF23Ijp7ktmd39rawIA==}
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
rimraf@3.0.2:
resolution: {integrity: sha512-JZkJMZkAGFFPP2YqXZXPbMlMBgsxzE8ILs4lMIX/2o0L9UBw9O/Y3o6wFw/i9YLapcUJWwqbi3kdxIPdC62TIA==}
deprecated: Rimraf versions prior to v4 are no longer supported
hasBin: true
first commit
2026-02-21 03:10:39 +00:00
rimraf@6.1.3:
resolution: {integrity: sha512-LKg+Cr2ZF61fkcaK1UdkH2yEBBKnYjTyWzTJT6KNPcSPaiT7HSdhtMXQuN5wkTX0Xu72KQ1l8S42rlmexS2hSA==}
engines: {node: 20 || >=22}
hasBin: true
rollup@4.58.0:
resolution: {integrity: sha512-wbT0mBmWbIvvq8NeEYWWvevvxnOyhKChir47S66WCxw1SXqhw7ssIYejnQEVt7XYQpsj2y8F9PM+Cr3SNEa0gw==}
engines: {node: '>=18.0.0', npm: '>=8.0.0'}
hasBin: true
router@2.2.0:
resolution: {integrity: sha512-nLTrUKm2UyiL7rlhapu/Zl45FwNgkZGaCpZbIHajDYgwlJCOzLSk+cIPAnsEqV955GjILJnKbdQC1nVPz+gAYQ==}
engines: {node: '>= 18'}
run-async@4.0.6:
resolution: {integrity: sha512-IoDlSLTs3Yq593mb3ZoKWKXMNu3UpObxhgA/Xuid5p4bbfi2jdY1Hj0m1K+0/tEuQTxIGMhQDqGjKb7RuxGpAQ==}
engines: {node: '>=0.12.0'}
rxjs@7.8.2:
resolution: {integrity: sha512-dhKf903U/PQZY6boNNtAGdWbG85WAbjT/1xYoZIC7FAY0yWapOBQVsVrDl58W86//e1VpMNBtRV4MaXfdMySFA==}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
s-js@0.4.9:
resolution: {integrity: sha512-RtpOm+cM6O0sHg6IA70wH+UC3FZcND+rccBZpBAHzlUgNO2Bm5BN+FnM8+OBxzXdwpKWFwX11JGF0MFRkhSoIQ==}
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
safe-buffer@5.2.1:
resolution: {integrity: sha512-rp3So07KcdmmKbGvgaNxQSJr7bGVSVk5S9Eq1F+ppbRo70+YeaDxkw5Dd8NPN+GD6bjnYm2VuPuCXmpuYvmCXQ==}
first commit
2026-02-21 03:10:39 +00:00
safe-regex2@5.0.0:
resolution: {integrity: sha512-YwJwe5a51WlK7KbOJREPdjNrpViQBI3p4T50lfwPuDhZnE3XGVTlGvi+aolc5+RvxDD6bnUmjVsU9n1eboLUYw==}
safe-stable-stringify@2.5.0:
resolution: {integrity: sha512-b3rppTKm9T+PsVCBEOUR46GWI7fdOs00VKZ1+9c1EWDaDMvjQc6tUwuFyIprgGgTcWoVHSKrU8H31ZHA2e0RHA==}
engines: {node: '>=10'}
safer-buffer@2.1.2:
resolution: {integrity: sha512-YZo3K82SD7Riyi0E1EQPojLz7kpepnSQI9IyPbHHg1XXXevb5dJI7tpyN2ADxGcQbHG7vcyRHk0cbwqcQriUtg==}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
saxes@6.0.0:
resolution: {integrity: sha512-xAg7SOnEhrm5zI3puOOKyy1OMcMlIJZYNJY7xLBwSze0UjhPLnWfj2GF2EpT0jmzaJKIWKHLsaSSajf35bcYnA==}
engines: {node: '>=v12.22.7'}
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
scheduler@0.27.0:
resolution: {integrity: sha512-eNv+WrVbKu1f3vbYJT/xtiF5syA5HPIMtf9IgY/nKg0sWqzAUEvqY/xm7OcZc/qafLx/iO9FgOmeSAp4v5ti/Q==}
first commit
2026-02-21 03:10:39 +00:00
secure-json-parse@4.1.0:
resolution: {integrity: sha512-l4KnYfEyqYJxDwlNVyRfO2E4NTHfMKAWdUuA8J0yve2Dz/E/PdBepY03RvyJpssIpRFwJoCD55wA+mEDs6ByWA==}
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
semver@6.3.1:
resolution: {integrity: sha512-BR7VvDCVHO+q2xBEWskxS6DJE1qRnb7DxzUrogb71CWoSficBxYsiAGd+Kl0mmq/MprG9yArRkyrQxTO6XjMzA==}
hasBin: true
first commit
2026-02-21 03:10:39 +00:00
semver@7.7.4:
resolution: {integrity: sha512-vFKC2IEtQnVhpT78h1Yp8wzwrf8CM+MzKMHGJZfBtzhZNycRFnXsHk6E5TxIkkMsgNS7mdX3AGB7x2QM2di4lA==}
engines: {node: '>=10'}
hasBin: true
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
semver@7.8.0:
resolution: {integrity: sha512-AcM7dV/5ul4EekoQ29Agm5vri8JNqRyj39o0qpX6vDF2GZrtutZl5RwgD1XnZjiTAfncsJhMI48QQH3sN87YNA==}
engines: {node: '>=10'}
hasBin: true
first commit
2026-02-21 03:10:39 +00:00
send@1.2.1:
resolution: {integrity: sha512-1gnZf7DFcoIcajTjTwjwuDjzuz4PPcY2StKPlsGAQ1+YH20IRVrBaXSWmdjowTJ6u8Rc01PoYOGHXfP1mYcZNQ==}
engines: {node: '>= 18'}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
seroval-plugins@1.5.6:
resolution: {integrity: sha512-HXuLAX2pu/UByPpaeo/TaMfvMIi+1QqIoPJYCcAtU8QkVNwgR6MPlGuCQTErV1JwraaMbYaWVIBX7mppzGLATQ==}
engines: {node: '>=10'}
peerDependencies:
seroval: ^1.0
seroval@1.5.6:
resolution: {integrity: sha512-rVQVWjjSvlINzaQPZH5JFqsqEsIWdTxY3iJZCnTL/5gQbXIRooVZKI60tVCkOVfzcRPejboxO2t0P89dg5mQaA==}
engines: {node: '>=10'}
first commit
2026-02-21 03:10:39 +00:00
serve-static@2.2.1:
resolution: {integrity: sha512-xRXBn0pPqQTVQiC8wyQrKs2MOlX24zQ0POGaj0kultvoOCstBQM5yvOhAVSUwOMjQtTvsPWoNCHfPGwaaQJhTw==}
engines: {node: '>= 18'}
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
set-blocking@2.0.0:
resolution: {integrity: sha512-KiKBS8AnWGEyLzofFfmvKwpdPzqiy16LvQfK3yv/fVH7Bj13/wl3JSR1J+rfgRE9q7xUJK4qvgS8raSOeLUehw==}
first commit
2026-02-21 03:10:39 +00:00
set-cookie-parser@2.7.2:
resolution: {integrity: sha512-oeM1lpU/UvhTxw+g3cIfxXHyJRc/uidd3yK1P242gzHds0udQBYzs3y8j4gCCW+ZJ7ad0yctld8RYO+bdurlvw==}
setprototypeof@1.2.0:
resolution: {integrity: sha512-E5LDX7Wrp85Kil5bhZv46j8jOeboKq5JMmYM3gVGdGH8xFpPWXUMsNrlODCrkoxMEeNi/XZIwuRvY4XNwYMJpw==}
feat(web): bold redesign — Tailwind v4 + shadcn-style primitives + Skills/Proposals/Revisions UI Phase 6 of the Skills + Revisions + Proposals work. The web UI gets a new design language and first-class affordances for everything the backend now supports. ## Visual direction - Tailwind v4 with custom @theme block (oklch tokens). Dark-mode-only (internal tool — light mode doubles QA surface). - Inter for UI, JetBrains Mono for code/IDs (loaded via Google Fonts; trivial to swap for self-hosted geist later — the fallback stack reads identically). - Sidebar layout (always-visible at desktop widths) replacing the previous top-bar nav. Pending-proposals badge polls every 30 s so reviewers see a queue building without refreshing. - Lucide icons throughout. - Spacing and radii on Tailwind defaults. Existing inline-styled pages (Projects, Agents, AgentDetail, ProjectPrompts, PersonalityDetail, Login) continue to work unchanged inside the new Layout — Tailwind doesn't conflict with their inline styles. A follow-up can migrate them incrementally. ## What's added ### Build infra (src/web/) - package.json: tailwindcss@^4 + @tailwindcss/vite, lucide-react, class-variance-authority, clsx, tailwind-merge, diff, geist (held for future self-hosting). - vite.config.ts: registers the @tailwindcss/vite plugin. - src/index.css: Tailwind import + @theme tokens + @layer base. - src/main.tsx: imports index.css. - src/lib/utils.ts: shadcn-style cn() helper. ### shadcn-style primitives (src/components/ui/) Hand-written rather than generated via `npx shadcn` so the repo doesn't depend on a CLI tool that needs interactive runtime: - button.tsx — variants: primary / secondary / ghost / danger / link; sizes: sm / md / lg / icon. - card.tsx — Card + Header/Title/Description/Content/Footer subparts. - badge.tsx — variants: default / info / success / warning / danger / outline. - input.tsx — Input + Textarea + Label. - tabs.tsx — no-dep accessible Tabs (no Radix needed for our use). - separator.tsx — h/v separator with role=separator. ### Diff component (src/components/Diff.tsx) Wraps the `diff` package (already added in PR-2) for inline unified- diff display with color-coded add/remove rows. Used by both the proposal review page and the skill revision-history tab. ### New pages (src/pages/) - Dashboard.tsx — at-a-glance home. Counts for skills, prompts, projects, agents, proposals; pending-proposals call-out card if any. - Skills.tsx — list view, separated into Global vs Project/Agent- scoped sections. - SkillDetail.tsx — name + semver + description; tabs for SKILL.md / Files / Metadata / History. History tab shows revisions with click-to-diff against the live body. - Proposals.tsx — queue with Pending/Approved/Rejected tabs. Pending count is highlighted in amber. - ProposalDetail.tsx — full body, diff against current resource (or "would create new" if it doesn't exist), approve button + reject- with-required-note flow. ### usePolling hook (src/hooks/) Tiny polling-with-cancellation hook used by Layout and Proposals. ### Layout rewrite (src/components/Layout.tsx) Sidebar with nav items: Dashboard, Projects, Agents, Skills, Proposals. Lucide icons. Active-route highlighting via NavLink. Pending-proposals warning badge on the Proposals item. ### Routes (src/App.tsx) New routes: /dashboard, /skills, /skills/:name, /proposals, /proposals/:id. Default redirects to /dashboard. ### API types (src/api.ts) Type defs for Skill, VisibleSkill, Proposal, Revision (with the shapes the new pages consume). ## Tests Existing 7 web tests still pass (Login + api). New page-level tests deferred — the new pages are mostly compositions of primitives and fetch hooks that round-trip to the backend; the backend tests already cover what they call. PR-7 polish can add render-and-click tests if coverage drift surfaces. ## Verification - `pnpm --filter @mcpctl/web build` clean, no warnings. - `pnpm test:run` whole monorepo: 162 test files / 2157 tests green. - Visual smoke deferred — needs a running mcpd to populate the fixtures. Manual smoke tested locally is the next step. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-07 17:54:55 +01:00
sharp@0.34.5:
resolution: {integrity: sha512-Ou9I5Ft9WNcCbXrU9cMgPBcCK8LiwLqcbywW3t4oDV37n1pzpuNLsYiAV8eODnjbtQlSDwZ2cUEeQz4E54Hltg==}
engines: {node: ^18.17.0 || ^20.3.0 || >=21.0.0}
first commit
2026-02-21 03:10:39 +00:00
shebang-command@2.0.0:
resolution: {integrity: sha512-kHxr2zZpYtdmrN1qDjrrX/Z1rR1kG8Dx+gkpK1G4eXmvXswmcE1hTWBWYUzlraYw1/yZp6YuDY77YtvbN0dmDA==}
engines: {node: '>=8'}
shebang-regex@3.0.0:
resolution: {integrity: sha512-7++dFhtcx3353uBaq8DDR4NuxBetBzC7ZQOhmTQInHEd6bSrXdiEyzCvG07Z44UYdLShWUyXt5M/yhz8ekcb1A==}
engines: {node: '>=8'}
side-channel-list@1.0.0:
resolution: {integrity: sha512-FCLHtRD/gnpCiCHEiJLOwdmFP+wzCmDEkc9y7NsYxeF4u7Btsn1ZuwgwJGxImImHicJArLP4R0yX4c2KCrMrTA==}
engines: {node: '>= 0.4'}
side-channel-map@1.0.1:
resolution: {integrity: sha512-VCjCNfgMsby3tTdo02nbjtM/ewra6jPHmpThenkTYh8pG9ucZ/1P8So4u4FGBek/BjpOVsDCMoLA/iuBKIFXRA==}
engines: {node: '>= 0.4'}
side-channel-weakmap@1.0.2:
resolution: {integrity: sha512-WPS/HvHQTYnHisLo9McqBHOJk2FkHO/tlpvldyrnem4aeQp4hai3gythswg6p01oSoTl58rcpiFAjF2br2Ak2A==}
engines: {node: '>= 0.4'}
side-channel@1.1.0:
resolution: {integrity: sha512-ZX99e6tRweoUXqR+VBrslhda51Nh5MTQwou5tnUDgbtyM0dBgmhEDtWGP/xbKn6hqfPRHujUNwz5fy/wbbhnpw==}
engines: {node: '>= 0.4'}
siginfo@2.0.0:
resolution: {integrity: sha512-ybx0WO1/8bSBLEWXZvEd7gMW3Sn3JFlW3TvX1nREbDLRNQNaeNN8WK0meBwPdAaOI7TtRRRJn/Es1zhrrCHu7g==}
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
signal-exit@3.0.7:
resolution: {integrity: sha512-wnD2ZE+l+SPC/uoS0vXeE9L1+0wuaMqKlfz9AMUo38JsyLSBWSFcHR1Rri62LZc12vLr1gb3jl7iwQhgwpAbGQ==}
first commit
2026-02-21 03:10:39 +00:00
signal-exit@4.1.0:
resolution: {integrity: sha512-bzyZ1e88w9O1iNJbKnOlvYTrWPDl46O1bG0D3XInv+9tkPrxrN8jUUTiFlDkkmKWgn1M6CfIA13SuGqOa9Korw==}
engines: {node: '>=14'}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
sigstore@4.1.1:
resolution: {integrity: sha512-endqECJkfhozrXMK5ngu/UAA0xVcVEFdnHJCElGaExypjW+HK5i6zu3NteLoaX/iFbRUbC3+DjttQs0GARr+5w==}
engines: {node: ^20.17.0 || >=22.9.0}
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
slice-ansi@7.1.2:
resolution: {integrity: sha512-iOBWFgUX7caIZiuutICxVgX1SdxwAVFFKwt1EvMYYec/NWO5meOJ6K5uQxhrYBdQJne4KxiqZc+KptFOWFSI9w==}
engines: {node: '>=18'}
slice-ansi@8.0.0:
resolution: {integrity: sha512-stxByr12oeeOyY2BlviTNQlYV5xOj47GirPr4yA1hE9JCtxfQN0+tVbkxwCtYDQWhEKWFHsEK48ORg5jrouCAg==}
engines: {node: '>=20'}
feat: add Kubernetes orchestrator for MCP server pod management mcpd can now deploy MCP server instances as Kubernetes pods instead of Docker containers. Set MCPD_ORCHESTRATOR=kubernetes to enable. - Add @kubernetes/client-node with thin wrapper (context enforcement via MCPD_K8S_CONTEXT to prevent multi-cluster mishaps) - Rewrite KubernetesOrchestrator: pod CRUD, pod IP extraction, exec via SPDY (one-shot + interactive), log streaming - Manifest generator: stdin:true for STDIO servers, args (not command) to preserve runner image entrypoint, security hardening - Orchestrator selection in main.ts via MCPD_ORCHESTRATOR env var - 25 unit tests for k8s orchestrator, all 624 tests pass Tested end-to-end on local k3s: - mcpd deployed via Pulumi, creates pods in mcpctl-servers namespace - NetworkPolicy verified: only mcpd can reach MCP server pods - Python runner (uvx) successfully runs aws-documentation-mcp-server Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-08 01:55:13 +01:00
smart-buffer@4.2.0:
resolution: {integrity: sha512-94hK0Hh8rPqQl2xXc3HsaBoOXKV20MToPkcXvwbISWLEs+64sBq5kFgn2kJDHb1Pry9yrP0dxrCI9RRci7RXKg==}
engines: {node: '>= 6.0.0', npm: '>= 3.0.0'}
socks-proxy-agent@8.0.5:
resolution: {integrity: sha512-HehCEsotFqbPW9sJ8WVYB6UbmIMv7kUUORIF2Nncq4VQvBfNBLibW9YZR5dlYCSUhwcD628pRllm7n+E+YTzJw==}
engines: {node: '>= 14'}
socks@2.8.7:
resolution: {integrity: sha512-HLpt+uLy/pxB+bum/9DzAgiKS8CX1EvbWxI4zlmgGCExImLdiad2iCwXT5Z4c9c3Eq8rP2318mPW2c+QbtjK8A==}
engines: {node: '>= 10.0.0', npm: '>= 3.0.0'}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
solid-js@1.9.12:
resolution: {integrity: sha512-QzKaSJq2/iDrWR1As6MHZQ8fQkdOBf8GReYb7L5iKwMGceg7HxDcaOHk0at66tNgn9U2U7dXo8ZZpLIAmGMzgw==}
first commit
2026-02-21 03:10:39 +00:00
sonic-boom@4.2.1:
resolution: {integrity: sha512-w6AxtubXa2wTXAUsZMMWERrsIRAdrK0Sc+FUytWvYAhBJLyuI4llrMIC1DtlNSdI99EI86KZum2MMq3EAZlF9Q==}
source-map-js@1.2.1:
resolution: {integrity: sha512-UXWMKhLOwVKb728IUtQPXxfYU+usdybtUrK/8uGE8CQMvrhOpwvzDBwj0QhSL7MQc7vIsISBG8VQ8+IDQxpfQA==}
engines: {node: '>=0.10.0'}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
source-map-support@0.5.21:
resolution: {integrity: sha512-uBHU3L3czsIyYXKX88fdrGovxdSCoTGDRZ6SYXtSRxLZUzHg5P/66Ht6uoUlHu9EZod+inXhKo3qQgwXUT/y1w==}
source-map@0.6.1:
resolution: {integrity: sha512-UjgapumWlbMhkBgzT7Ykc5YXUT46F0iKu8SGXq0bcwP5dz/h0Plj6enJqjz1Zbq2l5WaqYnrVbwWOWMyF3F47g==}
engines: {node: '>=0.10.0'}
spdx-correct@3.2.0:
resolution: {integrity: sha512-kN9dJbvnySHULIluDHy32WHRUu3Og7B9sbY7tsFLctQkIqnMh3hErYgdMjTYuqmcXX+lK5T1lnUt3G7zNswmZA==}
spdx-exceptions@2.5.0:
resolution: {integrity: sha512-PiU42r+xO4UbUS1buo3LPJkjlO7430Xn5SVAhdpzzsPHsjbYVflnnFdATgabnLude+Cqu25p6N+g2lw/PFsa4w==}
spdx-expression-parse@3.0.1:
resolution: {integrity: sha512-cbqHunsQWnJNE6KhVSMsMeH5H/L9EpymbzqTQ3uLwNCLZ1Q481oWaofqH7nO6V07xlXwY6PhQdQ2IedWx/ZK4Q==}
spdx-expression-parse@4.0.0:
resolution: {integrity: sha512-Clya5JIij/7C6bRR22+tnGXbc4VKlibKSVj2iHvVeX5iMW7s1SIQlqu699JkODJJIhh/pUu8L0/VLh8xflD+LQ==}
spdx-license-ids@3.0.23:
resolution: {integrity: sha512-CWLcCCH7VLu13TgOH+r8p1O/Znwhqv/dbb6lqWy67G+pT1kHmeD/+V36AVb/vq8QMIQwVShJ6Ssl5FPh0fuSdw==}
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
split-ca@1.0.1:
resolution: {integrity: sha512-Q5thBSxp5t8WPTTJQS59LrGqOZqOsrhDGDVm8azCqIBjSBd7nd9o2PM+mDulQQkh8h//4U6hFZnc/mul8t5pWQ==}
first commit
2026-02-21 03:10:39 +00:00
split2@4.2.0:
resolution: {integrity: sha512-UcjcJOWknrNkF6PLX83qcHM6KHgVKNkV62Y8a5uYDVv9ydGQVwAHMKqHdJje1VTWpljG0WYpCDhrCdAOYH4TWg==}
engines: {node: '>= 10.x'}
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
ssh2@1.17.0:
resolution: {integrity: sha512-wPldCk3asibAjQ/kziWQQt1Wh3PgDFpC0XpwclzKcdT1vql6KeYxf5LIt4nlFkUeR8WuphYMKqUA56X4rjbfgQ==}
engines: {node: '>=10.16.0'}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
ssri@13.0.1:
resolution: {integrity: sha512-QUiRf1+u9wPTL/76GTYlKttDEBWV1ga9ZXW8BG6kfdeyyM8LGPix9gROyg9V2+P0xNyF3X2Go526xKFdMZrHSQ==}
engines: {node: ^20.17.0 || >=22.9.0}
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
stack-utils@2.0.6:
resolution: {integrity: sha512-XlkWvfIm6RmsWtNJx+uqtKLS8eqFbxUg0ZzLXqY0caEy9l7hruX8IpiDnjsLavoBgqCCR71TqWO8MaXYheJ3RQ==}
engines: {node: '>=10'}
first commit
2026-02-21 03:10:39 +00:00
stackback@0.0.2:
resolution: {integrity: sha512-1XMJE5fQo1jGH6Y/7ebnwPOBEkIEnT4QF32d5R1+VXdXveM0IBMJt8zfaxX1P3QhVwrYe+576+jkANtSS2mBbw==}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
state-local@1.0.7:
resolution: {integrity: sha512-HTEHMNieakEnoe33shBYcZ7NX83ACUjCu8c40iOGEZsngj9zRnkqS9j1pqQPXwobB0ZcVTk27REb7COQ0UR59w==}
first commit
2026-02-21 03:10:39 +00:00
statuses@2.0.2:
resolution: {integrity: sha512-DvEy55V3DB7uknRo+4iOGT5fP1slR8wQohVdknigZPMpMstaKJQWhwiYBACJE3Ul2pTnATihhBYnRhZQHGBiRw==}
engines: {node: '>= 0.8'}
std-env@3.10.0:
resolution: {integrity: sha512-5GS12FdOZNliM5mAOxFRg7Ir0pWz8MdpYm6AY6VPkGpbA7ZzmbzNcBJQ0GPvvyWgcY7QAhCgf9Uy89I03faLkg==}
feat: add Kubernetes orchestrator for MCP server pod management mcpd can now deploy MCP server instances as Kubernetes pods instead of Docker containers. Set MCPD_ORCHESTRATOR=kubernetes to enable. - Add @kubernetes/client-node with thin wrapper (context enforcement via MCPD_K8S_CONTEXT to prevent multi-cluster mishaps) - Rewrite KubernetesOrchestrator: pod CRUD, pod IP extraction, exec via SPDY (one-shot + interactive), log streaming - Manifest generator: stdin:true for STDIO servers, args (not command) to preserve runner image entrypoint, security hardening - Orchestrator selection in main.ts via MCPD_ORCHESTRATOR env var - 25 unit tests for k8s orchestrator, all 624 tests pass Tested end-to-end on local k3s: - mcpd deployed via Pulumi, creates pods in mcpctl-servers namespace - NetworkPolicy verified: only mcpd can reach MCP server pods - Python runner (uvx) successfully runs aws-documentation-mcp-server Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-08 01:55:13 +01:00
stream-buffers@3.0.3:
resolution: {integrity: sha512-pqMqwQCso0PBJt2PQmDO0cFj0lyqmiwOMiMSkVtRokl7e+ZTRYgDHKnuZNbqjiJXgsg4nuqtD/zxuo9KqTp0Yw==}
engines: {node: '>= 0.10.0'}
streamx@2.25.0:
resolution: {integrity: sha512-0nQuG6jf1w+wddNEEXCF4nTg3LtufWINB5eFEN+5TNZW7KWJp6x87+JFL43vaAUPyCfH1wID+mNVyW6OHtFamg==}
first commit
2026-02-21 03:10:39 +00:00
string-width@4.2.3:
resolution: {integrity: sha512-wKyQRQpjJ0sIp62ErSZdGsjMJWsap5oRNihHhu6G7JVO/9jIB6UyevL+tXuOqrng8j/cxKTWyWUwvSTriiZz/g==}
engines: {node: '>=8'}
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
string-width@7.2.0:
resolution: {integrity: sha512-tsaTIkKW9b4N+AEj+SVA+WhJzV7/zMhcSu78mLKWSk7cXMOSHsBKFWUs0fWwq8QyK3MgJBQRX6Gbi4kYbdvGkQ==}
engines: {node: '>=18'}
string-width@8.2.0:
resolution: {integrity: sha512-6hJPQ8N0V0P3SNmP6h2J99RLuzrWz2gvT7VnK5tKvrNqJoyS9W4/Fb8mo31UiPvy00z7DQXkP2hnKBVav76thw==}
engines: {node: '>=20'}
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
string_decoder@1.3.0:
resolution: {integrity: sha512-hkRX8U1WjJFd8LsDJ2yQ/wWWxaopEsABU1XfkM8A+j0+85JAGppt16cr1Whg6KIbb4okU6Mql6BOj+uup/wKeA==}
first commit
2026-02-21 03:10:39 +00:00
strip-ansi@6.0.1:
resolution: {integrity: sha512-Y38VPSHcqkFrCpFnQ9vuSXmquuv5oXOKpGeT6aGrr3o3Gc9AlVa6JBfUSOCnbxGGZF+/0ooI7KrPuUSztUdU5A==}
engines: {node: '>=8'}
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
strip-ansi@7.1.2:
resolution: {integrity: sha512-gmBGslpoQJtgnMAvOVqGZpEz9dyoKTCzy2nfz/n8aIFhN/jCE/rCmcxabB6jOOHV+0WNnylOxaxBQPSvcWklhA==}
engines: {node: '>=12'}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
strip-final-newline@2.0.0:
resolution: {integrity: sha512-BrpvfNAE3dcvq7ll3xVumzjKjZQ5tI1sEUIKr3Uoks0XUl45St3FlatVqef9prk4jRDzhW6WZg+3bk93y6pLjA==}
engines: {node: '>=6'}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
strip-indent@3.0.0:
resolution: {integrity: sha512-laJTa3Jb+VQpaC6DseHhF7dXVqHTfJPCRDaEbid/drOhgitgYku/letMUqOXFoWV0zIIUbjpdH2t+tYj4bQMRQ==}
engines: {node: '>=8'}
feat(web): bold redesign — Tailwind v4 + shadcn-style primitives + Skills/Proposals/Revisions UI Phase 6 of the Skills + Revisions + Proposals work. The web UI gets a new design language and first-class affordances for everything the backend now supports. ## Visual direction - Tailwind v4 with custom @theme block (oklch tokens). Dark-mode-only (internal tool — light mode doubles QA surface). - Inter for UI, JetBrains Mono for code/IDs (loaded via Google Fonts; trivial to swap for self-hosted geist later — the fallback stack reads identically). - Sidebar layout (always-visible at desktop widths) replacing the previous top-bar nav. Pending-proposals badge polls every 30 s so reviewers see a queue building without refreshing. - Lucide icons throughout. - Spacing and radii on Tailwind defaults. Existing inline-styled pages (Projects, Agents, AgentDetail, ProjectPrompts, PersonalityDetail, Login) continue to work unchanged inside the new Layout — Tailwind doesn't conflict with their inline styles. A follow-up can migrate them incrementally. ## What's added ### Build infra (src/web/) - package.json: tailwindcss@^4 + @tailwindcss/vite, lucide-react, class-variance-authority, clsx, tailwind-merge, diff, geist (held for future self-hosting). - vite.config.ts: registers the @tailwindcss/vite plugin. - src/index.css: Tailwind import + @theme tokens + @layer base. - src/main.tsx: imports index.css. - src/lib/utils.ts: shadcn-style cn() helper. ### shadcn-style primitives (src/components/ui/) Hand-written rather than generated via `npx shadcn` so the repo doesn't depend on a CLI tool that needs interactive runtime: - button.tsx — variants: primary / secondary / ghost / danger / link; sizes: sm / md / lg / icon. - card.tsx — Card + Header/Title/Description/Content/Footer subparts. - badge.tsx — variants: default / info / success / warning / danger / outline. - input.tsx — Input + Textarea + Label. - tabs.tsx — no-dep accessible Tabs (no Radix needed for our use). - separator.tsx — h/v separator with role=separator. ### Diff component (src/components/Diff.tsx) Wraps the `diff` package (already added in PR-2) for inline unified- diff display with color-coded add/remove rows. Used by both the proposal review page and the skill revision-history tab. ### New pages (src/pages/) - Dashboard.tsx — at-a-glance home. Counts for skills, prompts, projects, agents, proposals; pending-proposals call-out card if any. - Skills.tsx — list view, separated into Global vs Project/Agent- scoped sections. - SkillDetail.tsx — name + semver + description; tabs for SKILL.md / Files / Metadata / History. History tab shows revisions with click-to-diff against the live body. - Proposals.tsx — queue with Pending/Approved/Rejected tabs. Pending count is highlighted in amber. - ProposalDetail.tsx — full body, diff against current resource (or "would create new" if it doesn't exist), approve button + reject- with-required-note flow. ### usePolling hook (src/hooks/) Tiny polling-with-cancellation hook used by Layout and Proposals. ### Layout rewrite (src/components/Layout.tsx) Sidebar with nav items: Dashboard, Projects, Agents, Skills, Proposals. Lucide icons. Active-route highlighting via NavLink. Pending-proposals warning badge on the Proposals item. ### Routes (src/App.tsx) New routes: /dashboard, /skills, /skills/:name, /proposals, /proposals/:id. Default redirects to /dashboard. ### API types (src/api.ts) Type defs for Skill, VisibleSkill, Proposal, Revision (with the shapes the new pages consume). ## Tests Existing 7 web tests still pass (Login + api). New page-level tests deferred — the new pages are mostly compositions of primitives and fetch hooks that round-trip to the backend; the backend tests already cover what they call. PR-7 polish can add render-and-click tests if coverage drift surfaces. ## Verification - `pnpm --filter @mcpctl/web build` clean, no warnings. - `pnpm test:run` whole monorepo: 162 test files / 2157 tests green. - Visual smoke deferred — needs a running mcpd to populate the fixtures. Manual smoke tested locally is the next step. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-07 17:54:55 +01:00
styled-jsx@5.1.6:
resolution: {integrity: sha512-qSVyDTeMotdvQYoHWLNGwRFJHC+i+ZvdBRYosOFgC+Wg1vx4frN2/RG/NA7SYqqvKNLf39P2LSRA2pu6n0XYZA==}
engines: {node: '>= 12.0.0'}
peerDependencies:
'@babel/core': '*'
babel-plugin-macros: '*'
react: '>= 16.8.0 || 17.x.x || ^18.0.0-0 || ^19.0.0-0'
peerDependenciesMeta:
'@babel/core':
optional: true
babel-plugin-macros:
optional: true
first commit
2026-02-21 03:10:39 +00:00
supports-color@7.2.0:
resolution: {integrity: sha512-qpCAvRl9stuOHveKsn7HncJRvv501qIacKzQlO/+Lwxc9+0q2wLyv4Dfvt80/DPn2pqOBsJdDiogXGR9+OvwRw==}
engines: {node: '>=8'}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
supports-preserve-symlinks-flag@1.0.0:
resolution: {integrity: sha512-ot0WnXS9fgdkgIcePe6RHNk1WA8+muPa6cSjeR3V8K27q9BB1rTE3R1p7Hv0z1ZyAc8s6Vvv8DIyWf681MAt0w==}
engines: {node: '>= 0.4'}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
symbol-tree@3.2.4:
resolution: {integrity: sha512-9QNk5KwDF+Bvz+PyObkmSYjI5ksVUYtjW7AU22r2NKcfLJcXp96hkDWU3+XndOsUb+AQ9QhfzfCT2O+CNWT5Tw==}
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
tagged-tag@1.0.0:
resolution: {integrity: sha512-yEFYrVhod+hdNyx7g5Bnkkb0G6si8HJurOoOEgC8B/O0uXLHlaey/65KRv6cuWBNhBgHKAROVpc7QyYqE5gFng==}
engines: {node: '>=20'}
feat(web): bold redesign — Tailwind v4 + shadcn-style primitives + Skills/Proposals/Revisions UI Phase 6 of the Skills + Revisions + Proposals work. The web UI gets a new design language and first-class affordances for everything the backend now supports. ## Visual direction - Tailwind v4 with custom @theme block (oklch tokens). Dark-mode-only (internal tool — light mode doubles QA surface). - Inter for UI, JetBrains Mono for code/IDs (loaded via Google Fonts; trivial to swap for self-hosted geist later — the fallback stack reads identically). - Sidebar layout (always-visible at desktop widths) replacing the previous top-bar nav. Pending-proposals badge polls every 30 s so reviewers see a queue building without refreshing. - Lucide icons throughout. - Spacing and radii on Tailwind defaults. Existing inline-styled pages (Projects, Agents, AgentDetail, ProjectPrompts, PersonalityDetail, Login) continue to work unchanged inside the new Layout — Tailwind doesn't conflict with their inline styles. A follow-up can migrate them incrementally. ## What's added ### Build infra (src/web/) - package.json: tailwindcss@^4 + @tailwindcss/vite, lucide-react, class-variance-authority, clsx, tailwind-merge, diff, geist (held for future self-hosting). - vite.config.ts: registers the @tailwindcss/vite plugin. - src/index.css: Tailwind import + @theme tokens + @layer base. - src/main.tsx: imports index.css. - src/lib/utils.ts: shadcn-style cn() helper. ### shadcn-style primitives (src/components/ui/) Hand-written rather than generated via `npx shadcn` so the repo doesn't depend on a CLI tool that needs interactive runtime: - button.tsx — variants: primary / secondary / ghost / danger / link; sizes: sm / md / lg / icon. - card.tsx — Card + Header/Title/Description/Content/Footer subparts. - badge.tsx — variants: default / info / success / warning / danger / outline. - input.tsx — Input + Textarea + Label. - tabs.tsx — no-dep accessible Tabs (no Radix needed for our use). - separator.tsx — h/v separator with role=separator. ### Diff component (src/components/Diff.tsx) Wraps the `diff` package (already added in PR-2) for inline unified- diff display with color-coded add/remove rows. Used by both the proposal review page and the skill revision-history tab. ### New pages (src/pages/) - Dashboard.tsx — at-a-glance home. Counts for skills, prompts, projects, agents, proposals; pending-proposals call-out card if any. - Skills.tsx — list view, separated into Global vs Project/Agent- scoped sections. - SkillDetail.tsx — name + semver + description; tabs for SKILL.md / Files / Metadata / History. History tab shows revisions with click-to-diff against the live body. - Proposals.tsx — queue with Pending/Approved/Rejected tabs. Pending count is highlighted in amber. - ProposalDetail.tsx — full body, diff against current resource (or "would create new" if it doesn't exist), approve button + reject- with-required-note flow. ### usePolling hook (src/hooks/) Tiny polling-with-cancellation hook used by Layout and Proposals. ### Layout rewrite (src/components/Layout.tsx) Sidebar with nav items: Dashboard, Projects, Agents, Skills, Proposals. Lucide icons. Active-route highlighting via NavLink. Pending-proposals warning badge on the Proposals item. ### Routes (src/App.tsx) New routes: /dashboard, /skills, /skills/:name, /proposals, /proposals/:id. Default redirects to /dashboard. ### API types (src/api.ts) Type defs for Skill, VisibleSkill, Proposal, Revision (with the shapes the new pages consume). ## Tests Existing 7 web tests still pass (Login + api). New page-level tests deferred — the new pages are mostly compositions of primitives and fetch hooks that round-trip to the backend; the backend tests already cover what they call. PR-7 polish can add render-and-click tests if coverage drift surfaces. ## Verification - `pnpm --filter @mcpctl/web build` clean, no warnings. - `pnpm test:run` whole monorepo: 162 test files / 2157 tests green. - Visual smoke deferred — needs a running mcpd to populate the fixtures. Manual smoke tested locally is the next step. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-07 17:54:55 +01:00
tailwind-merge@3.5.0:
resolution: {integrity: sha512-I8K9wewnVDkL1NTGoqWmVEIlUcB9gFriAEkXkfCjX5ib8ezGxtR3xD7iZIxrfArjEsH7F1CHD4RFUtxefdqV/A==}
tailwindcss@4.2.4:
resolution: {integrity: sha512-HhKppgO81FQof5m6TEnuBWCZGgfRAWbaeOaGT00KOy/Pf/j6oUihdvBpA7ltCeAvZpFhW3j0PTclkxsd4IXYDA==}
tapable@2.3.3:
resolution: {integrity: sha512-uxc/zpqFg6x7C8vOE7lh6Lbda8eEL9zmVm/PLeTPBRhh1xCgdWaQ+J1CUieGpIfm2HdtsUpRv+HshiasBMcc6A==}
engines: {node: '>=6'}
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
tar-fs@2.1.4:
resolution: {integrity: sha512-mDAjwmZdh7LTT6pNleZ05Yt65HC3E+NiQzl672vQG38jIrehtJk/J3mNwIg+vShQPcLF/LV7CMnDW6vjj6sfYQ==}
feat: add Kubernetes orchestrator for MCP server pod management mcpd can now deploy MCP server instances as Kubernetes pods instead of Docker containers. Set MCPD_ORCHESTRATOR=kubernetes to enable. - Add @kubernetes/client-node with thin wrapper (context enforcement via MCPD_K8S_CONTEXT to prevent multi-cluster mishaps) - Rewrite KubernetesOrchestrator: pod CRUD, pod IP extraction, exec via SPDY (one-shot + interactive), log streaming - Manifest generator: stdin:true for STDIO servers, args (not command) to preserve runner image entrypoint, security hardening - Orchestrator selection in main.ts via MCPD_ORCHESTRATOR env var - 25 unit tests for k8s orchestrator, all 624 tests pass Tested end-to-end on local k3s: - mcpd deployed via Pulumi, creates pods in mcpctl-servers namespace - NetworkPolicy verified: only mcpd can reach MCP server pods - Python runner (uvx) successfully runs aws-documentation-mcp-server Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-08 01:55:13 +01:00
tar-fs@3.1.2:
resolution: {integrity: sha512-QGxxTxxyleAdyM3kpFs14ymbYmNFrfY+pHj7Z8FgtbZ7w2//VAgLMac7sT6nRpIHjppXO2AwwEOg0bPFVRcmXw==}
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
tar-stream@2.2.0:
resolution: {integrity: sha512-ujeqbceABgwMZxEJnk2HDY2DlnUZ+9oEcb1KzTVfYHio0UE6dG71n60d8D2I4qNvleWrrXpmjpt7vZeF1LnMZQ==}
engines: {node: '>=6'}
feat: add Kubernetes orchestrator for MCP server pod management mcpd can now deploy MCP server instances as Kubernetes pods instead of Docker containers. Set MCPD_ORCHESTRATOR=kubernetes to enable. - Add @kubernetes/client-node with thin wrapper (context enforcement via MCPD_K8S_CONTEXT to prevent multi-cluster mishaps) - Rewrite KubernetesOrchestrator: pod CRUD, pod IP extraction, exec via SPDY (one-shot + interactive), log streaming - Manifest generator: stdin:true for STDIO servers, args (not command) to preserve runner image entrypoint, security hardening - Orchestrator selection in main.ts via MCPD_ORCHESTRATOR env var - 25 unit tests for k8s orchestrator, all 624 tests pass Tested end-to-end on local k3s: - mcpd deployed via Pulumi, creates pods in mcpctl-servers namespace - NetworkPolicy verified: only mcpd can reach MCP server pods - Python runner (uvx) successfully runs aws-documentation-mcp-server Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-08 01:55:13 +01:00
tar-stream@3.1.8:
resolution: {integrity: sha512-U6QpVRyCGHva435KoNWy9PRoi2IFYCgtEhq9nmrPPpbRacPs9IH4aJ3gbrFC8dPcXvdSZ4XXfXT5Fshbp2MtlQ==}
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
tar@6.2.1:
resolution: {integrity: sha512-DZ4yORTwrbTj/7MZYq2w+/ZFdI6OZ/f9SFHR+71gIVUZhOQPHzVCLpvRnPgyaMpfWxxk/4ONva3GQSyNIKRv6A==}
engines: {node: '>=10'}
deprecated: Old versions of tar are not supported, and contain widely publicized security vulnerabilities, which have been fixed in the current version. Please update. Support for old versions may be purchased (at exorbitant rates) by contacting i@izs.me
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
tar@7.5.20:
resolution: {integrity: sha512-9FcyK4PA6+WbzlTM9WhQm6vB5W7cP7dUiPsv1g7YDwEQnQ1CGpK3MGlKk/ITVWMk05kHZuBhmVhiv8LZoy/PFQ==}
engines: {node: '>=18'}
feat: add Kubernetes orchestrator for MCP server pod management mcpd can now deploy MCP server instances as Kubernetes pods instead of Docker containers. Set MCPD_ORCHESTRATOR=kubernetes to enable. - Add @kubernetes/client-node with thin wrapper (context enforcement via MCPD_K8S_CONTEXT to prevent multi-cluster mishaps) - Rewrite KubernetesOrchestrator: pod CRUD, pod IP extraction, exec via SPDY (one-shot + interactive), log streaming - Manifest generator: stdin:true for STDIO servers, args (not command) to preserve runner image entrypoint, security hardening - Orchestrator selection in main.ts via MCPD_ORCHESTRATOR env var - 25 unit tests for k8s orchestrator, all 624 tests pass Tested end-to-end on local k3s: - mcpd deployed via Pulumi, creates pods in mcpctl-servers namespace - NetworkPolicy verified: only mcpd can reach MCP server pods - Python runner (uvx) successfully runs aws-documentation-mcp-server Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-08 01:55:13 +01:00
teex@1.0.1:
resolution: {integrity: sha512-eYE6iEI62Ni1H8oIa7KlDU6uQBtqr4Eajni3wX7rpfXD8ysFx8z0+dri+KWEPWpBsxXfxu58x/0jvTVT1ekOSg==}
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
terminal-size@4.0.1:
resolution: {integrity: sha512-avMLDQpUI9I5XFrklECw1ZEUPJhqzcwSWsyyI8blhRLT+8N1jLJWLWWYQpB2q2xthq8xDvjZPISVh53T/+CLYQ==}
engines: {node: '>=18'}
feat: add Kubernetes orchestrator for MCP server pod management mcpd can now deploy MCP server instances as Kubernetes pods instead of Docker containers. Set MCPD_ORCHESTRATOR=kubernetes to enable. - Add @kubernetes/client-node with thin wrapper (context enforcement via MCPD_K8S_CONTEXT to prevent multi-cluster mishaps) - Rewrite KubernetesOrchestrator: pod CRUD, pod IP extraction, exec via SPDY (one-shot + interactive), log streaming - Manifest generator: stdin:true for STDIO servers, args (not command) to preserve runner image entrypoint, security hardening - Orchestrator selection in main.ts via MCPD_ORCHESTRATOR env var - 25 unit tests for k8s orchestrator, all 624 tests pass Tested end-to-end on local k3s: - mcpd deployed via Pulumi, creates pods in mcpctl-servers namespace - NetworkPolicy verified: only mcpd can reach MCP server pods - Python runner (uvx) successfully runs aws-documentation-mcp-server Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-08 01:55:13 +01:00
text-decoder@1.2.7:
resolution: {integrity: sha512-vlLytXkeP4xvEq2otHeJfSQIRyWxo/oZGEbXrtEEF9Hnmrdly59sUbzZ/QgyWuLYHctCHxFF4tRQZNQ9k60ExQ==}
first commit
2026-02-21 03:10:39 +00:00
thread-stream@4.0.0:
resolution: {integrity: sha512-4iMVL6HAINXWf1ZKZjIPcz5wYaOdPhtO8ATvZ+Xqp3BTdaqtAwQkNmKORqcIo5YkQqGXq5cwfswDwMqqQNrpJA==}
engines: {node: '>=20'}
tinybench@2.9.0:
resolution: {integrity: sha512-0+DUvqWMValLmha6lr4kD8iAMK1HzV0/aKnCtWb9v9641TnP/MFb7Pc2bxoxQjTXAErryXVgUOfv2YqNllqGeg==}
tinyexec@1.0.2:
resolution: {integrity: sha512-W/KYk+NFhkmsYpuHq5JykngiOCnxeVL8v8dFnqxSD8qEEdRfXk1SDM6JzNqcERbcGYj9tMrDQBYV9cjgnunFIg==}
engines: {node: '>=18'}
tinyglobby@0.2.15:
resolution: {integrity: sha512-j2Zq4NyQYG5XMST4cbs02Ak8iJUdxRM0XI5QyxXuZOzKOINmWurp3smXu3y5wDcJrptwpSjgXHzIQxR0omXljQ==}
engines: {node: '>=12.0.0'}
tinyrainbow@3.0.3:
resolution: {integrity: sha512-PSkbLUoxOFRzJYjjxHJt9xro7D+iilgMX/C9lawzVuYiIdcihh9DXmVibBe8lmcFrRi/VzlPjBxbN7rH24q8/Q==}
engines: {node: '>=14.0.0'}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
tldts-core@7.0.28:
resolution: {integrity: sha512-7W5Efjhsc3chVdFhqtaU0KtK32J37Zcr9RKtID54nG+tIpcY79CQK/veYPODxtD/LJ4Lue66jvrQzIX2Z2/pUQ==}
tldts@7.0.28:
resolution: {integrity: sha512-+Zg3vWhRUv8B1maGSTFdev9mjoo8Etn2Ayfs4cnjlD3CsGkxXX4QyW3j2WJ0wdjYcYmy7Lx2RDsZMhgCWafKIw==}
hasBin: true
first commit
2026-02-21 03:10:39 +00:00
toad-cache@3.7.0:
resolution: {integrity: sha512-/m8M+2BJUpoJdgAHoG+baCwBT+tf2VraSfkBgl0Y00qIWt41DJ8R5B8nsEw0I58YwF5IZH6z24/2TobDKnqSWw==}
engines: {node: '>=12'}
toidentifier@1.0.1:
resolution: {integrity: sha512-o5sSPKEkg/DIQNmH43V0/uerLrpzVedkUh8tGNvaeXpfpuwjKenlSox/2O/BTlZUtEe+JG7s5YhEz608PlAHRA==}
engines: {node: '>=0.6'}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
toml@4.3.0:
resolution: {integrity: sha512-lVb8X9BsPVuH0M4BKeS91tXAmJvCjQ5UIyAbQFaxkKGyUFK2RPkhwaFSQH8vbpl1d23eu/IBH+dwVMHWaq9A5A==}
engines: {node: '>=20'}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
tough-cookie@6.0.1:
resolution: {integrity: sha512-LktZQb3IeoUWB9lqR5EWTHgW/VTITCXg4D21M+lvybRVdylLrRMnqaIONLVb5mav8vM19m44HIcGq4qASeu2Qw==}
engines: {node: '>=16'}
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
tr46@0.0.3:
resolution: {integrity: sha512-N3WMsuqV66lT30CrXNbEjx4GEwlow3v6rr4mCcv6prnfwhS01rkgyFdjPNBYd9br7LpXV1+Emh01fHnq2Gdgrw==}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
tr46@6.0.0:
resolution: {integrity: sha512-bLVMLPtstlZ4iMQHpFHTR7GAGj2jxi8Dg0s2h2MafAE4uSWF98FC/3MomU51iQAMf8/qDUbKWf5GxuvvVcXEhw==}
engines: {node: '>=20'}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
treeverse@3.0.0:
resolution: {integrity: sha512-gcANaAnd2QDZFmHFEOF4k7uc1J/6a6z3DJMd/QwEyxLoKGiptJRwid582r7QIsFlFMIZ3SnxfS52S4hm2DHkuQ==}
engines: {node: ^14.17.0 || ^16.13.0 || >=18.0.0}
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
ts-algebra@2.0.0:
resolution: {integrity: sha512-FPAhNPFMrkwz76P7cdjdmiShwMynZYN6SgOujD1urY4oNm80Ou9oMdmbR45LotcKOXoy7wSmHkRFE6Mxbrhefw==}
first commit
2026-02-21 03:10:39 +00:00
ts-api-utils@2.4.0:
resolution: {integrity: sha512-3TaVTaAv2gTiMB35i3FiGJaRfwb3Pyn/j3m/bfAvGe8FB7CF6u+LMYqYlDh7reQf7UNvoTvdfAqHGmPGOSsPmA==}
engines: {node: '>=18.12'}
peerDependencies:
typescript: '>=4.8.4'
tslib@2.8.1:
resolution: {integrity: sha512-oJFu94HQb+KVduSUQL7wnpmqnfmLsOA/nAh6b6EH0wCEoK0/mPeXU6c3wKDV83MkOuHPRHtSXKKU99IBazS/2w==}
tsx@4.21.0:
resolution: {integrity: sha512-5C1sg4USs1lfG0GFb2RLXsdpXqBSEhAaA/0kPL01wxzpMqLILNxIxIOKiILz+cdg/pLnOUxFYOR5yhHU666wbw==}
engines: {node: '>=18.0.0'}
hasBin: true
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
tuf-js@4.1.0:
resolution: {integrity: sha512-50QV99kCKH5P/Vs4E2Gzp7BopNV+KzTXqWeaxrfu5IQJBOULRsTIS9seSsOVT8ZnGXzCyx55nYWAi4qJzpZKEQ==}
engines: {node: ^20.17.0 || >=22.9.0}
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
tweetnacl@0.14.5:
resolution: {integrity: sha512-KXXFFdAbFXY4geFIwoyNK+f5Z1b7swfXABfL7HXCmoIWMKU3dmS26672A4EeQtDzLKy7SXmfBu51JolvEKwtGA==}
first commit
2026-02-21 03:10:39 +00:00
type-check@0.4.0:
resolution: {integrity: sha512-XleUoc9uwGXqjWwXaUTZAmzMcFZ5858QA2vvx1Ur5xIcixXIP+8LnFDgRplU30us6teqdlskFfu+ae4K79Ooew==}
engines: {node: '>= 0.8.0'}
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
type-fest@5.4.4:
resolution: {integrity: sha512-JnTrzGu+zPV3aXIUhnyWJj4z/wigMsdYajGLIYakqyOW1nPllzXEJee0QQbHj+CTIQtXGlAjuK0UY+2xTyjVAw==}
engines: {node: '>=20'}
first commit
2026-02-21 03:10:39 +00:00
type-is@2.0.1:
resolution: {integrity: sha512-OZs6gsjF4vMp32qrCbiVSkrFmXtG/AZhY3t0iAMrMBiAZyV9oALtXO8hsrHbMXF9x6L3grlFuwW2oAz7cav+Gw==}
engines: {node: '>= 0.6'}
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
typebox@1.3.11:
resolution: {integrity: sha512-tZGKIS02Opbh4EYMmAEVuXl+y3EQJ9ZlkitLZ1CmFCY4ZOqr/xWR9dDSCFmIjNDICGMWkOqMh6WxGTyRXqcSGQ==}
typebox@1.3.7:
resolution: {integrity: sha512-meKuifc33Pccx0O6PdIzYMq3Og8zvP4TIi/a+Bw3AEMZMxOD0+RHGQvpglEe6Zdy3wZ8nqn/j95h8LUZLk/6Hg==}
first commit
2026-02-21 03:10:39 +00:00
typescript@5.9.3:
resolution: {integrity: sha512-jl1vZzPDinLr9eUt3J/t7V6FgNEw9QjvBPdysz9KfQDD41fQrC2Y4vKQdiaUpFT4bXlb1RHhLpp8wtm6M5TgSw==}
engines: {node: '>=14.17'}
hasBin: true
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
undici-types@5.26.5:
resolution: {integrity: sha512-JlCMO+ehdEIKqlFxk6IfVoAUVmgz7cU7zD/h9XZ0qzeosSHmUJVOzSQvvYSYWXkFXC+IfLKSIffhv0sVZup6pA==}
feat: add Kubernetes orchestrator for MCP server pod management mcpd can now deploy MCP server instances as Kubernetes pods instead of Docker containers. Set MCPD_ORCHESTRATOR=kubernetes to enable. - Add @kubernetes/client-node with thin wrapper (context enforcement via MCPD_K8S_CONTEXT to prevent multi-cluster mishaps) - Rewrite KubernetesOrchestrator: pod CRUD, pod IP extraction, exec via SPDY (one-shot + interactive), log streaming - Manifest generator: stdin:true for STDIO servers, args (not command) to preserve runner image entrypoint, security hardening - Orchestrator selection in main.ts via MCPD_ORCHESTRATOR env var - 25 unit tests for k8s orchestrator, all 624 tests pass Tested end-to-end on local k3s: - mcpd deployed via Pulumi, creates pods in mcpctl-servers namespace - NetworkPolicy verified: only mcpd can reach MCP server pods - Python runner (uvx) successfully runs aws-documentation-mcp-server Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-08 01:55:13 +01:00
undici-types@7.16.0:
resolution: {integrity: sha512-Zz+aZWSj8LE6zoxD+xrjh4VfkIG8Ya6LvYkZqtUQGJPZjYl53ypCaUwWqo7eI0x66KBGeRo+mlBEkMSeSZ38Nw==}
feat: implement MCP registry client with multi-source search Add registry client that queries Official, Glama, and Smithery MCP registries with caching, request deduplication, retry logic, and result ranking/dedup. Includes 53 tests covering all components. Also fix null priority values in cancelled tasks (19-21) that broke Task Master, and add new tasks 25-27 for registry completion and CLI discover/install commands. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 03:46:14 +00:00
undici-types@7.18.2:
resolution: {integrity: sha512-AsuCzffGHJybSaRrmr5eHr81mwJU3kjw6M+uprWvCXiNeN9SOGwQ3Jn8jb8m3Z6izVgknn1R0FTCEAP2QrLY/w==}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
undici@6.27.0:
resolution: {integrity: sha512-YmfV3YnEDzXRC5lZ2jWtWWHKGUm1zIt8AhesR1tens+HTNv+YZlN/dp6G727LOvMJ8xjP9Be7Y2Sdr96LDm+pg==}
engines: {node: '>=18.17'}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
undici@7.25.0:
resolution: {integrity: sha512-xXnp4kTyor2Zq+J1FfPI6Eq3ew5h6Vl0F/8d9XU5zZQf1tX9s2Su1/3PiMmUANFULpmksxkClamIZcaUqryHsQ==}
engines: {node: '>=20.18.1'}
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
undici@8.9.0:
resolution: {integrity: sha512-aWZpUj7XoGonMClx4gdDRfgBjqeA+F473aDmROQQbM9n6PRfK/u1q/a0X4wMTgcHfT8H6fpbt98PFuDUwFg2YA==}
engines: {node: '>=22.19.0'}
first commit
2026-02-21 03:10:39 +00:00
unpipe@1.0.0:
resolution: {integrity: sha512-pjy2bYhSsufwWlKwPc+l3cN7+wuJlK6uz0YdJEOlQDbl6jo/YlPi4mb8agUkVC8BF7V8NuzeyPNqRksA3hztKQ==}
engines: {node: '>= 0.8'}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
upath@1.2.0:
resolution: {integrity: sha512-aZwGpamFO61g3OlfT7OQCHqhGnW43ieH9WZeP7QxN/G/jS4jfqUkZxoryvJgVPEcrl5NL/ggHsSmLMHuH64Lhg==}
engines: {node: '>=4'}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
update-browserslist-db@1.2.3:
resolution: {integrity: sha512-Js0m9cx+qOgDxo0eMiFGEueWztz+d4+M3rGlmKPT+T4IS/jP4ylw3Nwpu6cpTTP8R1MAC1kF4VbdLt3ARf209w==}
hasBin: true
peerDependencies:
browserslist: '>= 4.21.0'
first commit
2026-02-21 03:10:39 +00:00
uri-js@4.4.1:
resolution: {integrity: sha512-7rKUyy33Q1yc98pQ1DAmLtwX109F7TIfWlW1Ydo8Wl1ii1SeHieeh0HHfPeL2fMXK6z0s8ecKs9frCuLJvndBg==}
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
util-deprecate@1.0.2:
resolution: {integrity: sha512-EPD5q1uXyFxJpCrLnCc1nHnq3gOa6DZBocAIiI2TaSCA7VCJ1UJDMagCzIkXNsUYfD1daK//LTEQ8xiIbrHtcw==}
uuid@10.0.0:
resolution: {integrity: sha512-8XkAphELsDnEGrDxUOHB3RGvXz6TeuYSGEZBOjtTtPm2lwhGBjLgOzLHB63IUWfBpNucQjND6d3AOudO+H3RWQ==}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
deprecated: uuid@10 and below is no longer supported. For ESM codebases, update to uuid@latest. For CommonJS codebases, use uuid@11 (but be aware this version will likely be deprecated in 2028).
hasBin: true
uuid@14.0.1:
resolution: {integrity: sha512-6ZxzVpzDXDa3bJWaHilVayA+BH/1zmxCJoVgvmqJnid/gPoKHxUrS/aC/T6LGQtNHT+XHG9fXPJB4d+IrU30Ew==}
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
hasBin: true
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
validate-npm-package-license@3.0.4:
resolution: {integrity: sha512-DpKm2Ui/xN7/HQKCtpZxoRWBhZ9Z0kqtygG8XCgNQ8ZlDnxuQmWhj566j8fN4Cu3/JmbhsDo7fcAJq4s9h27Ew==}
validate-npm-package-name@7.0.2:
resolution: {integrity: sha512-hVDIBwsRruT73PbK7uP5ebUt+ezEtCmzZz3F59BSr2F6OVFnJ/6h8liuvdLrQ88Xmnk6/+xGGuq+pG9WwTuy3A==}
engines: {node: ^20.17.0 || >=22.9.0}
first commit
2026-02-21 03:10:39 +00:00
vary@1.1.2:
resolution: {integrity: sha512-BNGbWLfd0eUPabhkXUVm0j8uuvREyTh5ovRa/dyow/BqAbZJyC+5fU+IzQOzmAKzYqYRAISoRhdQr3eIZ/PXqg==}
engines: {node: '>= 0.8'}
vite@7.3.1:
resolution: {integrity: sha512-w+N7Hifpc3gRjZ63vYBXA56dvvRlNWRczTdmCBBa+CotUzAPf5b7YMdMR/8CQoeYE5LX3W4wj6RYTgonm1b9DA==}
engines: {node: ^20.19.0 || >=22.12.0}
hasBin: true
peerDependencies:
'@types/node': ^20.19.0 || >=22.12.0
jiti: '>=1.21.0'
less: ^4.0.0
lightningcss: ^1.21.0
sass: ^1.70.0
sass-embedded: ^1.70.0
stylus: '>=0.54.8'
sugarss: ^5.0.0
terser: ^5.16.0
tsx: ^4.8.1
yaml: ^2.4.2
peerDependenciesMeta:
'@types/node':
optional: true
jiti:
optional: true
less:
optional: true
lightningcss:
optional: true
sass:
optional: true
sass-embedded:
optional: true
stylus:
optional: true
sugarss:
optional: true
terser:
optional: true
tsx:
optional: true
yaml:
optional: true
vitest@4.0.18:
resolution: {integrity: sha512-hOQuK7h0FGKgBAas7v0mSAsnvrIgAvWmRFjmzpJ7SwFHH3g1k2u37JtYwOwmEKhK6ZO3v9ggDBBm0La1LCK4uQ==}
engines: {node: ^20.0.0 || ^22.0.0 || >=24.0.0}
hasBin: true
peerDependencies:
'@edge-runtime/vm': '*'
'@opentelemetry/api': ^1.9.0
'@types/node': ^20.0.0 || ^22.0.0 || >=24.0.0
'@vitest/browser-playwright': 4.0.18
'@vitest/browser-preview': 4.0.18
'@vitest/browser-webdriverio': 4.0.18
'@vitest/ui': 4.0.18
happy-dom: '*'
jsdom: '*'
peerDependenciesMeta:
'@edge-runtime/vm':
optional: true
'@opentelemetry/api':
optional: true
'@types/node':
optional: true
'@vitest/browser-playwright':
optional: true
'@vitest/browser-preview':
optional: true
'@vitest/browser-webdriverio':
optional: true
'@vitest/ui':
optional: true
happy-dom:
optional: true
jsdom:
optional: true
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
w3c-xmlserializer@5.0.0:
resolution: {integrity: sha512-o8qghlI8NZHU1lLPrpi2+Uq7abh4GGPpYANlalzWxyWteJOCsr/P+oPBA49TOLu5FTZO4d3F9MnWJfiMo4BkmA==}
engines: {node: '>=18'}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
walk-up-path@4.0.0:
resolution: {integrity: sha512-3hu+tD8YzSLGuFYtPRb48vdhKMi0KQV5sn+uWr8+7dMEq/2G/dtLrdDinkLjqq5TIbIBjYJ4Ax/n3YiaW7QM8A==}
engines: {node: 20 || >=22}
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
web-streams-polyfill@3.3.3:
resolution: {integrity: sha512-d2JWLCivmZYTSIoge9MsgFCZrt571BikcWGYkjC1khllbTeDlGqZ2D8vD8E/lJa8WGWbb7Plm8/XJYV7IJHZZw==}
engines: {node: '>= 8'}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
web-tree-sitter@0.25.10:
resolution: {integrity: sha512-Y09sF44/13XvgVKgO2cNDw5rGk6s26MgoZPXLESvMXeefBf7i6/73eFurre0IsTW6E14Y0ArIzhUMmjoc7xyzA==}
peerDependencies:
'@types/emscripten': ^1.40.0
peerDependenciesMeta:
'@types/emscripten':
optional: true
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
webidl-conversions@3.0.1:
resolution: {integrity: sha512-2JAn3z8AR6rjK8Sm8orRC0h/bcl/DqL7tRPdGZ4I1CjdF+EaMLmYxBHyXuKL849eucPFhvBoxMsflfOb8kxaeQ==}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
webidl-conversions@8.0.1:
resolution: {integrity: sha512-BMhLD/Sw+GbJC21C/UgyaZX41nPt8bUTg+jWyDeg7e7YN4xOM05YPSIXceACnXVtqyEw/LMClUQMtMZ+PGGpqQ==}
engines: {node: '>=20'}
whatwg-mimetype@5.0.0:
resolution: {integrity: sha512-sXcNcHOC51uPGF0P/D4NVtrkjSU2fNsm9iog4ZvZJsL3rjoDAzXZhkm2MWt1y+PUdggKAYVoMAIYcs78wJ51Cw==}
engines: {node: '>=20'}
whatwg-url@16.0.1:
resolution: {integrity: sha512-1to4zXBxmXHV3IiSSEInrreIlu02vUOvrhxJJH5vcxYTBDAx51cqZiKdyTxlecdKNSjj8EcxGBxNf6Vg+945gw==}
engines: {node: ^20.19.0 || ^22.12.0 || >=24.0.0}
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
whatwg-url@5.0.0:
resolution: {integrity: sha512-saE57nupxk6v3HY35+jzBwYa0rKSy0XR8JSxZPwgLr7ys0IBzhGviA1/TUGJLmSVqs8pb9AnvICXEuOHLprYTw==}
first commit
2026-02-21 03:10:39 +00:00
which@2.0.2:
resolution: {integrity: sha512-BLI3Tl1TW3Pvl70l3yq3Y64i+awpwXqsGBYWkkqMtnbXgrMD+yj7rhW0kuEDxzJaYXGjEW5ogapKNMEKNMjibA==}
engines: {node: '>= 8'}
hasBin: true
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
which@6.0.1:
resolution: {integrity: sha512-oGLe46MIrCRqX7ytPUf66EAYvdeMIZYn3WaocqqKZAxrBpkqHfL/qvTyJ/bTk5+AqHCjXmrv3CEWgy368zhRUg==}
engines: {node: ^20.17.0 || >=22.9.0}
hasBin: true
first commit
2026-02-21 03:10:39 +00:00
why-is-node-running@2.3.0:
resolution: {integrity: sha512-hUrmaWBdVDcxvYqnyh09zunKzROWjbZTiNy8dBEjkS7ehEDQibXJ7XvlmtbwuTclUiIyN+CyXQD4Vmko8fNm8w==}
engines: {node: '>=8'}
hasBin: true
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
wide-align@1.1.5:
resolution: {integrity: sha512-eDMORYaPNZ4sQIuuYPDHdQvf4gyCF9rEEV/yPxGfwPkRodwEgiMUUXTx/dex+Me0wxx53S+NgUHaP7y3MGlDmg==}
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
widest-line@6.0.0:
resolution: {integrity: sha512-U89AsyEeAsyoF0zVJBkG9zBgekjgjK7yk9sje3F4IQpXBJ10TF6ByLlIfjMhcmHMJgHZI4KHt4rdNfktzxIAMA==}
engines: {node: '>=20'}
first commit
2026-02-21 03:10:39 +00:00
word-wrap@1.2.5:
resolution: {integrity: sha512-BN22B5eaMMI9UMtjrGd5g5eCYPpCPDUy0FJXbYsaT5zYxjFOckS53SQDE3pWkVoWpHXVb3BrYcEN4Twa55B5cA==}
engines: {node: '>=0.10.0'}
wrap-ansi@6.2.0:
resolution: {integrity: sha512-r6lPcBGxZXlIcymEu7InxDMhdW0KDxpLgoFLcguasxCaJ/SOIZwINatK9KY/tf+ZrlywOKU0UDj3ATXUBfxJXA==}
engines: {node: '>=8'}
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
wrap-ansi@7.0.0:
resolution: {integrity: sha512-YVGIj2kamLSTxw6NsZjoBxfSwsn0ycdesmc4p+Q21c5zPuZ1pl+NfxVdxPtdHvmNVOQ6XSYG4AUtyt/Fi7D16Q==}
engines: {node: '>=10'}
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
wrap-ansi@9.0.2:
resolution: {integrity: sha512-42AtmgqjV+X1VpdOfyTGOYRi0/zsoLqtXQckTmqTeybT+BDIbM/Guxo7x3pE2vtpr1ok6xRqM9OpBe+Jyoqyww==}
engines: {node: '>=18'}
first commit
2026-02-21 03:10:39 +00:00
wrappy@1.0.2:
resolution: {integrity: sha512-l4Sp/DRseor9wL6EvV2+TuQn63dMkPjZ/sp9XkghTEbV9KlPS1xUsZ3u7/IQO4wxtcFB4bgpQPRcR3QCvezPcQ==}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
write-file-atomic@7.0.1:
resolution: {integrity: sha512-OTIk8iR8/aCRWBqvxrzxR0hgxWpnYBblY1S5hDWBQfk/VFmJwzmJgQFN3WsoUKHISv2eAwe+PpbUzyL1CKTLXg==}
engines: {node: ^20.17.0 || >=22.9.0}
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
ws@8.19.0:
resolution: {integrity: sha512-blAT2mjOEIi0ZzruJfIhb3nps74PRWTCz1IjglWEEpQl5XS/UNama6u2/rjFkDDouqr4L67ry+1aGIALViWjDg==}
engines: {node: '>=10.0.0'}
peerDependencies:
bufferutil: ^4.0.1
utf-8-validate: '>=5.0.2'
peerDependenciesMeta:
bufferutil:
optional: true
utf-8-validate:
optional: true
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
xml-name-validator@5.0.0:
resolution: {integrity: sha512-EvGK8EJ3DhaHfbRlETOWAS5pO9MZITeauHKJyb8wyajUfQUenkIg2MvLDTZ4T/TgIcm3HU0TFBgWWboAZ30UHg==}
engines: {node: '>=18'}
xmlchars@2.2.0:
resolution: {integrity: sha512-JZnDKK8B0RCDw84FNdDAIpZK+JuJw+s7Lz8nksI7SIuU3UXJJslUthsi+uWBUYOwPFwW7W7PRLRfUKpxjtjFCw==}
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
y18n@5.0.8:
resolution: {integrity: sha512-0pfFzegeDWJHJIAmTLRP2DwHjdF5s7jo9tuztdQxAhINCdvS+3nGINqPd00AphqJR/0LhANUS6/+7SCb98YOfA==}
engines: {node: '>=10'}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
yallist@3.1.1:
resolution: {integrity: sha512-a4UGQaWPH59mOXUYnAG2ewncQS4i4F43Tv3JoAM+s2VDAmS9NsK8GpDMLrCHPksFT7h3K6TOoUNn2pb7RoXx4g==}
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
yallist@4.0.0:
resolution: {integrity: sha512-3wdGidZyq5PB084XLES5TpOSRA3wjXAlIWMhum2kRcv/41Sn2emQ0dycQW4uZXLejwKvg6EsvbdlVL+FYEct7A==}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
yallist@5.0.0:
resolution: {integrity: sha512-YgvUTfwqyc7UXVMrB+SImsVYSmTS8X/tSrtdNZMImM+n7+QTriRXyXim0mBrTXNeqzVF0KWGgHPeiyViFFrNDw==}
engines: {node: '>=18'}
feat: eager vLLM warmup and smart page titles in paginate stage - Add warmup() to LlmProvider interface for eager subprocess startup - ManagedVllmProvider.warmup() starts vLLM in background on project load - ProviderRegistry.warmupAll() triggers all managed providers - NamedProvider proxies warmup() to inner provider - paginate stage generates LLM-powered descriptive page titles when available, cached by content hash, falls back to generic "Page N" - project-mcp-endpoint calls warmupAll() on router creation so vLLM is loading while the session initializes Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-03-03 19:07:39 +00:00
yaml@2.8.2:
resolution: {integrity: sha512-mplynKqc1C2hTVYxd0PU2xQAc22TI1vShAYGksCCfxbn/dFwnHTNi1bvYsBTkhdUNtGIf5xNOg938rrSSYvS9A==}
engines: {node: '>= 14.6'}
hasBin: true
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
yaml@2.9.0:
resolution: {integrity: sha512-2AvhNX3mb8zd6Zy7INTtSpl1F15HW6Wnqj0srWlkKLcpYl/gMIMJiyuGq2KeI2YFxUPjdlB+3Lc10seMLtL4cA==}
engines: {node: '>= 14.6'}
hasBin: true
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
yargs-parser@21.1.1:
resolution: {integrity: sha512-tVpsJW7DdjecAiFpbIB1e3qxIQsE6NoPc5/eTdrbbIC4h0LVsWhnoa3g+m2HclBIujHzsxZ4VJVA+GUuc2/LBw==}
engines: {node: '>=12'}
yargs@17.7.2:
resolution: {integrity: sha512-7dSzzRQ++CKnNI/krKnYRV7JKKPUXMEh61soaHKg9mrWEhzFWhFnxPxGl+69cD1Ou63C13NUPCnmIcrvqCuM6w==}
engines: {node: '>=12'}
first commit
2026-02-21 03:10:39 +00:00
yocto-queue@0.1.0:
resolution: {integrity: sha512-rVksvsnNCdJ/ohGc6xgPwyN8eheCxsiLM8mxuE/t/mOVqJewPuO1miLpTHQiRgTKCLexL4MeAFVagts7HmNZ2Q==}
engines: {node: '>=10'}
yoctocolors-cjs@2.1.3:
resolution: {integrity: sha512-U/PBtDf35ff0D8X8D0jfdzHYEPFxAI7jJlxZXwCSez5M3190m+QobIfh+sWDWSHMCWWJN2AWamkegn6vr6YBTw==}
engines: {node: '>=18'}
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
yoga-layout@3.2.1:
resolution: {integrity: sha512-0LPOt3AxKqMdFBZA3HBAt/t/8vIKq7VaQYbuA8WxCgung+p9TVyKRYdpvCb80HcdTN2NkbIKbhNwKUfm3tQywQ==}
first commit
2026-02-21 03:10:39 +00:00
zod-to-json-schema@3.25.1:
resolution: {integrity: sha512-pM/SU9d3YAggzi6MtR4h7ruuQlqKtad8e9S0fmxcMi+ueAK5Korys/aWcV9LIIHTVbj01NdzxcnXSN+O74ZIVA==}
peerDependencies:
zod: ^3.25 || ^4
zod@3.25.76:
resolution: {integrity: sha512-gzUt/qt81nXsFGKIFcC3YnfEAx5NkunCfnDlvuBSSFS02bcXu4Lmea0AFIUwbLWxWPx3d9p8S5QoaujKcNQxcQ==}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
zod@4.1.8:
resolution: {integrity: sha512-5R1P+WwQqmmMIEACyzSvo4JXHY5WiAFHRMg+zBZKgKS+Q1viRa0C1hmUKtHltoIFKtIdki3pRxkmpP74jnNYHQ==}
first commit
2026-02-21 03:10:39 +00:00
snapshots:
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
'@acemir/cssom@0.9.31': {}
'@adobe/css-tools@4.4.4': {}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
'@ai-sdk/provider@3.0.8':
dependencies:
json-schema: 0.4.0
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
'@alcalzone/ansi-tokenize@0.2.5':
dependencies:
ansi-styles: 6.2.3
is-fullwidth-code-point: 5.1.0
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
'@ampproject/remapping@2.3.0':
dependencies:
'@jridgewell/gen-mapping': 0.3.13
'@jridgewell/trace-mapping': 0.3.31
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
'@anthropic-ai/sdk@0.91.1(zod@3.25.76)':
dependencies:
json-schema-to-ts: 3.1.1
optionalDependencies:
zod: 3.25.76
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
'@asamuzakjp/css-color@5.1.11':
dependencies:
'@asamuzakjp/generational-cache': 1.0.1
'@csstools/css-calc': 3.2.0(@csstools/css-parser-algorithms@4.0.0(@csstools/css-tokenizer@4.0.0))(@csstools/css-tokenizer@4.0.0)
'@csstools/css-color-parser': 4.1.0(@csstools/css-parser-algorithms@4.0.0(@csstools/css-tokenizer@4.0.0))(@csstools/css-tokenizer@4.0.0)
'@csstools/css-parser-algorithms': 4.0.0(@csstools/css-tokenizer@4.0.0)
'@csstools/css-tokenizer': 4.0.0
'@asamuzakjp/dom-selector@6.8.1':
dependencies:
'@asamuzakjp/nwsapi': 2.3.9
bidi-js: 1.0.3
css-tree: 3.2.1
is-potential-custom-element-name: 1.0.1
lru-cache: 11.2.6
'@asamuzakjp/generational-cache@1.0.1': {}
'@asamuzakjp/nwsapi@2.3.9': {}
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
'@aws-crypto/sha256-browser@5.2.0':
dependencies:
'@aws-crypto/sha256-js': 5.2.0
'@aws-crypto/supports-web-crypto': 5.2.0
'@aws-crypto/util': 5.2.0
'@aws-sdk/types': 3.974.2
'@aws-sdk/util-locate-window': 3.965.8
'@smithy/util-utf8': 2.3.0
tslib: 2.8.1
'@aws-crypto/sha256-js@5.2.0':
dependencies:
'@aws-crypto/util': 5.2.0
'@aws-sdk/types': 3.974.2
tslib: 2.8.1
'@aws-crypto/supports-web-crypto@5.2.0':
dependencies:
tslib: 2.8.1
'@aws-crypto/util@5.2.0':
dependencies:
'@aws-sdk/types': 3.974.2
'@smithy/util-utf8': 2.3.0
tslib: 2.8.1
'@aws-sdk/client-bedrock-runtime@3.1048.0':
dependencies:
'@aws-crypto/sha256-browser': 5.2.0
'@aws-crypto/sha256-js': 5.2.0
'@aws-sdk/core': 3.977.6
'@aws-sdk/credential-provider-node': 3.972.78
'@aws-sdk/eventstream-handler-node': 3.972.31
'@aws-sdk/middleware-eventstream': 3.972.26
'@aws-sdk/middleware-websocket': 3.972.49
'@aws-sdk/token-providers': 3.1048.0
'@aws-sdk/types': 3.974.2
'@smithy/core': 3.31.1
'@smithy/fetch-http-handler': 5.6.13
'@smithy/node-http-handler': 4.9.13
'@smithy/types': 4.16.1
tslib: 2.8.1
'@aws-sdk/core@3.977.6':
dependencies:
'@aws-sdk/types': 3.974.2
'@aws-sdk/xml-builder': 3.972.37
'@aws/lambda-invoke-store': 0.3.0
'@smithy/core': 3.31.1
'@smithy/signature-v4': 5.6.12
'@smithy/types': 4.16.1
bowser: 2.14.1
tslib: 2.8.1
'@aws-sdk/credential-provider-env@3.972.67':
dependencies:
'@aws-sdk/core': 3.977.6
'@aws-sdk/types': 3.974.2
'@smithy/core': 3.31.1
'@smithy/types': 4.16.1
tslib: 2.8.1
'@aws-sdk/credential-provider-http@3.972.69':
dependencies:
'@aws-sdk/core': 3.977.6
'@aws-sdk/types': 3.974.2
'@smithy/core': 3.31.1
'@smithy/fetch-http-handler': 5.6.13
'@smithy/node-http-handler': 4.9.13
'@smithy/types': 4.16.1
tslib: 2.8.1
'@aws-sdk/credential-provider-ini@3.973.12':
dependencies:
'@aws-sdk/core': 3.977.6
'@aws-sdk/credential-provider-env': 3.972.67
'@aws-sdk/credential-provider-http': 3.972.69
'@aws-sdk/credential-provider-login': 3.972.74
'@aws-sdk/credential-provider-process': 3.972.67
'@aws-sdk/credential-provider-sso': 3.973.11
'@aws-sdk/credential-provider-web-identity': 3.972.73
'@aws-sdk/nested-clients': 3.997.41
'@aws-sdk/types': 3.974.2
'@smithy/core': 3.31.1
'@smithy/credential-provider-imds': 4.4.16
'@smithy/types': 4.16.1
tslib: 2.8.1
'@aws-sdk/credential-provider-login@3.972.74':
dependencies:
'@aws-sdk/core': 3.977.6
'@aws-sdk/nested-clients': 3.997.41
'@aws-sdk/types': 3.974.2
'@smithy/core': 3.31.1
'@smithy/types': 4.16.1
tslib: 2.8.1
'@aws-sdk/credential-provider-node@3.972.78':
dependencies:
'@aws-sdk/credential-provider-env': 3.972.67
'@aws-sdk/credential-provider-http': 3.972.69
'@aws-sdk/credential-provider-ini': 3.973.12
'@aws-sdk/credential-provider-process': 3.972.67
'@aws-sdk/credential-provider-sso': 3.973.11
'@aws-sdk/credential-provider-web-identity': 3.972.73
'@aws-sdk/types': 3.974.2
'@smithy/core': 3.31.1
'@smithy/credential-provider-imds': 4.4.16
'@smithy/types': 4.16.1
tslib: 2.8.1
'@aws-sdk/credential-provider-process@3.972.67':
dependencies:
'@aws-sdk/core': 3.977.6
'@aws-sdk/types': 3.974.2
'@smithy/core': 3.31.1
'@smithy/types': 4.16.1
tslib: 2.8.1
'@aws-sdk/credential-provider-sso@3.973.11':
dependencies:
'@aws-sdk/core': 3.977.6
'@aws-sdk/nested-clients': 3.997.41
'@aws-sdk/token-providers': 3.1103.0
'@aws-sdk/types': 3.974.2
'@smithy/core': 3.31.1
'@smithy/types': 4.16.1
tslib: 2.8.1
'@aws-sdk/credential-provider-web-identity@3.972.73':
dependencies:
'@aws-sdk/core': 3.977.6
'@aws-sdk/nested-clients': 3.997.41
'@aws-sdk/types': 3.974.2
'@smithy/core': 3.31.1
'@smithy/types': 4.16.1
tslib: 2.8.1
'@aws-sdk/eventstream-handler-node@3.972.31':
dependencies:
'@aws-sdk/types': 3.974.2
'@smithy/core': 3.31.1
'@smithy/types': 4.16.1
tslib: 2.8.1
'@aws-sdk/middleware-eventstream@3.972.26':
dependencies:
'@aws-sdk/types': 3.974.2
'@smithy/core': 3.31.1
'@smithy/types': 4.16.1
tslib: 2.8.1
'@aws-sdk/middleware-websocket@3.972.49':
dependencies:
'@aws-sdk/core': 3.977.6
'@aws-sdk/types': 3.974.2
'@smithy/core': 3.31.1
'@smithy/fetch-http-handler': 5.6.13
'@smithy/signature-v4': 5.6.12
'@smithy/types': 4.16.1
tslib: 2.8.1
'@aws-sdk/nested-clients@3.997.41':
dependencies:
'@aws-sdk/core': 3.977.6
'@aws-sdk/signature-v4-multi-region': 3.996.43
'@aws-sdk/types': 3.974.2
'@smithy/core': 3.31.1
'@smithy/fetch-http-handler': 5.6.13
'@smithy/node-http-handler': 4.9.13
'@smithy/types': 4.16.1
tslib: 2.8.1
'@aws-sdk/signature-v4-multi-region@3.996.43':
dependencies:
'@aws-sdk/types': 3.974.2
'@smithy/signature-v4': 5.6.12
'@smithy/types': 4.16.1
tslib: 2.8.1
'@aws-sdk/token-providers@3.1048.0':
dependencies:
'@aws-sdk/core': 3.977.6
'@aws-sdk/nested-clients': 3.997.41
'@aws-sdk/types': 3.974.2
'@smithy/core': 3.31.1
'@smithy/types': 4.16.1
tslib: 2.8.1
'@aws-sdk/token-providers@3.1103.0':
dependencies:
'@aws-sdk/core': 3.977.6
'@aws-sdk/nested-clients': 3.997.41
'@aws-sdk/types': 3.974.2
'@smithy/core': 3.31.1
'@smithy/types': 4.16.1
tslib: 2.8.1
'@aws-sdk/types@3.974.2':
dependencies:
'@smithy/types': 4.16.1
tslib: 2.8.1
'@aws-sdk/util-locate-window@3.965.8':
dependencies:
tslib: 2.8.1
'@aws-sdk/xml-builder@3.972.37':
dependencies:
'@smithy/types': 4.16.1
tslib: 2.8.1
'@aws/lambda-invoke-store@0.3.0': {}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
'@babel/code-frame@7.29.0':
dependencies:
'@babel/helper-validator-identifier': 7.28.5
js-tokens: 4.0.0
picocolors: 1.1.1
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
'@babel/code-frame@7.29.7':
dependencies:
'@babel/helper-validator-identifier': 7.29.7
js-tokens: 4.0.0
picocolors: 1.1.1
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
'@babel/compat-data@7.29.0': {}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
'@babel/core@7.28.0':
dependencies:
'@ampproject/remapping': 2.3.0
'@babel/code-frame': 7.29.7
'@babel/generator': 7.29.8
'@babel/helper-compilation-targets': 7.28.6
'@babel/helper-module-transforms': 7.29.7(@babel/core@7.28.0)
'@babel/helpers': 7.29.2
'@babel/parser': 7.29.8
'@babel/template': 7.29.7
'@babel/traverse': 7.29.8
'@babel/types': 7.29.8
convert-source-map: 2.0.0
debug: 4.4.3
gensync: 1.0.0-beta.2
json5: 2.2.3
semver: 6.3.1
transitivePeerDependencies:
- supports-color
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
'@babel/core@7.29.0':
dependencies:
'@babel/code-frame': 7.29.0
'@babel/generator': 7.29.1
'@babel/helper-compilation-targets': 7.28.6
'@babel/helper-module-transforms': 7.28.6(@babel/core@7.29.0)
'@babel/helpers': 7.29.2
'@babel/parser': 7.29.0
'@babel/template': 7.28.6
'@babel/traverse': 7.29.0
'@babel/types': 7.29.0
'@jridgewell/remapping': 2.3.5
convert-source-map: 2.0.0
debug: 4.4.3
gensync: 1.0.0-beta.2
json5: 2.2.3
semver: 6.3.1
transitivePeerDependencies:
- supports-color
'@babel/generator@7.29.1':
dependencies:
'@babel/parser': 7.29.0
'@babel/types': 7.29.0
'@jridgewell/gen-mapping': 0.3.13
'@jridgewell/trace-mapping': 0.3.31
jsesc: 3.1.0
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
'@babel/generator@7.29.8':
dependencies:
'@babel/parser': 7.29.8
'@babel/types': 7.29.8
'@jridgewell/gen-mapping': 0.3.13
'@jridgewell/trace-mapping': 0.3.31
jsesc: 3.1.0
'@babel/helper-annotate-as-pure@7.29.7':
dependencies:
'@babel/types': 7.29.8
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
'@babel/helper-compilation-targets@7.28.6':
dependencies:
'@babel/compat-data': 7.29.0
'@babel/helper-validator-option': 7.27.1
browserslist: 4.28.2
lru-cache: 5.1.1
semver: 6.3.1
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
'@babel/helper-create-class-features-plugin@7.29.7(@babel/core@7.28.0)':
dependencies:
'@babel/core': 7.28.0
'@babel/helper-annotate-as-pure': 7.29.7
'@babel/helper-member-expression-to-functions': 7.29.7
'@babel/helper-optimise-call-expression': 7.29.7
'@babel/helper-replace-supers': 7.29.7(@babel/core@7.28.0)
'@babel/helper-skip-transparent-expression-wrappers': 7.29.7
'@babel/traverse': 7.29.8
semver: 6.3.1
transitivePeerDependencies:
- supports-color
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
'@babel/helper-globals@7.28.0': {}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
'@babel/helper-globals@7.29.7': {}
'@babel/helper-member-expression-to-functions@7.29.7':
dependencies:
'@babel/traverse': 7.29.8
'@babel/types': 7.29.8
transitivePeerDependencies:
- supports-color
'@babel/helper-module-imports@7.18.6':
dependencies:
'@babel/types': 7.29.8
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
'@babel/helper-module-imports@7.28.6':
dependencies:
'@babel/traverse': 7.29.0
'@babel/types': 7.29.0
transitivePeerDependencies:
- supports-color
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
'@babel/helper-module-imports@7.29.7':
dependencies:
'@babel/traverse': 7.29.8
'@babel/types': 7.29.8
transitivePeerDependencies:
- supports-color
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
'@babel/helper-module-transforms@7.28.6(@babel/core@7.29.0)':
dependencies:
'@babel/core': 7.29.0
'@babel/helper-module-imports': 7.28.6
'@babel/helper-validator-identifier': 7.28.5
'@babel/traverse': 7.29.0
transitivePeerDependencies:
- supports-color
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
'@babel/helper-module-transforms@7.29.7(@babel/core@7.28.0)':
dependencies:
'@babel/core': 7.28.0
'@babel/helper-module-imports': 7.29.7
'@babel/helper-validator-identifier': 7.29.7
'@babel/traverse': 7.29.8
transitivePeerDependencies:
- supports-color
'@babel/helper-optimise-call-expression@7.29.7':
dependencies:
'@babel/types': 7.29.8
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
'@babel/helper-plugin-utils@7.28.6': {}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
'@babel/helper-plugin-utils@7.29.7': {}
'@babel/helper-replace-supers@7.29.7(@babel/core@7.28.0)':
dependencies:
'@babel/core': 7.28.0
'@babel/helper-member-expression-to-functions': 7.29.7
'@babel/helper-optimise-call-expression': 7.29.7
'@babel/traverse': 7.29.8
transitivePeerDependencies:
- supports-color
'@babel/helper-skip-transparent-expression-wrappers@7.29.7':
dependencies:
'@babel/traverse': 7.29.8
'@babel/types': 7.29.8
transitivePeerDependencies:
- supports-color
first commit
2026-02-21 03:10:39 +00:00
'@babel/helper-string-parser@7.27.1': {}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
'@babel/helper-string-parser@7.29.7': {}
first commit
2026-02-21 03:10:39 +00:00
'@babel/helper-validator-identifier@7.28.5': {}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
'@babel/helper-validator-identifier@7.29.7': {}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
'@babel/helper-validator-option@7.27.1': {}
'@babel/helpers@7.29.2':
dependencies:
'@babel/template': 7.28.6
'@babel/types': 7.29.0
first commit
2026-02-21 03:10:39 +00:00
'@babel/parser@7.29.0':
dependencies:
'@babel/types': 7.29.0
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
'@babel/parser@7.29.8':
dependencies:
'@babel/types': 7.29.8
'@babel/plugin-syntax-jsx@7.29.7(@babel/core@7.28.0)':
dependencies:
'@babel/core': 7.28.0
'@babel/helper-plugin-utils': 7.29.7
'@babel/plugin-syntax-typescript@7.29.7(@babel/core@7.28.0)':
dependencies:
'@babel/core': 7.28.0
'@babel/helper-plugin-utils': 7.29.7
'@babel/plugin-transform-modules-commonjs@7.29.7(@babel/core@7.28.0)':
dependencies:
'@babel/core': 7.28.0
'@babel/helper-module-transforms': 7.29.7(@babel/core@7.28.0)
'@babel/helper-plugin-utils': 7.29.7
transitivePeerDependencies:
- supports-color
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
'@babel/plugin-transform-react-jsx-self@7.27.1(@babel/core@7.29.0)':
dependencies:
'@babel/core': 7.29.0
'@babel/helper-plugin-utils': 7.28.6
'@babel/plugin-transform-react-jsx-source@7.27.1(@babel/core@7.29.0)':
dependencies:
'@babel/core': 7.29.0
'@babel/helper-plugin-utils': 7.28.6
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
'@babel/plugin-transform-typescript@7.29.7(@babel/core@7.28.0)':
dependencies:
'@babel/core': 7.28.0
'@babel/helper-annotate-as-pure': 7.29.7
'@babel/helper-create-class-features-plugin': 7.29.7(@babel/core@7.28.0)
'@babel/helper-plugin-utils': 7.29.7
'@babel/helper-skip-transparent-expression-wrappers': 7.29.7
'@babel/plugin-syntax-typescript': 7.29.7(@babel/core@7.28.0)
transitivePeerDependencies:
- supports-color
'@babel/preset-typescript@7.27.1(@babel/core@7.28.0)':
dependencies:
'@babel/core': 7.28.0
'@babel/helper-plugin-utils': 7.29.7
'@babel/helper-validator-option': 7.27.1
'@babel/plugin-syntax-jsx': 7.29.7(@babel/core@7.28.0)
'@babel/plugin-transform-modules-commonjs': 7.29.7(@babel/core@7.28.0)
'@babel/plugin-transform-typescript': 7.29.7(@babel/core@7.28.0)
transitivePeerDependencies:
- supports-color
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
'@babel/runtime@7.29.2': {}
'@babel/template@7.28.6':
dependencies:
'@babel/code-frame': 7.29.0
'@babel/parser': 7.29.0
'@babel/types': 7.29.0
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
'@babel/template@7.29.7':
dependencies:
'@babel/code-frame': 7.29.7
'@babel/parser': 7.29.8
'@babel/types': 7.29.8
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
'@babel/traverse@7.29.0':
dependencies:
'@babel/code-frame': 7.29.0
'@babel/generator': 7.29.1
'@babel/helper-globals': 7.28.0
'@babel/parser': 7.29.0
'@babel/template': 7.28.6
'@babel/types': 7.29.0
debug: 4.4.3
transitivePeerDependencies:
- supports-color
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
'@babel/traverse@7.29.8':
dependencies:
'@babel/code-frame': 7.29.7
'@babel/generator': 7.29.8
'@babel/helper-globals': 7.29.7
'@babel/parser': 7.29.8
'@babel/template': 7.29.7
'@babel/types': 7.29.8
debug: 4.4.3
transitivePeerDependencies:
- supports-color
first commit
2026-02-21 03:10:39 +00:00
'@babel/types@7.29.0':
dependencies:
'@babel/helper-string-parser': 7.27.1
'@babel/helper-validator-identifier': 7.28.5
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
'@babel/types@7.29.8':
dependencies:
'@babel/helper-string-parser': 7.29.7
'@babel/helper-validator-identifier': 7.29.7
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
'@balena/dockerignore@1.0.2': {}
first commit
2026-02-21 03:10:39 +00:00
'@bcoe/v8-coverage@1.0.2': {}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
'@bramus/specificity@2.4.2':
dependencies:
css-tree: 3.2.1
'@csstools/color-helpers@6.0.2': {}
'@csstools/css-calc@3.2.0(@csstools/css-parser-algorithms@4.0.0(@csstools/css-tokenizer@4.0.0))(@csstools/css-tokenizer@4.0.0)':
dependencies:
'@csstools/css-parser-algorithms': 4.0.0(@csstools/css-tokenizer@4.0.0)
'@csstools/css-tokenizer': 4.0.0
'@csstools/css-color-parser@4.1.0(@csstools/css-parser-algorithms@4.0.0(@csstools/css-tokenizer@4.0.0))(@csstools/css-tokenizer@4.0.0)':
dependencies:
'@csstools/color-helpers': 6.0.2
'@csstools/css-calc': 3.2.0(@csstools/css-parser-algorithms@4.0.0(@csstools/css-tokenizer@4.0.0))(@csstools/css-tokenizer@4.0.0)
'@csstools/css-parser-algorithms': 4.0.0(@csstools/css-tokenizer@4.0.0)
'@csstools/css-tokenizer': 4.0.0
'@csstools/css-parser-algorithms@4.0.0(@csstools/css-tokenizer@4.0.0)':
dependencies:
'@csstools/css-tokenizer': 4.0.0
'@csstools/css-syntax-patches-for-csstree@1.1.3(css-tree@3.2.1)':
optionalDependencies:
css-tree: 3.2.1
'@csstools/css-tokenizer@4.0.0': {}
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
'@earendil-works/pi-agent-core@0.84.1(@modelcontextprotocol/sdk@1.26.0(zod@3.25.76))(ws@8.19.0)(zod@3.25.76)':
dependencies:
'@earendil-works/pi-ai': 0.84.1(@modelcontextprotocol/sdk@1.26.0(zod@3.25.76))(ws@8.19.0)(zod@3.25.76)
'@earendil-works/pi-telemetry': 0.84.1
diff: 8.0.4
ignore: 7.0.5
typebox: 1.3.7
yaml: 2.9.0
transitivePeerDependencies:
- '@modelcontextprotocol/sdk'
- bufferutil
- supports-color
- utf-8-validate
- ws
- zod
'@earendil-works/pi-ai@0.84.1(@modelcontextprotocol/sdk@1.26.0(zod@3.25.76))(ws@8.19.0)(zod@3.25.76)':
dependencies:
'@anthropic-ai/sdk': 0.91.1(zod@3.25.76)
'@aws-sdk/client-bedrock-runtime': 3.1048.0
'@earendil-works/pi-telemetry': 0.84.1
'@google/genai': 1.52.0(@modelcontextprotocol/sdk@1.26.0(zod@3.25.76))
'@mistralai/mistralai': 2.2.6(@opentelemetry/api@1.9.0)
'@opentelemetry/api': 1.9.0
'@smithy/node-http-handler': 4.7.3
http-proxy-agent: 7.0.2
https-proxy-agent: 7.0.6
openai: 6.26.0(ws@8.19.0)(zod@3.25.76)
partial-json: 0.1.7
typebox: 1.3.7
transitivePeerDependencies:
- '@modelcontextprotocol/sdk'
- bufferutil
- supports-color
- utf-8-validate
- ws
- zod
'@earendil-works/pi-client@0.84.1':
dependencies:
'@earendil-works/pi-protocol': 0.84.1
'@earendil-works/pi-coding-agent@0.84.1(@modelcontextprotocol/sdk@1.26.0(zod@3.25.76))(ws@8.19.0)(zod@3.25.76)':
dependencies:
'@earendil-works/pi-agent-core': 0.84.1(@modelcontextprotocol/sdk@1.26.0(zod@3.25.76))(ws@8.19.0)(zod@3.25.76)
'@earendil-works/pi-ai': 0.84.1(@modelcontextprotocol/sdk@1.26.0(zod@3.25.76))(ws@8.19.0)(zod@3.25.76)
'@earendil-works/pi-client': 0.84.1
'@earendil-works/pi-protocol': 0.84.1
'@earendil-works/pi-tui': 0.84.1
'@silvia-odwyer/photon-node': 0.3.4
chalk: 5.6.2
cross-spawn: 7.0.6
diff: 8.0.4
glob: 13.0.6
grok-mermaid: 0.2.2
highlight.js: 10.7.3
hosted-git-info: 9.0.3
ignore: 7.0.5
jiti: 2.7.0
minimatch: 10.2.5
proper-lockfile: 4.1.2
semver: 7.8.0
typebox: 1.3.7
undici: 8.9.0
yaml: 2.9.0
optionalDependencies:
'@mariozechner/clipboard': 0.3.9
transitivePeerDependencies:
- '@modelcontextprotocol/sdk'
- bufferutil
- supports-color
- utf-8-validate
- ws
- zod
'@earendil-works/pi-protocol@0.84.1':
dependencies:
typebox: 1.3.7
'@earendil-works/pi-telemetry@0.84.1': {}
'@earendil-works/pi-tui@0.84.1':
dependencies:
get-east-asian-width: 1.6.0
marked: 18.0.5
feat(web): bold redesign — Tailwind v4 + shadcn-style primitives + Skills/Proposals/Revisions UI Phase 6 of the Skills + Revisions + Proposals work. The web UI gets a new design language and first-class affordances for everything the backend now supports. ## Visual direction - Tailwind v4 with custom @theme block (oklch tokens). Dark-mode-only (internal tool — light mode doubles QA surface). - Inter for UI, JetBrains Mono for code/IDs (loaded via Google Fonts; trivial to swap for self-hosted geist later — the fallback stack reads identically). - Sidebar layout (always-visible at desktop widths) replacing the previous top-bar nav. Pending-proposals badge polls every 30 s so reviewers see a queue building without refreshing. - Lucide icons throughout. - Spacing and radii on Tailwind defaults. Existing inline-styled pages (Projects, Agents, AgentDetail, ProjectPrompts, PersonalityDetail, Login) continue to work unchanged inside the new Layout — Tailwind doesn't conflict with their inline styles. A follow-up can migrate them incrementally. ## What's added ### Build infra (src/web/) - package.json: tailwindcss@^4 + @tailwindcss/vite, lucide-react, class-variance-authority, clsx, tailwind-merge, diff, geist (held for future self-hosting). - vite.config.ts: registers the @tailwindcss/vite plugin. - src/index.css: Tailwind import + @theme tokens + @layer base. - src/main.tsx: imports index.css. - src/lib/utils.ts: shadcn-style cn() helper. ### shadcn-style primitives (src/components/ui/) Hand-written rather than generated via `npx shadcn` so the repo doesn't depend on a CLI tool that needs interactive runtime: - button.tsx — variants: primary / secondary / ghost / danger / link; sizes: sm / md / lg / icon. - card.tsx — Card + Header/Title/Description/Content/Footer subparts. - badge.tsx — variants: default / info / success / warning / danger / outline. - input.tsx — Input + Textarea + Label. - tabs.tsx — no-dep accessible Tabs (no Radix needed for our use). - separator.tsx — h/v separator with role=separator. ### Diff component (src/components/Diff.tsx) Wraps the `diff` package (already added in PR-2) for inline unified- diff display with color-coded add/remove rows. Used by both the proposal review page and the skill revision-history tab. ### New pages (src/pages/) - Dashboard.tsx — at-a-glance home. Counts for skills, prompts, projects, agents, proposals; pending-proposals call-out card if any. - Skills.tsx — list view, separated into Global vs Project/Agent- scoped sections. - SkillDetail.tsx — name + semver + description; tabs for SKILL.md / Files / Metadata / History. History tab shows revisions with click-to-diff against the live body. - Proposals.tsx — queue with Pending/Approved/Rejected tabs. Pending count is highlighted in amber. - ProposalDetail.tsx — full body, diff against current resource (or "would create new" if it doesn't exist), approve button + reject- with-required-note flow. ### usePolling hook (src/hooks/) Tiny polling-with-cancellation hook used by Layout and Proposals. ### Layout rewrite (src/components/Layout.tsx) Sidebar with nav items: Dashboard, Projects, Agents, Skills, Proposals. Lucide icons. Active-route highlighting via NavLink. Pending-proposals warning badge on the Proposals item. ### Routes (src/App.tsx) New routes: /dashboard, /skills, /skills/:name, /proposals, /proposals/:id. Default redirects to /dashboard. ### API types (src/api.ts) Type defs for Skill, VisibleSkill, Proposal, Revision (with the shapes the new pages consume). ## Tests Existing 7 web tests still pass (Login + api). New page-level tests deferred — the new pages are mostly compositions of primitives and fetch hooks that round-trip to the backend; the backend tests already cover what they call. PR-7 polish can add render-and-click tests if coverage drift surfaces. ## Verification - `pnpm --filter @mcpctl/web build` clean, no warnings. - `pnpm test:run` whole monorepo: 162 test files / 2157 tests green. - Visual smoke deferred — needs a running mcpd to populate the fixtures. Manual smoke tested locally is the next step. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-07 17:54:55 +01:00
'@emnapi/runtime@1.10.0':
dependencies:
tslib: 2.8.1
optional: true
first commit
2026-02-21 03:10:39 +00:00
'@esbuild/aix-ppc64@0.27.3':
optional: true
'@esbuild/android-arm64@0.27.3':
optional: true
'@esbuild/android-arm@0.27.3':
optional: true
'@esbuild/android-x64@0.27.3':
optional: true
'@esbuild/darwin-arm64@0.27.3':
optional: true
'@esbuild/darwin-x64@0.27.3':
optional: true
'@esbuild/freebsd-arm64@0.27.3':
optional: true
'@esbuild/freebsd-x64@0.27.3':
optional: true
'@esbuild/linux-arm64@0.27.3':
optional: true
'@esbuild/linux-arm@0.27.3':
optional: true
'@esbuild/linux-ia32@0.27.3':
optional: true
'@esbuild/linux-loong64@0.27.3':
optional: true
'@esbuild/linux-mips64el@0.27.3':
optional: true
'@esbuild/linux-ppc64@0.27.3':
optional: true
'@esbuild/linux-riscv64@0.27.3':
optional: true
'@esbuild/linux-s390x@0.27.3':
optional: true
'@esbuild/linux-x64@0.27.3':
optional: true
'@esbuild/netbsd-arm64@0.27.3':
optional: true
'@esbuild/netbsd-x64@0.27.3':
optional: true
'@esbuild/openbsd-arm64@0.27.3':
optional: true
'@esbuild/openbsd-x64@0.27.3':
optional: true
'@esbuild/openharmony-arm64@0.27.3':
optional: true
'@esbuild/sunos-x64@0.27.3':
optional: true
'@esbuild/win32-arm64@0.27.3':
optional: true
'@esbuild/win32-ia32@0.27.3':
optional: true
'@esbuild/win32-x64@0.27.3':
optional: true
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
'@eslint-community/eslint-utils@4.9.1(eslint@10.0.1(jiti@2.7.0))':
first commit
2026-02-21 03:10:39 +00:00
dependencies:
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
eslint: 10.0.1(jiti@2.7.0)
first commit
2026-02-21 03:10:39 +00:00
eslint-visitor-keys: 3.4.3
'@eslint-community/regexpp@4.12.2': {}
'@eslint/config-array@0.23.2':
dependencies:
'@eslint/object-schema': 3.0.2
debug: 4.4.3
minimatch: 10.2.2
transitivePeerDependencies:
- supports-color
'@eslint/config-helpers@0.5.2':
dependencies:
'@eslint/core': 1.1.0
'@eslint/core@1.1.0':
dependencies:
'@types/json-schema': 7.0.15
'@eslint/object-schema@3.0.2': {}
'@eslint/plugin-kit@0.6.0':
dependencies:
'@eslint/core': 1.1.0
levn: 0.4.1
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
'@exodus/bytes@1.15.0': {}
feat(mcpd+deploy): serve web UI at /ui + smoke tests + docs (Stage 6) The closing stage. mcpd now hosts the Stage 5 SPA, the Docker image bundles the build artifact, a smoke test exercises the personality HTTP surface end-to-end, and the user-facing docs spell out the mental model. mcpd: - Add @fastify/static dep. - New routes/web-ui.ts: registers /ui/* against a static bundle. Looks for the bundle at $MCPD_WEB_ROOT, then /usr/share/mcpd/web (the Docker image path), then a dev-tree fallback. Logs and skips cleanly if missing — API-only deploys keep working. - SPA fallback: any /ui/<path> that doesn't match a file falls through to index.html so direct hits to react-router URLs work. - /ui/* falls through to `kind: skip` in mapUrlToPermission, so the static assets are served unauthenticated. Each API call from the SPA still carries the bearer token. Deploy: - Dockerfile.mcpd builds the @mcpctl/web bundle in the same builder stage and copies dist/ to /usr/share/mcpd/web in the runtime image. Smoke (personality.smoke.test.ts): - Live mcpd flow: create secret/llm/agent/personality, attach an agent-direct prompt, verify the binding listing, reject double- attach (409) + foreign-agent prompt (400), set defaultPersonality by name, detach + delete cleanup. Docs: - New docs/personalities.md: VLAN-on-ethernet model, system-block ordering table, three prompt scopes, CLI walkthrough, web UI walkthrough, full API surface, RBAC notes. - agents.md and chat.md cross-link. - README's Agents section gains a Personalities subsection. Test count after Stage 6: mcpd: 801/801 cli: 430/430 web: 7/7 db: 58/62 (4 pre-existing) Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:48:43 +01:00
'@fastify/accept-negotiator@2.0.1': {}
first commit
2026-02-21 03:10:39 +00:00
'@fastify/ajv-compiler@4.0.5':
dependencies:
ajv: 8.18.0
ajv-formats: 3.0.1(ajv@8.18.0)
fast-uri: 3.1.0
'@fastify/cors@10.1.0':
dependencies:
fastify-plugin: 5.1.0
mnemonist: 0.40.0
'@fastify/error@4.2.0': {}
'@fastify/fast-json-stringify-compiler@5.0.3':
dependencies:
fast-json-stringify: 6.3.0
'@fastify/forwarded@3.0.1': {}
'@fastify/helmet@12.0.1':
dependencies:
fastify-plugin: 5.1.0
helmet: 7.2.0
'@fastify/merge-json-schemas@0.2.1':
dependencies:
dequal: 2.0.3
'@fastify/proxy-addr@5.1.0':
dependencies:
'@fastify/forwarded': 3.0.1
ipaddr.js: 2.3.0
'@fastify/rate-limit@10.3.0':
dependencies:
'@lukeed/ms': 2.0.2
fastify-plugin: 5.1.0
toad-cache: 3.7.0
feat(mcpd+deploy): serve web UI at /ui + smoke tests + docs (Stage 6) The closing stage. mcpd now hosts the Stage 5 SPA, the Docker image bundles the build artifact, a smoke test exercises the personality HTTP surface end-to-end, and the user-facing docs spell out the mental model. mcpd: - Add @fastify/static dep. - New routes/web-ui.ts: registers /ui/* against a static bundle. Looks for the bundle at $MCPD_WEB_ROOT, then /usr/share/mcpd/web (the Docker image path), then a dev-tree fallback. Logs and skips cleanly if missing — API-only deploys keep working. - SPA fallback: any /ui/<path> that doesn't match a file falls through to index.html so direct hits to react-router URLs work. - /ui/* falls through to `kind: skip` in mapUrlToPermission, so the static assets are served unauthenticated. Each API call from the SPA still carries the bearer token. Deploy: - Dockerfile.mcpd builds the @mcpctl/web bundle in the same builder stage and copies dist/ to /usr/share/mcpd/web in the runtime image. Smoke (personality.smoke.test.ts): - Live mcpd flow: create secret/llm/agent/personality, attach an agent-direct prompt, verify the binding listing, reject double- attach (409) + foreign-agent prompt (400), set defaultPersonality by name, detach + delete cleanup. Docs: - New docs/personalities.md: VLAN-on-ethernet model, system-block ordering table, three prompt scopes, CLI walkthrough, web UI walkthrough, full API surface, RBAC notes. - agents.md and chat.md cross-link. - README's Agents section gains a Personalities subsection. Test count after Stage 6: mcpd: 801/801 cli: 430/430 web: 7/7 db: 58/62 (4 pre-existing) Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:48:43 +01:00
'@fastify/send@4.1.0':
dependencies:
'@lukeed/ms': 2.0.2
escape-html: 1.0.3
fast-decode-uri-component: 1.0.1
http-errors: 2.0.1
mime: 3.0.0
'@fastify/static@8.3.0':
dependencies:
'@fastify/accept-negotiator': 2.0.1
'@fastify/send': 4.1.0
content-disposition: 0.5.4
fastify-plugin: 5.1.0
fastq: 1.20.1
glob: 11.1.0
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
'@gar/promise-retry@1.0.3': {}
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
'@google/genai@1.52.0(@modelcontextprotocol/sdk@1.26.0(zod@3.25.76))':
dependencies:
google-auth-library: 10.9.1
p-retry: 4.6.2
protobufjs: 7.5.4
ws: 8.19.0
optionalDependencies:
'@modelcontextprotocol/sdk': 1.26.0(zod@3.25.76)
transitivePeerDependencies:
- bufferutil
- supports-color
- utf-8-validate
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
'@grpc/grpc-js@1.14.3':
dependencies:
'@grpc/proto-loader': 0.8.0
'@js-sdsl/ordered-map': 4.4.2
'@grpc/proto-loader@0.7.15':
dependencies:
lodash.camelcase: 4.3.0
long: 5.3.2
protobufjs: 7.5.4
yargs: 17.7.2
'@grpc/proto-loader@0.8.0':
dependencies:
lodash.camelcase: 4.3.0
long: 5.3.2
protobufjs: 7.5.4
yargs: 17.7.2
first commit
2026-02-21 03:10:39 +00:00
'@hono/node-server@1.19.9(hono@4.12.0)':
dependencies:
hono: 4.12.0
'@humanfs/core@0.19.1': {}
'@humanfs/node@0.16.7':
dependencies:
'@humanfs/core': 0.19.1
'@humanwhocodes/retry': 0.4.3
'@humanwhocodes/module-importer@1.0.1': {}
'@humanwhocodes/retry@0.4.3': {}
feat(web): bold redesign — Tailwind v4 + shadcn-style primitives + Skills/Proposals/Revisions UI Phase 6 of the Skills + Revisions + Proposals work. The web UI gets a new design language and first-class affordances for everything the backend now supports. ## Visual direction - Tailwind v4 with custom @theme block (oklch tokens). Dark-mode-only (internal tool — light mode doubles QA surface). - Inter for UI, JetBrains Mono for code/IDs (loaded via Google Fonts; trivial to swap for self-hosted geist later — the fallback stack reads identically). - Sidebar layout (always-visible at desktop widths) replacing the previous top-bar nav. Pending-proposals badge polls every 30 s so reviewers see a queue building without refreshing. - Lucide icons throughout. - Spacing and radii on Tailwind defaults. Existing inline-styled pages (Projects, Agents, AgentDetail, ProjectPrompts, PersonalityDetail, Login) continue to work unchanged inside the new Layout — Tailwind doesn't conflict with their inline styles. A follow-up can migrate them incrementally. ## What's added ### Build infra (src/web/) - package.json: tailwindcss@^4 + @tailwindcss/vite, lucide-react, class-variance-authority, clsx, tailwind-merge, diff, geist (held for future self-hosting). - vite.config.ts: registers the @tailwindcss/vite plugin. - src/index.css: Tailwind import + @theme tokens + @layer base. - src/main.tsx: imports index.css. - src/lib/utils.ts: shadcn-style cn() helper. ### shadcn-style primitives (src/components/ui/) Hand-written rather than generated via `npx shadcn` so the repo doesn't depend on a CLI tool that needs interactive runtime: - button.tsx — variants: primary / secondary / ghost / danger / link; sizes: sm / md / lg / icon. - card.tsx — Card + Header/Title/Description/Content/Footer subparts. - badge.tsx — variants: default / info / success / warning / danger / outline. - input.tsx — Input + Textarea + Label. - tabs.tsx — no-dep accessible Tabs (no Radix needed for our use). - separator.tsx — h/v separator with role=separator. ### Diff component (src/components/Diff.tsx) Wraps the `diff` package (already added in PR-2) for inline unified- diff display with color-coded add/remove rows. Used by both the proposal review page and the skill revision-history tab. ### New pages (src/pages/) - Dashboard.tsx — at-a-glance home. Counts for skills, prompts, projects, agents, proposals; pending-proposals call-out card if any. - Skills.tsx — list view, separated into Global vs Project/Agent- scoped sections. - SkillDetail.tsx — name + semver + description; tabs for SKILL.md / Files / Metadata / History. History tab shows revisions with click-to-diff against the live body. - Proposals.tsx — queue with Pending/Approved/Rejected tabs. Pending count is highlighted in amber. - ProposalDetail.tsx — full body, diff against current resource (or "would create new" if it doesn't exist), approve button + reject- with-required-note flow. ### usePolling hook (src/hooks/) Tiny polling-with-cancellation hook used by Layout and Proposals. ### Layout rewrite (src/components/Layout.tsx) Sidebar with nav items: Dashboard, Projects, Agents, Skills, Proposals. Lucide icons. Active-route highlighting via NavLink. Pending-proposals warning badge on the Proposals item. ### Routes (src/App.tsx) New routes: /dashboard, /skills, /skills/:name, /proposals, /proposals/:id. Default redirects to /dashboard. ### API types (src/api.ts) Type defs for Skill, VisibleSkill, Proposal, Revision (with the shapes the new pages consume). ## Tests Existing 7 web tests still pass (Login + api). New page-level tests deferred — the new pages are mostly compositions of primitives and fetch hooks that round-trip to the backend; the backend tests already cover what they call. PR-7 polish can add render-and-click tests if coverage drift surfaces. ## Verification - `pnpm --filter @mcpctl/web build` clean, no warnings. - `pnpm test:run` whole monorepo: 162 test files / 2157 tests green. - Visual smoke deferred — needs a running mcpd to populate the fixtures. Manual smoke tested locally is the next step. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-07 17:54:55 +01:00
'@img/colour@1.1.0':
optional: true
'@img/sharp-darwin-arm64@0.34.5':
optionalDependencies:
'@img/sharp-libvips-darwin-arm64': 1.2.4
optional: true
'@img/sharp-darwin-x64@0.34.5':
optionalDependencies:
'@img/sharp-libvips-darwin-x64': 1.2.4
optional: true
'@img/sharp-libvips-darwin-arm64@1.2.4':
optional: true
'@img/sharp-libvips-darwin-x64@1.2.4':
optional: true
'@img/sharp-libvips-linux-arm64@1.2.4':
optional: true
'@img/sharp-libvips-linux-arm@1.2.4':
optional: true
'@img/sharp-libvips-linux-ppc64@1.2.4':
optional: true
'@img/sharp-libvips-linux-riscv64@1.2.4':
optional: true
'@img/sharp-libvips-linux-s390x@1.2.4':
optional: true
'@img/sharp-libvips-linux-x64@1.2.4':
optional: true
'@img/sharp-libvips-linuxmusl-arm64@1.2.4':
optional: true
'@img/sharp-libvips-linuxmusl-x64@1.2.4':
optional: true
'@img/sharp-linux-arm64@0.34.5':
optionalDependencies:
'@img/sharp-libvips-linux-arm64': 1.2.4
optional: true
'@img/sharp-linux-arm@0.34.5':
optionalDependencies:
'@img/sharp-libvips-linux-arm': 1.2.4
optional: true
'@img/sharp-linux-ppc64@0.34.5':
optionalDependencies:
'@img/sharp-libvips-linux-ppc64': 1.2.4
optional: true
'@img/sharp-linux-riscv64@0.34.5':
optionalDependencies:
'@img/sharp-libvips-linux-riscv64': 1.2.4
optional: true
'@img/sharp-linux-s390x@0.34.5':
optionalDependencies:
'@img/sharp-libvips-linux-s390x': 1.2.4
optional: true
'@img/sharp-linux-x64@0.34.5':
optionalDependencies:
'@img/sharp-libvips-linux-x64': 1.2.4
optional: true
'@img/sharp-linuxmusl-arm64@0.34.5':
optionalDependencies:
'@img/sharp-libvips-linuxmusl-arm64': 1.2.4
optional: true
'@img/sharp-linuxmusl-x64@0.34.5':
optionalDependencies:
'@img/sharp-libvips-linuxmusl-x64': 1.2.4
optional: true
'@img/sharp-wasm32@0.34.5':
dependencies:
'@emnapi/runtime': 1.10.0
optional: true
'@img/sharp-win32-arm64@0.34.5':
optional: true
'@img/sharp-win32-ia32@0.34.5':
optional: true
'@img/sharp-win32-x64@0.34.5':
optional: true
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
'@inkjs/ui@2.0.0(ink@6.8.0(@types/react@19.2.14)(react@19.2.4))':
dependencies:
chalk: 5.6.2
cli-spinners: 3.4.0
deepmerge: 4.3.1
figures: 6.1.0
ink: 6.8.0(@types/react@19.2.14)(react@19.2.4)
first commit
2026-02-21 03:10:39 +00:00
'@inquirer/ansi@1.0.2': {}
feat: implement MCP registry client with multi-source search Add registry client that queries Official, Glama, and Smithery MCP registries with caching, request deduplication, retry logic, and result ranking/dedup. Includes 53 tests covering all components. Also fix null priority values in cancelled tasks (19-21) that broke Task Master, and add new tasks 25-27 for registry completion and CLI discover/install commands. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 03:46:14 +00:00
'@inquirer/checkbox@4.3.2(@types/node@25.3.0)':
first commit
2026-02-21 03:10:39 +00:00
dependencies:
'@inquirer/ansi': 1.0.2
feat: implement MCP registry client with multi-source search Add registry client that queries Official, Glama, and Smithery MCP registries with caching, request deduplication, retry logic, and result ranking/dedup. Includes 53 tests covering all components. Also fix null priority values in cancelled tasks (19-21) that broke Task Master, and add new tasks 25-27 for registry completion and CLI discover/install commands. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 03:46:14 +00:00
'@inquirer/core': 10.3.2(@types/node@25.3.0)
first commit
2026-02-21 03:10:39 +00:00
'@inquirer/figures': 1.0.15
feat: implement MCP registry client with multi-source search Add registry client that queries Official, Glama, and Smithery MCP registries with caching, request deduplication, retry logic, and result ranking/dedup. Includes 53 tests covering all components. Also fix null priority values in cancelled tasks (19-21) that broke Task Master, and add new tasks 25-27 for registry completion and CLI discover/install commands. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 03:46:14 +00:00
'@inquirer/type': 3.0.10(@types/node@25.3.0)
first commit
2026-02-21 03:10:39 +00:00
yoctocolors-cjs: 2.1.3
feat: implement MCP registry client with multi-source search Add registry client that queries Official, Glama, and Smithery MCP registries with caching, request deduplication, retry logic, and result ranking/dedup. Includes 53 tests covering all components. Also fix null priority values in cancelled tasks (19-21) that broke Task Master, and add new tasks 25-27 for registry completion and CLI discover/install commands. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 03:46:14 +00:00
optionalDependencies:
'@types/node': 25.3.0
first commit
2026-02-21 03:10:39 +00:00
feat: implement MCP registry client with multi-source search Add registry client that queries Official, Glama, and Smithery MCP registries with caching, request deduplication, retry logic, and result ranking/dedup. Includes 53 tests covering all components. Also fix null priority values in cancelled tasks (19-21) that broke Task Master, and add new tasks 25-27 for registry completion and CLI discover/install commands. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 03:46:14 +00:00
'@inquirer/confirm@5.1.21(@types/node@25.3.0)':
first commit
2026-02-21 03:10:39 +00:00
dependencies:
feat: implement MCP registry client with multi-source search Add registry client that queries Official, Glama, and Smithery MCP registries with caching, request deduplication, retry logic, and result ranking/dedup. Includes 53 tests covering all components. Also fix null priority values in cancelled tasks (19-21) that broke Task Master, and add new tasks 25-27 for registry completion and CLI discover/install commands. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 03:46:14 +00:00
'@inquirer/core': 10.3.2(@types/node@25.3.0)
'@inquirer/type': 3.0.10(@types/node@25.3.0)
optionalDependencies:
'@types/node': 25.3.0
first commit
2026-02-21 03:10:39 +00:00
feat: implement MCP registry client with multi-source search Add registry client that queries Official, Glama, and Smithery MCP registries with caching, request deduplication, retry logic, and result ranking/dedup. Includes 53 tests covering all components. Also fix null priority values in cancelled tasks (19-21) that broke Task Master, and add new tasks 25-27 for registry completion and CLI discover/install commands. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 03:46:14 +00:00
'@inquirer/core@10.3.2(@types/node@25.3.0)':
first commit
2026-02-21 03:10:39 +00:00
dependencies:
'@inquirer/ansi': 1.0.2
'@inquirer/figures': 1.0.15
feat: implement MCP registry client with multi-source search Add registry client that queries Official, Glama, and Smithery MCP registries with caching, request deduplication, retry logic, and result ranking/dedup. Includes 53 tests covering all components. Also fix null priority values in cancelled tasks (19-21) that broke Task Master, and add new tasks 25-27 for registry completion and CLI discover/install commands. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 03:46:14 +00:00
'@inquirer/type': 3.0.10(@types/node@25.3.0)
first commit
2026-02-21 03:10:39 +00:00
cli-width: 4.1.0
mute-stream: 2.0.0
signal-exit: 4.1.0
wrap-ansi: 6.2.0
yoctocolors-cjs: 2.1.3
feat: implement MCP registry client with multi-source search Add registry client that queries Official, Glama, and Smithery MCP registries with caching, request deduplication, retry logic, and result ranking/dedup. Includes 53 tests covering all components. Also fix null priority values in cancelled tasks (19-21) that broke Task Master, and add new tasks 25-27 for registry completion and CLI discover/install commands. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 03:46:14 +00:00
optionalDependencies:
'@types/node': 25.3.0
first commit
2026-02-21 03:10:39 +00:00
feat: implement MCP registry client with multi-source search Add registry client that queries Official, Glama, and Smithery MCP registries with caching, request deduplication, retry logic, and result ranking/dedup. Includes 53 tests covering all components. Also fix null priority values in cancelled tasks (19-21) that broke Task Master, and add new tasks 25-27 for registry completion and CLI discover/install commands. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 03:46:14 +00:00
'@inquirer/editor@4.2.23(@types/node@25.3.0)':
first commit
2026-02-21 03:10:39 +00:00
dependencies:
feat: implement MCP registry client with multi-source search Add registry client that queries Official, Glama, and Smithery MCP registries with caching, request deduplication, retry logic, and result ranking/dedup. Includes 53 tests covering all components. Also fix null priority values in cancelled tasks (19-21) that broke Task Master, and add new tasks 25-27 for registry completion and CLI discover/install commands. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 03:46:14 +00:00
'@inquirer/core': 10.3.2(@types/node@25.3.0)
'@inquirer/external-editor': 1.0.3(@types/node@25.3.0)
'@inquirer/type': 3.0.10(@types/node@25.3.0)
optionalDependencies:
'@types/node': 25.3.0
first commit
2026-02-21 03:10:39 +00:00
feat: implement MCP registry client with multi-source search Add registry client that queries Official, Glama, and Smithery MCP registries with caching, request deduplication, retry logic, and result ranking/dedup. Includes 53 tests covering all components. Also fix null priority values in cancelled tasks (19-21) that broke Task Master, and add new tasks 25-27 for registry completion and CLI discover/install commands. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 03:46:14 +00:00
'@inquirer/expand@4.0.23(@types/node@25.3.0)':
first commit
2026-02-21 03:10:39 +00:00
dependencies:
feat: implement MCP registry client with multi-source search Add registry client that queries Official, Glama, and Smithery MCP registries with caching, request deduplication, retry logic, and result ranking/dedup. Includes 53 tests covering all components. Also fix null priority values in cancelled tasks (19-21) that broke Task Master, and add new tasks 25-27 for registry completion and CLI discover/install commands. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 03:46:14 +00:00
'@inquirer/core': 10.3.2(@types/node@25.3.0)
'@inquirer/type': 3.0.10(@types/node@25.3.0)
first commit
2026-02-21 03:10:39 +00:00
yoctocolors-cjs: 2.1.3
feat: implement MCP registry client with multi-source search Add registry client that queries Official, Glama, and Smithery MCP registries with caching, request deduplication, retry logic, and result ranking/dedup. Includes 53 tests covering all components. Also fix null priority values in cancelled tasks (19-21) that broke Task Master, and add new tasks 25-27 for registry completion and CLI discover/install commands. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 03:46:14 +00:00
optionalDependencies:
'@types/node': 25.3.0
first commit
2026-02-21 03:10:39 +00:00
feat: implement MCP registry client with multi-source search Add registry client that queries Official, Glama, and Smithery MCP registries with caching, request deduplication, retry logic, and result ranking/dedup. Includes 53 tests covering all components. Also fix null priority values in cancelled tasks (19-21) that broke Task Master, and add new tasks 25-27 for registry completion and CLI discover/install commands. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 03:46:14 +00:00
'@inquirer/external-editor@1.0.3(@types/node@25.3.0)':
first commit
2026-02-21 03:10:39 +00:00
dependencies:
chardet: 2.1.1
iconv-lite: 0.7.2
feat: implement MCP registry client with multi-source search Add registry client that queries Official, Glama, and Smithery MCP registries with caching, request deduplication, retry logic, and result ranking/dedup. Includes 53 tests covering all components. Also fix null priority values in cancelled tasks (19-21) that broke Task Master, and add new tasks 25-27 for registry completion and CLI discover/install commands. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 03:46:14 +00:00
optionalDependencies:
'@types/node': 25.3.0
first commit
2026-02-21 03:10:39 +00:00
'@inquirer/figures@1.0.15': {}
feat: implement MCP registry client with multi-source search Add registry client that queries Official, Glama, and Smithery MCP registries with caching, request deduplication, retry logic, and result ranking/dedup. Includes 53 tests covering all components. Also fix null priority values in cancelled tasks (19-21) that broke Task Master, and add new tasks 25-27 for registry completion and CLI discover/install commands. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 03:46:14 +00:00
'@inquirer/input@4.3.1(@types/node@25.3.0)':
first commit
2026-02-21 03:10:39 +00:00
dependencies:
feat: implement MCP registry client with multi-source search Add registry client that queries Official, Glama, and Smithery MCP registries with caching, request deduplication, retry logic, and result ranking/dedup. Includes 53 tests covering all components. Also fix null priority values in cancelled tasks (19-21) that broke Task Master, and add new tasks 25-27 for registry completion and CLI discover/install commands. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 03:46:14 +00:00
'@inquirer/core': 10.3.2(@types/node@25.3.0)
'@inquirer/type': 3.0.10(@types/node@25.3.0)
optionalDependencies:
'@types/node': 25.3.0
first commit
2026-02-21 03:10:39 +00:00
feat: implement MCP registry client with multi-source search Add registry client that queries Official, Glama, and Smithery MCP registries with caching, request deduplication, retry logic, and result ranking/dedup. Includes 53 tests covering all components. Also fix null priority values in cancelled tasks (19-21) that broke Task Master, and add new tasks 25-27 for registry completion and CLI discover/install commands. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 03:46:14 +00:00
'@inquirer/number@3.0.23(@types/node@25.3.0)':
first commit
2026-02-21 03:10:39 +00:00
dependencies:
feat: implement MCP registry client with multi-source search Add registry client that queries Official, Glama, and Smithery MCP registries with caching, request deduplication, retry logic, and result ranking/dedup. Includes 53 tests covering all components. Also fix null priority values in cancelled tasks (19-21) that broke Task Master, and add new tasks 25-27 for registry completion and CLI discover/install commands. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 03:46:14 +00:00
'@inquirer/core': 10.3.2(@types/node@25.3.0)
'@inquirer/type': 3.0.10(@types/node@25.3.0)
optionalDependencies:
'@types/node': 25.3.0
first commit
2026-02-21 03:10:39 +00:00
feat: implement MCP registry client with multi-source search Add registry client that queries Official, Glama, and Smithery MCP registries with caching, request deduplication, retry logic, and result ranking/dedup. Includes 53 tests covering all components. Also fix null priority values in cancelled tasks (19-21) that broke Task Master, and add new tasks 25-27 for registry completion and CLI discover/install commands. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 03:46:14 +00:00
'@inquirer/password@4.0.23(@types/node@25.3.0)':
first commit
2026-02-21 03:10:39 +00:00
dependencies:
'@inquirer/ansi': 1.0.2
feat: implement MCP registry client with multi-source search Add registry client that queries Official, Glama, and Smithery MCP registries with caching, request deduplication, retry logic, and result ranking/dedup. Includes 53 tests covering all components. Also fix null priority values in cancelled tasks (19-21) that broke Task Master, and add new tasks 25-27 for registry completion and CLI discover/install commands. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 03:46:14 +00:00
'@inquirer/core': 10.3.2(@types/node@25.3.0)
'@inquirer/type': 3.0.10(@types/node@25.3.0)
optionalDependencies:
'@types/node': 25.3.0
'@inquirer/prompts@7.10.1(@types/node@25.3.0)':
dependencies:
'@inquirer/checkbox': 4.3.2(@types/node@25.3.0)
'@inquirer/confirm': 5.1.21(@types/node@25.3.0)
'@inquirer/editor': 4.2.23(@types/node@25.3.0)
'@inquirer/expand': 4.0.23(@types/node@25.3.0)
'@inquirer/input': 4.3.1(@types/node@25.3.0)
'@inquirer/number': 3.0.23(@types/node@25.3.0)
'@inquirer/password': 4.0.23(@types/node@25.3.0)
'@inquirer/rawlist': 4.1.11(@types/node@25.3.0)
'@inquirer/search': 3.2.2(@types/node@25.3.0)
'@inquirer/select': 4.4.2(@types/node@25.3.0)
optionalDependencies:
'@types/node': 25.3.0
first commit
2026-02-21 03:10:39 +00:00
feat: implement MCP registry client with multi-source search Add registry client that queries Official, Glama, and Smithery MCP registries with caching, request deduplication, retry logic, and result ranking/dedup. Includes 53 tests covering all components. Also fix null priority values in cancelled tasks (19-21) that broke Task Master, and add new tasks 25-27 for registry completion and CLI discover/install commands. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 03:46:14 +00:00
'@inquirer/rawlist@4.1.11(@types/node@25.3.0)':
first commit
2026-02-21 03:10:39 +00:00
dependencies:
feat: implement MCP registry client with multi-source search Add registry client that queries Official, Glama, and Smithery MCP registries with caching, request deduplication, retry logic, and result ranking/dedup. Includes 53 tests covering all components. Also fix null priority values in cancelled tasks (19-21) that broke Task Master, and add new tasks 25-27 for registry completion and CLI discover/install commands. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 03:46:14 +00:00
'@inquirer/core': 10.3.2(@types/node@25.3.0)
'@inquirer/type': 3.0.10(@types/node@25.3.0)
first commit
2026-02-21 03:10:39 +00:00
yoctocolors-cjs: 2.1.3
feat: implement MCP registry client with multi-source search Add registry client that queries Official, Glama, and Smithery MCP registries with caching, request deduplication, retry logic, and result ranking/dedup. Includes 53 tests covering all components. Also fix null priority values in cancelled tasks (19-21) that broke Task Master, and add new tasks 25-27 for registry completion and CLI discover/install commands. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 03:46:14 +00:00
optionalDependencies:
'@types/node': 25.3.0
first commit
2026-02-21 03:10:39 +00:00
feat: implement MCP registry client with multi-source search Add registry client that queries Official, Glama, and Smithery MCP registries with caching, request deduplication, retry logic, and result ranking/dedup. Includes 53 tests covering all components. Also fix null priority values in cancelled tasks (19-21) that broke Task Master, and add new tasks 25-27 for registry completion and CLI discover/install commands. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 03:46:14 +00:00
'@inquirer/search@3.2.2(@types/node@25.3.0)':
first commit
2026-02-21 03:10:39 +00:00
dependencies:
feat: implement MCP registry client with multi-source search Add registry client that queries Official, Glama, and Smithery MCP registries with caching, request deduplication, retry logic, and result ranking/dedup. Includes 53 tests covering all components. Also fix null priority values in cancelled tasks (19-21) that broke Task Master, and add new tasks 25-27 for registry completion and CLI discover/install commands. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 03:46:14 +00:00
'@inquirer/core': 10.3.2(@types/node@25.3.0)
first commit
2026-02-21 03:10:39 +00:00
'@inquirer/figures': 1.0.15
feat: implement MCP registry client with multi-source search Add registry client that queries Official, Glama, and Smithery MCP registries with caching, request deduplication, retry logic, and result ranking/dedup. Includes 53 tests covering all components. Also fix null priority values in cancelled tasks (19-21) that broke Task Master, and add new tasks 25-27 for registry completion and CLI discover/install commands. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 03:46:14 +00:00
'@inquirer/type': 3.0.10(@types/node@25.3.0)
first commit
2026-02-21 03:10:39 +00:00
yoctocolors-cjs: 2.1.3
feat: implement MCP registry client with multi-source search Add registry client that queries Official, Glama, and Smithery MCP registries with caching, request deduplication, retry logic, and result ranking/dedup. Includes 53 tests covering all components. Also fix null priority values in cancelled tasks (19-21) that broke Task Master, and add new tasks 25-27 for registry completion and CLI discover/install commands. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 03:46:14 +00:00
optionalDependencies:
'@types/node': 25.3.0
first commit
2026-02-21 03:10:39 +00:00
feat: implement MCP registry client with multi-source search Add registry client that queries Official, Glama, and Smithery MCP registries with caching, request deduplication, retry logic, and result ranking/dedup. Includes 53 tests covering all components. Also fix null priority values in cancelled tasks (19-21) that broke Task Master, and add new tasks 25-27 for registry completion and CLI discover/install commands. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 03:46:14 +00:00
'@inquirer/select@4.4.2(@types/node@25.3.0)':
first commit
2026-02-21 03:10:39 +00:00
dependencies:
'@inquirer/ansi': 1.0.2
feat: implement MCP registry client with multi-source search Add registry client that queries Official, Glama, and Smithery MCP registries with caching, request deduplication, retry logic, and result ranking/dedup. Includes 53 tests covering all components. Also fix null priority values in cancelled tasks (19-21) that broke Task Master, and add new tasks 25-27 for registry completion and CLI discover/install commands. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 03:46:14 +00:00
'@inquirer/core': 10.3.2(@types/node@25.3.0)
first commit
2026-02-21 03:10:39 +00:00
'@inquirer/figures': 1.0.15
feat: implement MCP registry client with multi-source search Add registry client that queries Official, Glama, and Smithery MCP registries with caching, request deduplication, retry logic, and result ranking/dedup. Includes 53 tests covering all components. Also fix null priority values in cancelled tasks (19-21) that broke Task Master, and add new tasks 25-27 for registry completion and CLI discover/install commands. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 03:46:14 +00:00
'@inquirer/type': 3.0.10(@types/node@25.3.0)
first commit
2026-02-21 03:10:39 +00:00
yoctocolors-cjs: 2.1.3
feat: implement MCP registry client with multi-source search Add registry client that queries Official, Glama, and Smithery MCP registries with caching, request deduplication, retry logic, and result ranking/dedup. Includes 53 tests covering all components. Also fix null priority values in cancelled tasks (19-21) that broke Task Master, and add new tasks 25-27 for registry completion and CLI discover/install commands. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 03:46:14 +00:00
optionalDependencies:
'@types/node': 25.3.0
first commit
2026-02-21 03:10:39 +00:00
feat: implement MCP registry client with multi-source search Add registry client that queries Official, Glama, and Smithery MCP registries with caching, request deduplication, retry logic, and result ranking/dedup. Includes 53 tests covering all components. Also fix null priority values in cancelled tasks (19-21) that broke Task Master, and add new tasks 25-27 for registry completion and CLI discover/install commands. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 03:46:14 +00:00
'@inquirer/type@3.0.10(@types/node@25.3.0)':
optionalDependencies:
'@types/node': 25.3.0
first commit
2026-02-21 03:10:39 +00:00
feat(mcpd+deploy): serve web UI at /ui + smoke tests + docs (Stage 6) The closing stage. mcpd now hosts the Stage 5 SPA, the Docker image bundles the build artifact, a smoke test exercises the personality HTTP surface end-to-end, and the user-facing docs spell out the mental model. mcpd: - Add @fastify/static dep. - New routes/web-ui.ts: registers /ui/* against a static bundle. Looks for the bundle at $MCPD_WEB_ROOT, then /usr/share/mcpd/web (the Docker image path), then a dev-tree fallback. Logs and skips cleanly if missing — API-only deploys keep working. - SPA fallback: any /ui/<path> that doesn't match a file falls through to index.html so direct hits to react-router URLs work. - /ui/* falls through to `kind: skip` in mapUrlToPermission, so the static assets are served unauthenticated. Each API call from the SPA still carries the bearer token. Deploy: - Dockerfile.mcpd builds the @mcpctl/web bundle in the same builder stage and copies dist/ to /usr/share/mcpd/web in the runtime image. Smoke (personality.smoke.test.ts): - Live mcpd flow: create secret/llm/agent/personality, attach an agent-direct prompt, verify the binding listing, reject double- attach (409) + foreign-agent prompt (400), set defaultPersonality by name, detach + delete cleanup. Docs: - New docs/personalities.md: VLAN-on-ethernet model, system-block ordering table, three prompt scopes, CLI walkthrough, web UI walkthrough, full API surface, RBAC notes. - agents.md and chat.md cross-link. - README's Agents section gains a Personalities subsection. Test count after Stage 6: mcpd: 801/801 cli: 430/430 web: 7/7 db: 58/62 (4 pre-existing) Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:48:43 +01:00
'@isaacs/cliui@9.0.0': {}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
'@isaacs/fs-minipass@4.0.1':
dependencies:
minipass: 7.1.3
'@isaacs/string-locale-compare@1.1.0': {}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
'@jridgewell/gen-mapping@0.3.13':
dependencies:
'@jridgewell/sourcemap-codec': 1.5.5
'@jridgewell/trace-mapping': 0.3.31
'@jridgewell/remapping@2.3.5':
dependencies:
'@jridgewell/gen-mapping': 0.3.13
'@jridgewell/trace-mapping': 0.3.31
first commit
2026-02-21 03:10:39 +00:00
'@jridgewell/resolve-uri@3.1.2': {}
'@jridgewell/sourcemap-codec@1.5.5': {}
'@jridgewell/trace-mapping@0.3.31':
dependencies:
'@jridgewell/resolve-uri': 3.1.2
'@jridgewell/sourcemap-codec': 1.5.5
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
'@js-sdsl/ordered-map@4.4.2': {}
feat: add Kubernetes orchestrator for MCP server pod management mcpd can now deploy MCP server instances as Kubernetes pods instead of Docker containers. Set MCPD_ORCHESTRATOR=kubernetes to enable. - Add @kubernetes/client-node with thin wrapper (context enforcement via MCPD_K8S_CONTEXT to prevent multi-cluster mishaps) - Rewrite KubernetesOrchestrator: pod CRUD, pod IP extraction, exec via SPDY (one-shot + interactive), log streaming - Manifest generator: stdin:true for STDIO servers, args (not command) to preserve runner image entrypoint, security hardening - Orchestrator selection in main.ts via MCPD_ORCHESTRATOR env var - 25 unit tests for k8s orchestrator, all 624 tests pass Tested end-to-end on local k3s: - mcpd deployed via Pulumi, creates pods in mcpctl-servers namespace - NetworkPolicy verified: only mcpd can reach MCP server pods - Python runner (uvx) successfully runs aws-documentation-mcp-server Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-08 01:55:13 +01:00
'@jsep-plugin/assignment@1.3.0(jsep@1.4.0)':
dependencies:
jsep: 1.4.0
'@jsep-plugin/regex@1.0.4(jsep@1.4.0)':
dependencies:
jsep: 1.4.0
'@kubernetes/client-node@1.4.0':
dependencies:
'@types/js-yaml': 4.0.9
'@types/node': 24.12.2
'@types/node-fetch': 2.6.13
'@types/stream-buffers': 3.0.8
form-data: 4.0.5
hpagent: 1.2.0
isomorphic-ws: 5.0.0(ws@8.19.0)
js-yaml: 4.1.1
jsonpath-plus: 10.4.0
node-fetch: 2.7.0
openid-client: 6.8.2
rfc4648: 1.5.4
socks-proxy-agent: 8.0.5
stream-buffers: 3.0.3
tar-fs: 3.1.2
ws: 8.19.0
transitivePeerDependencies:
- bare-abort-controller
- bare-buffer
- bufferutil
- encoding
- react-native-b4a
- supports-color
- utf-8-validate
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
'@logdna/tail-file@2.2.0': {}
first commit
2026-02-21 03:10:39 +00:00
'@lukeed/ms@2.0.2': {}
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
'@mapbox/node-pre-gyp@1.0.11':
dependencies:
detect-libc: 2.1.2
https-proxy-agent: 5.0.1
make-dir: 3.1.0
node-fetch: 2.7.0
nopt: 5.0.0
npmlog: 5.0.1
rimraf: 3.0.2
semver: 7.7.4
tar: 6.2.1
transitivePeerDependencies:
- encoding
- supports-color
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
'@mariozechner/clipboard-darwin-arm64@0.3.9':
optional: true
'@mariozechner/clipboard-darwin-universal@0.3.9':
optional: true
'@mariozechner/clipboard-darwin-x64@0.3.9':
optional: true
'@mariozechner/clipboard-linux-arm64-gnu@0.3.9':
optional: true
'@mariozechner/clipboard-linux-arm64-musl@0.3.9':
optional: true
'@mariozechner/clipboard-linux-riscv64-gnu@0.3.9':
optional: true
'@mariozechner/clipboard-linux-x64-gnu@0.3.9':
optional: true
'@mariozechner/clipboard-linux-x64-musl@0.3.9':
optional: true
'@mariozechner/clipboard-win32-arm64-msvc@0.3.9':
optional: true
'@mariozechner/clipboard-win32-x64-msvc@0.3.9':
optional: true
'@mariozechner/clipboard@0.3.9':
optionalDependencies:
'@mariozechner/clipboard-darwin-arm64': 0.3.9
'@mariozechner/clipboard-darwin-universal': 0.3.9
'@mariozechner/clipboard-darwin-x64': 0.3.9
'@mariozechner/clipboard-linux-arm64-gnu': 0.3.9
'@mariozechner/clipboard-linux-arm64-musl': 0.3.9
'@mariozechner/clipboard-linux-riscv64-gnu': 0.3.9
'@mariozechner/clipboard-linux-x64-gnu': 0.3.9
'@mariozechner/clipboard-linux-x64-musl': 0.3.9
'@mariozechner/clipboard-win32-arm64-msvc': 0.3.9
'@mariozechner/clipboard-win32-x64-msvc': 0.3.9
optional: true
'@mistralai/mistralai@2.2.6(@opentelemetry/api@1.9.0)':
dependencies:
'@opentelemetry/semantic-conventions': 1.43.0
ws: 8.19.0
zod: 3.25.76
zod-to-json-schema: 3.25.1(zod@3.25.76)
optionalDependencies:
'@opentelemetry/api': 1.9.0
transitivePeerDependencies:
- bufferutil
- utf-8-validate
first commit
2026-02-21 03:10:39 +00:00
'@modelcontextprotocol/sdk@1.26.0(zod@3.25.76)':
dependencies:
'@hono/node-server': 1.19.9(hono@4.12.0)
ajv: 8.18.0
ajv-formats: 3.0.1(ajv@8.18.0)
content-type: 1.0.5
cors: 2.8.6
cross-spawn: 7.0.6
eventsource: 3.0.7
eventsource-parser: 3.0.6
express: 5.2.1
express-rate-limit: 8.2.1(express@5.2.1)
hono: 4.12.0
jose: 6.1.3
json-schema-typed: 8.0.2
pkce-challenge: 5.0.1
raw-body: 3.0.2
zod: 3.25.76
zod-to-json-schema: 3.25.1(zod@3.25.76)
transitivePeerDependencies:
- supports-color
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
optional: true
'@modelcontextprotocol/sdk@1.26.0(zod@4.1.8)':
dependencies:
'@hono/node-server': 1.19.9(hono@4.12.0)
ajv: 8.18.0
ajv-formats: 3.0.1(ajv@8.18.0)
content-type: 1.0.5
cors: 2.8.6
cross-spawn: 7.0.6
eventsource: 3.0.7
eventsource-parser: 3.0.6
express: 5.2.1
express-rate-limit: 8.2.1(express@5.2.1)
hono: 4.12.0
jose: 6.1.3
json-schema-typed: 8.0.2
pkce-challenge: 5.0.1
raw-body: 3.0.2
zod: 4.1.8
zod-to-json-schema: 3.25.1(zod@4.1.8)
transitivePeerDependencies:
- supports-color
first commit
2026-02-21 03:10:39 +00:00
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
'@monaco-editor/loader@1.7.0':
dependencies:
state-local: 1.0.7
'@monaco-editor/react@4.7.0(monaco-editor@0.55.1)(react-dom@19.2.5(react@19.2.5))(react@19.2.5)':
dependencies:
'@monaco-editor/loader': 1.7.0
monaco-editor: 0.55.1
react: 19.2.5
react-dom: 19.2.5(react@19.2.5)
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
'@msgpackr-extract/msgpackr-extract-darwin-arm64@3.0.4':
optional: true
'@msgpackr-extract/msgpackr-extract-darwin-x64@3.0.4':
optional: true
'@msgpackr-extract/msgpackr-extract-linux-arm64@3.0.4':
optional: true
'@msgpackr-extract/msgpackr-extract-linux-arm@3.0.4':
optional: true
'@msgpackr-extract/msgpackr-extract-linux-x64@3.0.4':
optional: true
'@msgpackr-extract/msgpackr-extract-win32-x64@3.0.4':
optional: true
feat(web): bold redesign — Tailwind v4 + shadcn-style primitives + Skills/Proposals/Revisions UI Phase 6 of the Skills + Revisions + Proposals work. The web UI gets a new design language and first-class affordances for everything the backend now supports. ## Visual direction - Tailwind v4 with custom @theme block (oklch tokens). Dark-mode-only (internal tool — light mode doubles QA surface). - Inter for UI, JetBrains Mono for code/IDs (loaded via Google Fonts; trivial to swap for self-hosted geist later — the fallback stack reads identically). - Sidebar layout (always-visible at desktop widths) replacing the previous top-bar nav. Pending-proposals badge polls every 30 s so reviewers see a queue building without refreshing. - Lucide icons throughout. - Spacing and radii on Tailwind defaults. Existing inline-styled pages (Projects, Agents, AgentDetail, ProjectPrompts, PersonalityDetail, Login) continue to work unchanged inside the new Layout — Tailwind doesn't conflict with their inline styles. A follow-up can migrate them incrementally. ## What's added ### Build infra (src/web/) - package.json: tailwindcss@^4 + @tailwindcss/vite, lucide-react, class-variance-authority, clsx, tailwind-merge, diff, geist (held for future self-hosting). - vite.config.ts: registers the @tailwindcss/vite plugin. - src/index.css: Tailwind import + @theme tokens + @layer base. - src/main.tsx: imports index.css. - src/lib/utils.ts: shadcn-style cn() helper. ### shadcn-style primitives (src/components/ui/) Hand-written rather than generated via `npx shadcn` so the repo doesn't depend on a CLI tool that needs interactive runtime: - button.tsx — variants: primary / secondary / ghost / danger / link; sizes: sm / md / lg / icon. - card.tsx — Card + Header/Title/Description/Content/Footer subparts. - badge.tsx — variants: default / info / success / warning / danger / outline. - input.tsx — Input + Textarea + Label. - tabs.tsx — no-dep accessible Tabs (no Radix needed for our use). - separator.tsx — h/v separator with role=separator. ### Diff component (src/components/Diff.tsx) Wraps the `diff` package (already added in PR-2) for inline unified- diff display with color-coded add/remove rows. Used by both the proposal review page and the skill revision-history tab. ### New pages (src/pages/) - Dashboard.tsx — at-a-glance home. Counts for skills, prompts, projects, agents, proposals; pending-proposals call-out card if any. - Skills.tsx — list view, separated into Global vs Project/Agent- scoped sections. - SkillDetail.tsx — name + semver + description; tabs for SKILL.md / Files / Metadata / History. History tab shows revisions with click-to-diff against the live body. - Proposals.tsx — queue with Pending/Approved/Rejected tabs. Pending count is highlighted in amber. - ProposalDetail.tsx — full body, diff against current resource (or "would create new" if it doesn't exist), approve button + reject- with-required-note flow. ### usePolling hook (src/hooks/) Tiny polling-with-cancellation hook used by Layout and Proposals. ### Layout rewrite (src/components/Layout.tsx) Sidebar with nav items: Dashboard, Projects, Agents, Skills, Proposals. Lucide icons. Active-route highlighting via NavLink. Pending-proposals warning badge on the Proposals item. ### Routes (src/App.tsx) New routes: /dashboard, /skills, /skills/:name, /proposals, /proposals/:id. Default redirects to /dashboard. ### API types (src/api.ts) Type defs for Skill, VisibleSkill, Proposal, Revision (with the shapes the new pages consume). ## Tests Existing 7 web tests still pass (Login + api). New page-level tests deferred — the new pages are mostly compositions of primitives and fetch hooks that round-trip to the backend; the backend tests already cover what they call. PR-7 polish can add render-and-click tests if coverage drift surfaces. ## Verification - `pnpm --filter @mcpctl/web build` clean, no warnings. - `pnpm test:run` whole monorepo: 162 test files / 2157 tests green. - Visual smoke deferred — needs a running mcpd to populate the fixtures. Manual smoke tested locally is the next step. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-07 17:54:55 +01:00
'@next/env@16.2.5': {}
'@next/swc-darwin-arm64@16.2.5':
optional: true
'@next/swc-darwin-x64@16.2.5':
optional: true
'@next/swc-linux-arm64-gnu@16.2.5':
optional: true
'@next/swc-linux-arm64-musl@16.2.5':
optional: true
'@next/swc-linux-x64-gnu@16.2.5':
optional: true
'@next/swc-linux-x64-musl@16.2.5':
optional: true
'@next/swc-win32-arm64-msvc@16.2.5':
optional: true
'@next/swc-win32-x64-msvc@16.2.5':
optional: true
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
'@npmcli/agent@4.0.2':
dependencies:
agent-base: 7.1.4
http-proxy-agent: 7.0.2
https-proxy-agent: 7.0.6
lru-cache: 11.2.6
socks-proxy-agent: 8.0.5
transitivePeerDependencies:
- supports-color
'@npmcli/arborist@9.9.0':
dependencies:
'@gar/promise-retry': 1.0.3
'@isaacs/string-locale-compare': 1.1.0
'@npmcli/fs': 5.0.0
'@npmcli/installed-package-contents': 4.0.0
'@npmcli/map-workspaces': 5.0.3
'@npmcli/metavuln-calculator': 9.0.3
'@npmcli/name-from-folder': 4.0.0
'@npmcli/node-gyp': 5.0.0
'@npmcli/package-json': 7.0.5
'@npmcli/query': 5.0.0
'@npmcli/redact': 4.0.0
'@npmcli/run-script': 10.0.4
bin-links: 6.0.2
cacache: 20.0.4
common-ancestor-path: 2.0.0
hosted-git-info: 9.0.3
json-stringify-nice: 1.1.4
lru-cache: 11.2.6
minimatch: 10.2.2
nopt: 9.0.0
npm-install-checks: 8.0.0
npm-package-arg: 13.0.2
npm-pick-manifest: 11.0.3
npm-registry-fetch: 19.1.1
pacote: 21.5.1
parse-conflict-json: 5.0.1
proc-log: 6.1.0
proggy: 4.0.0
promise-all-reject-late: 1.0.1
promise-call-limit: 3.0.2
semver: 7.7.4
ssri: 13.0.1
treeverse: 3.0.0
walk-up-path: 4.0.0
transitivePeerDependencies:
- supports-color
'@npmcli/fs@5.0.0':
dependencies:
semver: 7.7.4
'@npmcli/git@7.0.2':
dependencies:
'@gar/promise-retry': 1.0.3
'@npmcli/promise-spawn': 9.0.1
ini: 6.0.0
lru-cache: 11.2.6
npm-pick-manifest: 11.0.3
proc-log: 6.1.0
semver: 7.7.4
which: 6.0.1
'@npmcli/installed-package-contents@4.0.0':
dependencies:
npm-bundled: 5.0.0
npm-normalize-package-bin: 5.0.0
'@npmcli/map-workspaces@5.0.3':
dependencies:
'@npmcli/name-from-folder': 4.0.0
'@npmcli/package-json': 7.0.5
glob: 13.0.6
minimatch: 10.2.2
'@npmcli/metavuln-calculator@9.0.3':
dependencies:
cacache: 20.0.4
json-parse-even-better-errors: 5.0.0
pacote: 21.5.1
proc-log: 6.1.0
semver: 7.7.4
transitivePeerDependencies:
- supports-color
'@npmcli/name-from-folder@4.0.0': {}
'@npmcli/node-gyp@5.0.0': {}
'@npmcli/package-json@7.0.5':
dependencies:
'@npmcli/git': 7.0.2
glob: 13.0.6
hosted-git-info: 9.0.3
json-parse-even-better-errors: 5.0.0
proc-log: 6.1.0
semver: 7.7.4
spdx-expression-parse: 4.0.0
'@npmcli/promise-spawn@9.0.1':
dependencies:
which: 6.0.1
'@npmcli/query@5.0.0':
dependencies:
postcss-selector-parser: 7.1.4
'@npmcli/redact@4.0.0': {}
'@npmcli/run-script@10.0.4':
dependencies:
'@npmcli/node-gyp': 5.0.0
'@npmcli/package-json': 7.0.5
'@npmcli/promise-spawn': 9.0.1
node-gyp: 12.4.0
proc-log: 6.1.0
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
'@opencode-ai/plugin@1.18.15(@opentui/core@0.4.5(typescript@5.9.3)(web-tree-sitter@0.25.10))(@opentui/keymap@0.4.5(@opentui/solid@0.4.5(solid-js@1.9.12)(typescript@5.9.3)(web-tree-sitter@0.25.10))(react@19.2.5)(solid-js@1.9.12)(typescript@5.9.3)(web-tree-sitter@0.25.10))(@opentui/solid@0.4.5(solid-js@1.9.12)(typescript@5.9.3)(web-tree-sitter@0.25.10))':
dependencies:
'@ai-sdk/provider': 3.0.8
'@opencode-ai/sdk': 1.18.15
effect: 4.0.0-beta.83
zod: 4.1.8
optionalDependencies:
'@opentui/core': 0.4.5(typescript@5.9.3)(web-tree-sitter@0.25.10)
'@opentui/keymap': 0.4.5(@opentui/solid@0.4.5(solid-js@1.9.12)(typescript@5.9.3)(web-tree-sitter@0.25.10))(react@19.2.5)(solid-js@1.9.12)(typescript@5.9.3)(web-tree-sitter@0.25.10)
'@opentui/solid': 0.4.5(solid-js@1.9.12)(typescript@5.9.3)(web-tree-sitter@0.25.10)
'@opencode-ai/sdk@1.18.15':
dependencies:
cross-spawn: 7.0.6
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
'@opentelemetry/api-logs@0.220.0':
dependencies:
'@opentelemetry/api': 1.9.1
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
'@opentelemetry/api@1.9.0': {}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
'@opentelemetry/api@1.9.1': {}
'@opentelemetry/context-async-hooks@2.9.0(@opentelemetry/api@1.9.1)':
dependencies:
'@opentelemetry/api': 1.9.1
'@opentelemetry/core@2.9.0(@opentelemetry/api@1.9.1)':
dependencies:
'@opentelemetry/api': 1.9.1
'@opentelemetry/semantic-conventions': 1.43.0
'@opentelemetry/exporter-trace-otlp-grpc@0.220.0(@opentelemetry/api@1.9.1)':
dependencies:
'@grpc/grpc-js': 1.14.3
'@opentelemetry/api': 1.9.1
'@opentelemetry/otlp-exporter-base': 0.220.0(@opentelemetry/api@1.9.1)
'@opentelemetry/otlp-grpc-exporter-base': 0.220.0(@opentelemetry/api@1.9.1)
'@opentelemetry/otlp-transformer': 0.220.0(@opentelemetry/api@1.9.1)
'@opentelemetry/sdk-trace': 2.9.0(@opentelemetry/api@1.9.1)
'@opentelemetry/exporter-zipkin@2.9.0(@opentelemetry/api@1.9.1)':
dependencies:
'@opentelemetry/api': 1.9.1
'@opentelemetry/core': 2.9.0(@opentelemetry/api@1.9.1)
'@opentelemetry/resources': 2.9.0(@opentelemetry/api@1.9.1)
'@opentelemetry/sdk-trace': 2.9.0(@opentelemetry/api@1.9.1)
'@opentelemetry/semantic-conventions': 1.43.0
'@opentelemetry/instrumentation-grpc@0.220.0(@opentelemetry/api@1.9.1)':
dependencies:
'@opentelemetry/api': 1.9.1
'@opentelemetry/instrumentation': 0.220.0(@opentelemetry/api@1.9.1)
'@opentelemetry/semantic-conventions': 1.43.0
transitivePeerDependencies:
- supports-color
'@opentelemetry/instrumentation@0.220.0(@opentelemetry/api@1.9.1)':
dependencies:
'@opentelemetry/api': 1.9.1
'@opentelemetry/api-logs': 0.220.0
import-in-the-middle: 3.3.1
require-in-the-middle: 8.0.1
transitivePeerDependencies:
- supports-color
'@opentelemetry/otlp-exporter-base@0.220.0(@opentelemetry/api@1.9.1)':
dependencies:
'@opentelemetry/api': 1.9.1
'@opentelemetry/core': 2.9.0(@opentelemetry/api@1.9.1)
'@opentelemetry/otlp-transformer': 0.220.0(@opentelemetry/api@1.9.1)
'@opentelemetry/otlp-grpc-exporter-base@0.220.0(@opentelemetry/api@1.9.1)':
dependencies:
'@grpc/grpc-js': 1.14.3
'@opentelemetry/api': 1.9.1
'@opentelemetry/core': 2.9.0(@opentelemetry/api@1.9.1)
'@opentelemetry/otlp-exporter-base': 0.220.0(@opentelemetry/api@1.9.1)
'@opentelemetry/otlp-transformer': 0.220.0(@opentelemetry/api@1.9.1)
'@opentelemetry/otlp-transformer@0.220.0(@opentelemetry/api@1.9.1)':
dependencies:
'@opentelemetry/api': 1.9.1
'@opentelemetry/api-logs': 0.220.0
'@opentelemetry/core': 2.9.0(@opentelemetry/api@1.9.1)
'@opentelemetry/resources': 2.9.0(@opentelemetry/api@1.9.1)
'@opentelemetry/sdk-logs': 0.220.0(@opentelemetry/api@1.9.1)
'@opentelemetry/sdk-metrics': 2.9.0(@opentelemetry/api@1.9.1)
'@opentelemetry/sdk-trace': 2.9.0(@opentelemetry/api@1.9.1)
'@opentelemetry/resources@2.9.0(@opentelemetry/api@1.9.1)':
dependencies:
'@opentelemetry/api': 1.9.1
'@opentelemetry/core': 2.9.0(@opentelemetry/api@1.9.1)
'@opentelemetry/semantic-conventions': 1.43.0
'@opentelemetry/sdk-logs@0.220.0(@opentelemetry/api@1.9.1)':
dependencies:
'@opentelemetry/api': 1.9.1
'@opentelemetry/api-logs': 0.220.0
'@opentelemetry/core': 2.9.0(@opentelemetry/api@1.9.1)
'@opentelemetry/resources': 2.9.0(@opentelemetry/api@1.9.1)
'@opentelemetry/semantic-conventions': 1.43.0
'@opentelemetry/sdk-metrics@2.9.0(@opentelemetry/api@1.9.1)':
dependencies:
'@opentelemetry/api': 1.9.1
'@opentelemetry/core': 2.9.0(@opentelemetry/api@1.9.1)
'@opentelemetry/resources': 2.9.0(@opentelemetry/api@1.9.1)
'@opentelemetry/sdk-trace-base@2.9.0(@opentelemetry/api@1.9.1)':
dependencies:
'@opentelemetry/api': 1.9.1
'@opentelemetry/core': 2.9.0(@opentelemetry/api@1.9.1)
'@opentelemetry/resources': 2.9.0(@opentelemetry/api@1.9.1)
'@opentelemetry/sdk-trace': 2.9.0(@opentelemetry/api@1.9.1)
'@opentelemetry/semantic-conventions': 1.43.0
'@opentelemetry/sdk-trace-node@2.9.0(@opentelemetry/api@1.9.1)':
dependencies:
'@opentelemetry/api': 1.9.1
'@opentelemetry/context-async-hooks': 2.9.0(@opentelemetry/api@1.9.1)
'@opentelemetry/core': 2.9.0(@opentelemetry/api@1.9.1)
'@opentelemetry/sdk-trace-base': 2.9.0(@opentelemetry/api@1.9.1)
'@opentelemetry/sdk-trace@2.9.0(@opentelemetry/api@1.9.1)':
dependencies:
'@opentelemetry/api': 1.9.1
'@opentelemetry/core': 2.9.0(@opentelemetry/api@1.9.1)
'@opentelemetry/resources': 2.9.0(@opentelemetry/api@1.9.1)
'@opentelemetry/semantic-conventions': 1.43.0
'@opentelemetry/semantic-conventions@1.43.0': {}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
'@opentui/core-darwin-arm64@0.4.5':
optional: true
'@opentui/core-darwin-x64@0.4.5':
optional: true
'@opentui/core-linux-arm64-musl@0.4.5':
optional: true
'@opentui/core-linux-arm64@0.4.5':
optional: true
'@opentui/core-linux-x64-musl@0.4.5':
optional: true
'@opentui/core-linux-x64@0.4.5':
optional: true
'@opentui/core-win32-arm64@0.4.5':
optional: true
'@opentui/core-win32-x64@0.4.5':
optional: true
'@opentui/core@0.4.5(typescript@5.9.3)(web-tree-sitter@0.25.10)':
dependencies:
bun-ffi-structs: 0.2.4(typescript@5.9.3)
diff: 9.0.0
marked: 17.0.1
string-width: 7.2.0
strip-ansi: 7.1.2
web-tree-sitter: 0.25.10
optionalDependencies:
'@opentui/core-darwin-arm64': 0.4.5
'@opentui/core-darwin-x64': 0.4.5
'@opentui/core-linux-arm64': 0.4.5
'@opentui/core-linux-arm64-musl': 0.4.5
'@opentui/core-linux-x64': 0.4.5
'@opentui/core-linux-x64-musl': 0.4.5
'@opentui/core-win32-arm64': 0.4.5
'@opentui/core-win32-x64': 0.4.5
transitivePeerDependencies:
- typescript
'@opentui/keymap@0.4.5(@opentui/solid@0.4.5(solid-js@1.9.12)(typescript@5.9.3)(web-tree-sitter@0.25.10))(react@19.2.5)(solid-js@1.9.12)(typescript@5.9.3)(web-tree-sitter@0.25.10)':
dependencies:
'@opentui/core': 0.4.5(typescript@5.9.3)(web-tree-sitter@0.25.10)
optionalDependencies:
'@opentui/solid': 0.4.5(solid-js@1.9.12)(typescript@5.9.3)(web-tree-sitter@0.25.10)
react: 19.2.5
solid-js: 1.9.12
transitivePeerDependencies:
- typescript
- web-tree-sitter
'@opentui/solid@0.4.5(solid-js@1.9.12)(typescript@5.9.3)(web-tree-sitter@0.25.10)':
dependencies:
'@babel/core': 7.28.0
'@babel/preset-typescript': 7.27.1(@babel/core@7.28.0)
'@opentui/core': 0.4.5(typescript@5.9.3)(web-tree-sitter@0.25.10)
babel-plugin-module-resolver: 5.0.2
babel-preset-solid: 1.9.12(@babel/core@7.28.0)(solid-js@1.9.12)
entities: 7.0.1
s-js: 0.4.9
solid-js: 1.9.12
transitivePeerDependencies:
- supports-color
- typescript
- web-tree-sitter
first commit
2026-02-21 03:10:39 +00:00
'@pinojs/redact@0.4.0': {}
'@prisma/client@6.19.2(prisma@6.19.2(typescript@5.9.3))(typescript@5.9.3)':
optionalDependencies:
prisma: 6.19.2(typescript@5.9.3)
typescript: 5.9.3
'@prisma/config@6.19.2':
dependencies:
c12: 3.1.0
deepmerge-ts: 7.1.5
effect: 3.18.4
empathic: 2.0.0
transitivePeerDependencies:
- magicast
'@prisma/debug@6.19.2': {}
'@prisma/engines-version@7.1.1-3.c2990dca591cba766e3b7ef5d9e8a84796e47ab7': {}
'@prisma/engines@6.19.2':
dependencies:
'@prisma/debug': 6.19.2
'@prisma/engines-version': 7.1.1-3.c2990dca591cba766e3b7ef5d9e8a84796e47ab7
'@prisma/fetch-engine': 6.19.2
'@prisma/get-platform': 6.19.2
'@prisma/fetch-engine@6.19.2':
dependencies:
'@prisma/debug': 6.19.2
'@prisma/engines-version': 7.1.1-3.c2990dca591cba766e3b7ef5d9e8a84796e47ab7
'@prisma/get-platform': 6.19.2
'@prisma/get-platform@6.19.2':
dependencies:
'@prisma/debug': 6.19.2
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
'@protobufjs/aspromise@1.1.2': {}
'@protobufjs/base64@1.1.2': {}
'@protobufjs/codegen@2.0.4': {}
'@protobufjs/eventemitter@1.1.0': {}
'@protobufjs/fetch@1.1.0':
dependencies:
'@protobufjs/aspromise': 1.1.2
'@protobufjs/inquire': 1.1.0
'@protobufjs/float@1.0.2': {}
'@protobufjs/inquire@1.1.0': {}
'@protobufjs/path@1.1.2': {}
'@protobufjs/pool@1.1.0': {}
'@protobufjs/utf8@1.1.0': {}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
'@pulumi/pulumi@3.253.0(typescript@5.9.3)':
dependencies:
'@grpc/grpc-js': 1.14.3
'@logdna/tail-file': 2.2.0
'@npmcli/arborist': 9.9.0
'@opentelemetry/api': 1.9.1
'@opentelemetry/core': 2.9.0(@opentelemetry/api@1.9.1)
'@opentelemetry/exporter-trace-otlp-grpc': 0.220.0(@opentelemetry/api@1.9.1)
'@opentelemetry/exporter-zipkin': 2.9.0(@opentelemetry/api@1.9.1)
'@opentelemetry/instrumentation': 0.220.0(@opentelemetry/api@1.9.1)
'@opentelemetry/instrumentation-grpc': 0.220.0(@opentelemetry/api@1.9.1)
'@opentelemetry/resources': 2.9.0(@opentelemetry/api@1.9.1)
'@opentelemetry/sdk-trace-base': 2.9.0(@opentelemetry/api@1.9.1)
'@opentelemetry/sdk-trace-node': 2.9.0(@opentelemetry/api@1.9.1)
'@opentelemetry/semantic-conventions': 1.43.0
'@types/google-protobuf': 3.15.12
'@types/semver': 7.7.1
execa: 5.1.1
google-protobuf: 3.21.4
ini: 2.0.0
js-yaml: 4.1.1
minimist: 1.2.8
normalize-package-data: 6.0.2
require-from-string: 2.0.2
semver: 7.7.4
source-map-support: 0.5.21
upath: 1.2.0
optionalDependencies:
typescript: 5.9.3
transitivePeerDependencies:
- supports-color
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
'@rolldown/pluginutils@1.0.0-rc.3': {}
first commit
2026-02-21 03:10:39 +00:00
'@rollup/rollup-android-arm-eabi@4.58.0':
optional: true
'@rollup/rollup-android-arm64@4.58.0':
optional: true
'@rollup/rollup-darwin-arm64@4.58.0':
optional: true
'@rollup/rollup-darwin-x64@4.58.0':
optional: true
'@rollup/rollup-freebsd-arm64@4.58.0':
optional: true
'@rollup/rollup-freebsd-x64@4.58.0':
optional: true
'@rollup/rollup-linux-arm-gnueabihf@4.58.0':
optional: true
'@rollup/rollup-linux-arm-musleabihf@4.58.0':
optional: true
'@rollup/rollup-linux-arm64-gnu@4.58.0':
optional: true
'@rollup/rollup-linux-arm64-musl@4.58.0':
optional: true
'@rollup/rollup-linux-loong64-gnu@4.58.0':
optional: true
'@rollup/rollup-linux-loong64-musl@4.58.0':
optional: true
'@rollup/rollup-linux-ppc64-gnu@4.58.0':
optional: true
'@rollup/rollup-linux-ppc64-musl@4.58.0':
optional: true
'@rollup/rollup-linux-riscv64-gnu@4.58.0':
optional: true
'@rollup/rollup-linux-riscv64-musl@4.58.0':
optional: true
'@rollup/rollup-linux-s390x-gnu@4.58.0':
optional: true
'@rollup/rollup-linux-x64-gnu@4.58.0':
optional: true
'@rollup/rollup-linux-x64-musl@4.58.0':
optional: true
'@rollup/rollup-openbsd-x64@4.58.0':
optional: true
'@rollup/rollup-openharmony-arm64@4.58.0':
optional: true
'@rollup/rollup-win32-arm64-msvc@4.58.0':
optional: true
'@rollup/rollup-win32-ia32-msvc@4.58.0':
optional: true
'@rollup/rollup-win32-x64-gnu@4.58.0':
optional: true
'@rollup/rollup-win32-x64-msvc@4.58.0':
optional: true
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
'@sigstore/bundle@4.0.0':
dependencies:
'@sigstore/protobuf-specs': 0.5.1
'@sigstore/core@3.2.1': {}
'@sigstore/protobuf-specs@0.5.1': {}
'@sigstore/sign@4.1.1':
dependencies:
'@gar/promise-retry': 1.0.3
'@sigstore/bundle': 4.0.0
'@sigstore/core': 3.2.1
'@sigstore/protobuf-specs': 0.5.1
make-fetch-happen: 15.0.6
proc-log: 6.1.0
transitivePeerDependencies:
- supports-color
'@sigstore/tuf@4.0.2':
dependencies:
'@sigstore/protobuf-specs': 0.5.1
tuf-js: 4.1.0
transitivePeerDependencies:
- supports-color
'@sigstore/verify@3.1.1':
dependencies:
'@sigstore/bundle': 4.0.0
'@sigstore/core': 3.2.1
'@sigstore/protobuf-specs': 0.5.1
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
'@silvia-odwyer/photon-node@0.3.4': {}
'@smithy/core@3.31.1':
dependencies:
'@smithy/types': 4.16.1
tslib: 2.8.1
'@smithy/credential-provider-imds@4.4.16':
dependencies:
'@smithy/core': 3.31.1
'@smithy/types': 4.16.1
tslib: 2.8.1
'@smithy/fetch-http-handler@5.6.13':
dependencies:
'@smithy/core': 3.31.1
'@smithy/types': 4.16.1
tslib: 2.8.1
'@smithy/is-array-buffer@2.2.0':
dependencies:
tslib: 2.8.1
'@smithy/node-http-handler@4.7.3':
dependencies:
'@smithy/core': 3.31.1
'@smithy/types': 4.16.1
tslib: 2.8.1
'@smithy/node-http-handler@4.9.13':
dependencies:
'@smithy/core': 3.31.1
'@smithy/types': 4.16.1
tslib: 2.8.1
'@smithy/signature-v4@5.6.12':
dependencies:
'@smithy/core': 3.31.1
'@smithy/types': 4.16.1
tslib: 2.8.1
'@smithy/types@4.16.1':
dependencies:
tslib: 2.8.1
'@smithy/util-buffer-from@2.2.0':
dependencies:
'@smithy/is-array-buffer': 2.2.0
tslib: 2.8.1
'@smithy/util-utf8@2.3.0':
dependencies:
'@smithy/util-buffer-from': 2.2.0
tslib: 2.8.1
first commit
2026-02-21 03:10:39 +00:00
'@standard-schema/spec@1.1.0': {}
feat(web): bold redesign — Tailwind v4 + shadcn-style primitives + Skills/Proposals/Revisions UI Phase 6 of the Skills + Revisions + Proposals work. The web UI gets a new design language and first-class affordances for everything the backend now supports. ## Visual direction - Tailwind v4 with custom @theme block (oklch tokens). Dark-mode-only (internal tool — light mode doubles QA surface). - Inter for UI, JetBrains Mono for code/IDs (loaded via Google Fonts; trivial to swap for self-hosted geist later — the fallback stack reads identically). - Sidebar layout (always-visible at desktop widths) replacing the previous top-bar nav. Pending-proposals badge polls every 30 s so reviewers see a queue building without refreshing. - Lucide icons throughout. - Spacing and radii on Tailwind defaults. Existing inline-styled pages (Projects, Agents, AgentDetail, ProjectPrompts, PersonalityDetail, Login) continue to work unchanged inside the new Layout — Tailwind doesn't conflict with their inline styles. A follow-up can migrate them incrementally. ## What's added ### Build infra (src/web/) - package.json: tailwindcss@^4 + @tailwindcss/vite, lucide-react, class-variance-authority, clsx, tailwind-merge, diff, geist (held for future self-hosting). - vite.config.ts: registers the @tailwindcss/vite plugin. - src/index.css: Tailwind import + @theme tokens + @layer base. - src/main.tsx: imports index.css. - src/lib/utils.ts: shadcn-style cn() helper. ### shadcn-style primitives (src/components/ui/) Hand-written rather than generated via `npx shadcn` so the repo doesn't depend on a CLI tool that needs interactive runtime: - button.tsx — variants: primary / secondary / ghost / danger / link; sizes: sm / md / lg / icon. - card.tsx — Card + Header/Title/Description/Content/Footer subparts. - badge.tsx — variants: default / info / success / warning / danger / outline. - input.tsx — Input + Textarea + Label. - tabs.tsx — no-dep accessible Tabs (no Radix needed for our use). - separator.tsx — h/v separator with role=separator. ### Diff component (src/components/Diff.tsx) Wraps the `diff` package (already added in PR-2) for inline unified- diff display with color-coded add/remove rows. Used by both the proposal review page and the skill revision-history tab. ### New pages (src/pages/) - Dashboard.tsx — at-a-glance home. Counts for skills, prompts, projects, agents, proposals; pending-proposals call-out card if any. - Skills.tsx — list view, separated into Global vs Project/Agent- scoped sections. - SkillDetail.tsx — name + semver + description; tabs for SKILL.md / Files / Metadata / History. History tab shows revisions with click-to-diff against the live body. - Proposals.tsx — queue with Pending/Approved/Rejected tabs. Pending count is highlighted in amber. - ProposalDetail.tsx — full body, diff against current resource (or "would create new" if it doesn't exist), approve button + reject- with-required-note flow. ### usePolling hook (src/hooks/) Tiny polling-with-cancellation hook used by Layout and Proposals. ### Layout rewrite (src/components/Layout.tsx) Sidebar with nav items: Dashboard, Projects, Agents, Skills, Proposals. Lucide icons. Active-route highlighting via NavLink. Pending-proposals warning badge on the Proposals item. ### Routes (src/App.tsx) New routes: /dashboard, /skills, /skills/:name, /proposals, /proposals/:id. Default redirects to /dashboard. ### API types (src/api.ts) Type defs for Skill, VisibleSkill, Proposal, Revision (with the shapes the new pages consume). ## Tests Existing 7 web tests still pass (Login + api). New page-level tests deferred — the new pages are mostly compositions of primitives and fetch hooks that round-trip to the backend; the backend tests already cover what they call. PR-7 polish can add render-and-click tests if coverage drift surfaces. ## Verification - `pnpm --filter @mcpctl/web build` clean, no warnings. - `pnpm test:run` whole monorepo: 162 test files / 2157 tests green. - Visual smoke deferred — needs a running mcpd to populate the fixtures. Manual smoke tested locally is the next step. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-07 17:54:55 +01:00
'@swc/helpers@0.5.15':
dependencies:
tslib: 2.8.1
'@tailwindcss/node@4.2.4':
dependencies:
'@jridgewell/remapping': 2.3.5
enhanced-resolve: 5.21.0
jiti: 2.6.1
lightningcss: 1.32.0
magic-string: 0.30.21
source-map-js: 1.2.1
tailwindcss: 4.2.4
'@tailwindcss/oxide-android-arm64@4.2.4':
optional: true
'@tailwindcss/oxide-darwin-arm64@4.2.4':
optional: true
'@tailwindcss/oxide-darwin-x64@4.2.4':
optional: true
'@tailwindcss/oxide-freebsd-x64@4.2.4':
optional: true
'@tailwindcss/oxide-linux-arm-gnueabihf@4.2.4':
optional: true
'@tailwindcss/oxide-linux-arm64-gnu@4.2.4':
optional: true
'@tailwindcss/oxide-linux-arm64-musl@4.2.4':
optional: true
'@tailwindcss/oxide-linux-x64-gnu@4.2.4':
optional: true
'@tailwindcss/oxide-linux-x64-musl@4.2.4':
optional: true
'@tailwindcss/oxide-wasm32-wasi@4.2.4':
optional: true
'@tailwindcss/oxide-win32-arm64-msvc@4.2.4':
optional: true
'@tailwindcss/oxide-win32-x64-msvc@4.2.4':
optional: true
'@tailwindcss/oxide@4.2.4':
optionalDependencies:
'@tailwindcss/oxide-android-arm64': 4.2.4
'@tailwindcss/oxide-darwin-arm64': 4.2.4
'@tailwindcss/oxide-darwin-x64': 4.2.4
'@tailwindcss/oxide-freebsd-x64': 4.2.4
'@tailwindcss/oxide-linux-arm-gnueabihf': 4.2.4
'@tailwindcss/oxide-linux-arm64-gnu': 4.2.4
'@tailwindcss/oxide-linux-arm64-musl': 4.2.4
'@tailwindcss/oxide-linux-x64-gnu': 4.2.4
'@tailwindcss/oxide-linux-x64-musl': 4.2.4
'@tailwindcss/oxide-wasm32-wasi': 4.2.4
'@tailwindcss/oxide-win32-arm64-msvc': 4.2.4
'@tailwindcss/oxide-win32-x64-msvc': 4.2.4
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
'@tailwindcss/vite@4.2.4(vite@7.3.1(@types/node@25.3.0)(jiti@2.7.0)(lightningcss@1.32.0)(tsx@4.21.0)(yaml@2.9.0))':
feat(web): bold redesign — Tailwind v4 + shadcn-style primitives + Skills/Proposals/Revisions UI Phase 6 of the Skills + Revisions + Proposals work. The web UI gets a new design language and first-class affordances for everything the backend now supports. ## Visual direction - Tailwind v4 with custom @theme block (oklch tokens). Dark-mode-only (internal tool — light mode doubles QA surface). - Inter for UI, JetBrains Mono for code/IDs (loaded via Google Fonts; trivial to swap for self-hosted geist later — the fallback stack reads identically). - Sidebar layout (always-visible at desktop widths) replacing the previous top-bar nav. Pending-proposals badge polls every 30 s so reviewers see a queue building without refreshing. - Lucide icons throughout. - Spacing and radii on Tailwind defaults. Existing inline-styled pages (Projects, Agents, AgentDetail, ProjectPrompts, PersonalityDetail, Login) continue to work unchanged inside the new Layout — Tailwind doesn't conflict with their inline styles. A follow-up can migrate them incrementally. ## What's added ### Build infra (src/web/) - package.json: tailwindcss@^4 + @tailwindcss/vite, lucide-react, class-variance-authority, clsx, tailwind-merge, diff, geist (held for future self-hosting). - vite.config.ts: registers the @tailwindcss/vite plugin. - src/index.css: Tailwind import + @theme tokens + @layer base. - src/main.tsx: imports index.css. - src/lib/utils.ts: shadcn-style cn() helper. ### shadcn-style primitives (src/components/ui/) Hand-written rather than generated via `npx shadcn` so the repo doesn't depend on a CLI tool that needs interactive runtime: - button.tsx — variants: primary / secondary / ghost / danger / link; sizes: sm / md / lg / icon. - card.tsx — Card + Header/Title/Description/Content/Footer subparts. - badge.tsx — variants: default / info / success / warning / danger / outline. - input.tsx — Input + Textarea + Label. - tabs.tsx — no-dep accessible Tabs (no Radix needed for our use). - separator.tsx — h/v separator with role=separator. ### Diff component (src/components/Diff.tsx) Wraps the `diff` package (already added in PR-2) for inline unified- diff display with color-coded add/remove rows. Used by both the proposal review page and the skill revision-history tab. ### New pages (src/pages/) - Dashboard.tsx — at-a-glance home. Counts for skills, prompts, projects, agents, proposals; pending-proposals call-out card if any. - Skills.tsx — list view, separated into Global vs Project/Agent- scoped sections. - SkillDetail.tsx — name + semver + description; tabs for SKILL.md / Files / Metadata / History. History tab shows revisions with click-to-diff against the live body. - Proposals.tsx — queue with Pending/Approved/Rejected tabs. Pending count is highlighted in amber. - ProposalDetail.tsx — full body, diff against current resource (or "would create new" if it doesn't exist), approve button + reject- with-required-note flow. ### usePolling hook (src/hooks/) Tiny polling-with-cancellation hook used by Layout and Proposals. ### Layout rewrite (src/components/Layout.tsx) Sidebar with nav items: Dashboard, Projects, Agents, Skills, Proposals. Lucide icons. Active-route highlighting via NavLink. Pending-proposals warning badge on the Proposals item. ### Routes (src/App.tsx) New routes: /dashboard, /skills, /skills/:name, /proposals, /proposals/:id. Default redirects to /dashboard. ### API types (src/api.ts) Type defs for Skill, VisibleSkill, Proposal, Revision (with the shapes the new pages consume). ## Tests Existing 7 web tests still pass (Login + api). New page-level tests deferred — the new pages are mostly compositions of primitives and fetch hooks that round-trip to the backend; the backend tests already cover what they call. PR-7 polish can add render-and-click tests if coverage drift surfaces. ## Verification - `pnpm --filter @mcpctl/web build` clean, no warnings. - `pnpm test:run` whole monorepo: 162 test files / 2157 tests green. - Visual smoke deferred — needs a running mcpd to populate the fixtures. Manual smoke tested locally is the next step. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-07 17:54:55 +01:00
dependencies:
'@tailwindcss/node': 4.2.4
'@tailwindcss/oxide': 4.2.4
tailwindcss: 4.2.4
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
vite: 7.3.1(@types/node@25.3.0)(jiti@2.7.0)(lightningcss@1.32.0)(tsx@4.21.0)(yaml@2.9.0)
feat(web): bold redesign — Tailwind v4 + shadcn-style primitives + Skills/Proposals/Revisions UI Phase 6 of the Skills + Revisions + Proposals work. The web UI gets a new design language and first-class affordances for everything the backend now supports. ## Visual direction - Tailwind v4 with custom @theme block (oklch tokens). Dark-mode-only (internal tool — light mode doubles QA surface). - Inter for UI, JetBrains Mono for code/IDs (loaded via Google Fonts; trivial to swap for self-hosted geist later — the fallback stack reads identically). - Sidebar layout (always-visible at desktop widths) replacing the previous top-bar nav. Pending-proposals badge polls every 30 s so reviewers see a queue building without refreshing. - Lucide icons throughout. - Spacing and radii on Tailwind defaults. Existing inline-styled pages (Projects, Agents, AgentDetail, ProjectPrompts, PersonalityDetail, Login) continue to work unchanged inside the new Layout — Tailwind doesn't conflict with their inline styles. A follow-up can migrate them incrementally. ## What's added ### Build infra (src/web/) - package.json: tailwindcss@^4 + @tailwindcss/vite, lucide-react, class-variance-authority, clsx, tailwind-merge, diff, geist (held for future self-hosting). - vite.config.ts: registers the @tailwindcss/vite plugin. - src/index.css: Tailwind import + @theme tokens + @layer base. - src/main.tsx: imports index.css. - src/lib/utils.ts: shadcn-style cn() helper. ### shadcn-style primitives (src/components/ui/) Hand-written rather than generated via `npx shadcn` so the repo doesn't depend on a CLI tool that needs interactive runtime: - button.tsx — variants: primary / secondary / ghost / danger / link; sizes: sm / md / lg / icon. - card.tsx — Card + Header/Title/Description/Content/Footer subparts. - badge.tsx — variants: default / info / success / warning / danger / outline. - input.tsx — Input + Textarea + Label. - tabs.tsx — no-dep accessible Tabs (no Radix needed for our use). - separator.tsx — h/v separator with role=separator. ### Diff component (src/components/Diff.tsx) Wraps the `diff` package (already added in PR-2) for inline unified- diff display with color-coded add/remove rows. Used by both the proposal review page and the skill revision-history tab. ### New pages (src/pages/) - Dashboard.tsx — at-a-glance home. Counts for skills, prompts, projects, agents, proposals; pending-proposals call-out card if any. - Skills.tsx — list view, separated into Global vs Project/Agent- scoped sections. - SkillDetail.tsx — name + semver + description; tabs for SKILL.md / Files / Metadata / History. History tab shows revisions with click-to-diff against the live body. - Proposals.tsx — queue with Pending/Approved/Rejected tabs. Pending count is highlighted in amber. - ProposalDetail.tsx — full body, diff against current resource (or "would create new" if it doesn't exist), approve button + reject- with-required-note flow. ### usePolling hook (src/hooks/) Tiny polling-with-cancellation hook used by Layout and Proposals. ### Layout rewrite (src/components/Layout.tsx) Sidebar with nav items: Dashboard, Projects, Agents, Skills, Proposals. Lucide icons. Active-route highlighting via NavLink. Pending-proposals warning badge on the Proposals item. ### Routes (src/App.tsx) New routes: /dashboard, /skills, /skills/:name, /proposals, /proposals/:id. Default redirects to /dashboard. ### API types (src/api.ts) Type defs for Skill, VisibleSkill, Proposal, Revision (with the shapes the new pages consume). ## Tests Existing 7 web tests still pass (Login + api). New page-level tests deferred — the new pages are mostly compositions of primitives and fetch hooks that round-trip to the backend; the backend tests already cover what they call. PR-7 polish can add render-and-click tests if coverage drift surfaces. ## Verification - `pnpm --filter @mcpctl/web build` clean, no warnings. - `pnpm test:run` whole monorepo: 162 test files / 2157 tests green. - Visual smoke deferred — needs a running mcpd to populate the fixtures. Manual smoke tested locally is the next step. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-07 17:54:55 +01:00
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
'@testing-library/dom@10.4.1':
dependencies:
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
'@babel/code-frame': 7.29.7
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
'@babel/runtime': 7.29.2
'@types/aria-query': 5.0.4
aria-query: 5.3.0
dom-accessibility-api: 0.5.16
lz-string: 1.5.0
picocolors: 1.1.1
pretty-format: 27.5.1
'@testing-library/jest-dom@6.9.1':
dependencies:
'@adobe/css-tools': 4.4.4
aria-query: 5.3.2
css.escape: 1.5.1
dom-accessibility-api: 0.6.3
picocolors: 1.1.1
redent: 3.0.0
'@testing-library/react@16.3.2(@testing-library/dom@10.4.1)(@types/react-dom@19.2.3(@types/react@19.2.14))(@types/react@19.2.14)(react-dom@19.2.5(react@19.2.5))(react@19.2.5)':
dependencies:
'@babel/runtime': 7.29.2
'@testing-library/dom': 10.4.1
react: 19.2.5
react-dom: 19.2.5(react@19.2.5)
optionalDependencies:
'@types/react': 19.2.14
'@types/react-dom': 19.2.3(@types/react@19.2.14)
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
'@tufjs/canonical-json@2.0.0': {}
'@tufjs/models@4.1.0':
dependencies:
'@tufjs/canonical-json': 2.0.0
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
minimatch: 10.2.5
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
'@types/aria-query@5.0.4': {}
'@types/babel__core@7.20.5':
dependencies:
'@babel/parser': 7.29.0
'@babel/types': 7.29.0
'@types/babel__generator': 7.27.0
'@types/babel__template': 7.4.4
'@types/babel__traverse': 7.28.0
'@types/babel__generator@7.27.0':
dependencies:
'@babel/types': 7.29.0
'@types/babel__template@7.4.4':
dependencies:
'@babel/parser': 7.29.0
'@babel/types': 7.29.0
'@types/babel__traverse@7.28.0':
dependencies:
'@babel/types': 7.29.0
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
'@types/bcrypt@5.0.2':
dependencies:
'@types/node': 25.3.0
first commit
2026-02-21 03:10:39 +00:00
'@types/chai@5.2.3':
dependencies:
'@types/deep-eql': 4.0.2
assertion-error: 2.0.1
'@types/deep-eql@4.0.2': {}
feat(mcpd): ResourceRevision + ResourceProposal services + Prompt revision integration Phase 2 of the Skills + Revisions + Proposals work. Stands up the generic revision/proposal layer and wires Prompt into it. Skills will plug into the same infrastructure in PR-3 with no further service changes required. This PR is intentionally additive: PromptRequest table and routes are unchanged. The /api/v1/proposals API runs side-by-side with the legacy /api/v1/promptrequests API. The PromptRequest cutover (rename + backfill + mcplocal rewire) is deferred to a later PR so this one stays reviewable. ## What's added ### Repositories (src/mcpd/src/repositories/) - resource-revision.repository.ts — append-only revision log keyed by (resourceType, resourceId). Soft FK; no relations declared. Supports history listing, semver lookup, and contentHash cross-resource search. - resource-proposal.repository.ts — generic propose queue. Status lifecycle pending → approved | rejected. Mirrors Prompt's `?? ''` workaround for nullable-FK compound lookups. ### Services (src/mcpd/src/services/) - resource-revision.service.ts — record() inserts a revision with a stable sha256 contentHash computed from canonicalised JSON (key-sorted at every level so reordered objects produce the same hash). Caller passes a pre-computed semver; service does NOT decide bump policy. - resource-proposal.service.ts — propose / approve / reject / list, with a per-resourceType handler registry. PromptService registers the 'prompt' handler at construction; the SkillService will register 'skill' in PR-3. approve() runs in a Prisma $transaction so the resource update + revision insert + proposal status flip are atomic. ### Pure utility (src/mcpd/src/utils/semver.ts) - bumpSemver(current, kind) for major / minor / patch - compareSemver(a, b) — numeric, not lex (10 > 9) - isValidSemver(s) - Invalid input falls back to '0.1.0' rather than throwing — keeps the audit-write path from blowing up the prompt update if a row's semver ever drifts out of MAJOR.MINOR.PATCH shape. ### Routes (src/mcpd/src/routes/) - revisions.ts — GET /api/v1/revisions?resourceType=&resourceId=, GET /api/v1/revisions/:id, GET /api/v1/revisions/:id/diff?against=<id|live> (unified-format diff via the `diff` package), and POST /api/v1/prompts/:id/restore-revision { revisionId, note? }. - proposals.ts — GET / POST /api/v1/proposals, GET /api/v1/proposals/:id, PUT for body updates, POST .../approve and POST .../reject, plus DELETE. ## What's changed - PromptService.create / update now record a ResourceRevision when the revision service is wired. Update auto-bumps patch on content change; authors can override via `--bump major|minor|patch` or `--semver X.Y.Z` on the CLI (forwarded into the PUT body). Best-effort: revision write failures are swallowed so the prompt save still succeeds (revision is audit, not source of truth). - PromptService.setProposalService registers a 'prompt' approval handler with the proposal service. Approval runs in a Prisma transaction: upsert prompt → record revision → update currentRevisionId → flip proposal status. semver bumps to 0.1.0 on first approval, patch thereafter. - New CLI flags on `mcpctl edit prompt`: --bump, --semver, --note. They're prompt-only (validated client-side); other resources reject them. - Aliases in shared.ts: `proposal`/`prop` → proposals, `revision`/`rev` → revisions. - diff dependency added to mcpd. ## Tests - src/mcpd/tests/utils/semver.test.ts — covers bump/compare/validate including numeric (not lex) semver compare and invalid-input fallback. - prompt-service.test.ts updated: makePrompt fixture now sets semver + agentId + currentRevisionId; updatePrompt assertion expects the auto-bumped patch in the same update call. - prompt-routes.test.ts updated symmetrically. ## RBAC `proposals` and `revisions` URL segments map to the existing `prompts` permission for now. PR-7 may split if a "reviewer" role becomes useful. ## Verification Full suite: 158 test files / 2127 tests green. `pnpm build` clean across all 6 workspace packages. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-07 00:38:35 +01:00
'@types/diff@5.2.3': {}
feat: eager vLLM warmup and smart page titles in paginate stage - Add warmup() to LlmProvider interface for eager subprocess startup - ManagedVllmProvider.warmup() starts vLLM in background on project load - ProviderRegistry.warmupAll() triggers all managed providers - NamedProvider proxies warmup() to inner provider - paginate stage generates LLM-powered descriptive page titles when available, cached by content hash, falls back to generic "Page N" - project-mcp-endpoint calls warmupAll() on router creation so vLLM is loading while the session initializes Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-03-03 19:07:39 +00:00
'@types/diff@8.0.0':
dependencies:
diff: 8.0.3
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
'@types/docker-modem@3.0.6':
dependencies:
'@types/node': 25.3.0
'@types/ssh2': 1.15.5
'@types/dockerode@4.0.1':
dependencies:
'@types/docker-modem': 3.0.6
'@types/node': 25.3.0
'@types/ssh2': 1.15.5
first commit
2026-02-21 03:10:39 +00:00
'@types/esrecurse@4.3.1': {}
'@types/estree@1.0.8': {}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
'@types/google-protobuf@3.15.12': {}
feat: build CLI core framework with Commander.js Add CLI entry point with Commander.js, config management (~/.mcpctl/config.json with Zod validation), output formatters (table/json/yaml), config and status commands with dependency injection for testing. Fix sanitizeString regex ordering. 67 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:17:31 +00:00
'@types/js-yaml@4.0.9': {}
first commit
2026-02-21 03:10:39 +00:00
'@types/json-schema@7.0.15': {}
feat: add Kubernetes orchestrator for MCP server pod management mcpd can now deploy MCP server instances as Kubernetes pods instead of Docker containers. Set MCPD_ORCHESTRATOR=kubernetes to enable. - Add @kubernetes/client-node with thin wrapper (context enforcement via MCPD_K8S_CONTEXT to prevent multi-cluster mishaps) - Rewrite KubernetesOrchestrator: pod CRUD, pod IP extraction, exec via SPDY (one-shot + interactive), log streaming - Manifest generator: stdin:true for STDIO servers, args (not command) to preserve runner image entrypoint, security hardening - Orchestrator selection in main.ts via MCPD_ORCHESTRATOR env var - 25 unit tests for k8s orchestrator, all 624 tests pass Tested end-to-end on local k3s: - mcpd deployed via Pulumi, creates pods in mcpctl-servers namespace - NetworkPolicy verified: only mcpd can reach MCP server pods - Python runner (uvx) successfully runs aws-documentation-mcp-server Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-08 01:55:13 +01:00
'@types/node-fetch@2.6.13':
dependencies:
'@types/node': 25.3.0
form-data: 4.0.5
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
'@types/node@18.19.130':
dependencies:
undici-types: 5.26.5
feat: add Kubernetes orchestrator for MCP server pod management mcpd can now deploy MCP server instances as Kubernetes pods instead of Docker containers. Set MCPD_ORCHESTRATOR=kubernetes to enable. - Add @kubernetes/client-node with thin wrapper (context enforcement via MCPD_K8S_CONTEXT to prevent multi-cluster mishaps) - Rewrite KubernetesOrchestrator: pod CRUD, pod IP extraction, exec via SPDY (one-shot + interactive), log streaming - Manifest generator: stdin:true for STDIO servers, args (not command) to preserve runner image entrypoint, security hardening - Orchestrator selection in main.ts via MCPD_ORCHESTRATOR env var - 25 unit tests for k8s orchestrator, all 624 tests pass Tested end-to-end on local k3s: - mcpd deployed via Pulumi, creates pods in mcpctl-servers namespace - NetworkPolicy verified: only mcpd can reach MCP server pods - Python runner (uvx) successfully runs aws-documentation-mcp-server Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-08 01:55:13 +01:00
'@types/node@24.12.2':
dependencies:
undici-types: 7.16.0
feat: implement MCP registry client with multi-source search Add registry client that queries Official, Glama, and Smithery MCP registries with caching, request deduplication, retry logic, and result ranking/dedup. Includes 53 tests covering all components. Also fix null priority values in cancelled tasks (19-21) that broke Task Master, and add new tasks 25-27 for registry completion and CLI discover/install commands. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 03:46:14 +00:00
'@types/node@25.3.0':
dependencies:
undici-types: 7.18.2
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
'@types/react-dom@19.2.3(@types/react@19.2.14)':
dependencies:
'@types/react': 19.2.14
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
'@types/react@19.2.14':
dependencies:
csstype: 3.2.3
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
'@types/retry@0.12.0': {}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
'@types/semver@7.7.1': {}
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
'@types/ssh2@1.15.5':
dependencies:
'@types/node': 18.19.130
feat: add Kubernetes orchestrator for MCP server pod management mcpd can now deploy MCP server instances as Kubernetes pods instead of Docker containers. Set MCPD_ORCHESTRATOR=kubernetes to enable. - Add @kubernetes/client-node with thin wrapper (context enforcement via MCPD_K8S_CONTEXT to prevent multi-cluster mishaps) - Rewrite KubernetesOrchestrator: pod CRUD, pod IP extraction, exec via SPDY (one-shot + interactive), log streaming - Manifest generator: stdin:true for STDIO servers, args (not command) to preserve runner image entrypoint, security hardening - Orchestrator selection in main.ts via MCPD_ORCHESTRATOR env var - 25 unit tests for k8s orchestrator, all 624 tests pass Tested end-to-end on local k3s: - mcpd deployed via Pulumi, creates pods in mcpctl-servers namespace - NetworkPolicy verified: only mcpd can reach MCP server pods - Python runner (uvx) successfully runs aws-documentation-mcp-server Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-08 01:55:13 +01:00
'@types/stream-buffers@3.0.8':
dependencies:
'@types/node': 25.3.0
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
'@types/trusted-types@2.0.7':
optional: true
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
'@typescript-eslint/eslint-plugin@8.56.0(@typescript-eslint/parser@8.56.0(eslint@10.0.1(jiti@2.7.0))(typescript@5.9.3))(eslint@10.0.1(jiti@2.7.0))(typescript@5.9.3)':
first commit
2026-02-21 03:10:39 +00:00
dependencies:
'@eslint-community/regexpp': 4.12.2
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
'@typescript-eslint/parser': 8.56.0(eslint@10.0.1(jiti@2.7.0))(typescript@5.9.3)
first commit
2026-02-21 03:10:39 +00:00
'@typescript-eslint/scope-manager': 8.56.0
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
'@typescript-eslint/type-utils': 8.56.0(eslint@10.0.1(jiti@2.7.0))(typescript@5.9.3)
'@typescript-eslint/utils': 8.56.0(eslint@10.0.1(jiti@2.7.0))(typescript@5.9.3)
first commit
2026-02-21 03:10:39 +00:00
'@typescript-eslint/visitor-keys': 8.56.0
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
eslint: 10.0.1(jiti@2.7.0)
first commit
2026-02-21 03:10:39 +00:00
ignore: 7.0.5
natural-compare: 1.4.0
ts-api-utils: 2.4.0(typescript@5.9.3)
typescript: 5.9.3
transitivePeerDependencies:
- supports-color
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
'@typescript-eslint/parser@8.56.0(eslint@10.0.1(jiti@2.7.0))(typescript@5.9.3)':
first commit
2026-02-21 03:10:39 +00:00
dependencies:
'@typescript-eslint/scope-manager': 8.56.0
'@typescript-eslint/types': 8.56.0
'@typescript-eslint/typescript-estree': 8.56.0(typescript@5.9.3)
'@typescript-eslint/visitor-keys': 8.56.0
debug: 4.4.3
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
eslint: 10.0.1(jiti@2.7.0)
first commit
2026-02-21 03:10:39 +00:00
typescript: 5.9.3
transitivePeerDependencies:
- supports-color
'@typescript-eslint/project-service@8.56.0(typescript@5.9.3)':
dependencies:
'@typescript-eslint/tsconfig-utils': 8.56.0(typescript@5.9.3)
'@typescript-eslint/types': 8.56.0
debug: 4.4.3
typescript: 5.9.3
transitivePeerDependencies:
- supports-color
'@typescript-eslint/scope-manager@8.56.0':
dependencies:
'@typescript-eslint/types': 8.56.0
'@typescript-eslint/visitor-keys': 8.56.0
'@typescript-eslint/tsconfig-utils@8.56.0(typescript@5.9.3)':
dependencies:
typescript: 5.9.3
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
'@typescript-eslint/type-utils@8.56.0(eslint@10.0.1(jiti@2.7.0))(typescript@5.9.3)':
first commit
2026-02-21 03:10:39 +00:00
dependencies:
'@typescript-eslint/types': 8.56.0
'@typescript-eslint/typescript-estree': 8.56.0(typescript@5.9.3)
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
'@typescript-eslint/utils': 8.56.0(eslint@10.0.1(jiti@2.7.0))(typescript@5.9.3)
first commit
2026-02-21 03:10:39 +00:00
debug: 4.4.3
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
eslint: 10.0.1(jiti@2.7.0)
first commit
2026-02-21 03:10:39 +00:00
ts-api-utils: 2.4.0(typescript@5.9.3)
typescript: 5.9.3
transitivePeerDependencies:
- supports-color
'@typescript-eslint/types@8.56.0': {}
'@typescript-eslint/typescript-estree@8.56.0(typescript@5.9.3)':
dependencies:
'@typescript-eslint/project-service': 8.56.0(typescript@5.9.3)
'@typescript-eslint/tsconfig-utils': 8.56.0(typescript@5.9.3)
'@typescript-eslint/types': 8.56.0
'@typescript-eslint/visitor-keys': 8.56.0
debug: 4.4.3
minimatch: 9.0.5
semver: 7.7.4
tinyglobby: 0.2.15
ts-api-utils: 2.4.0(typescript@5.9.3)
typescript: 5.9.3
transitivePeerDependencies:
- supports-color
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
'@typescript-eslint/utils@8.56.0(eslint@10.0.1(jiti@2.7.0))(typescript@5.9.3)':
first commit
2026-02-21 03:10:39 +00:00
dependencies:
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
'@eslint-community/eslint-utils': 4.9.1(eslint@10.0.1(jiti@2.7.0))
first commit
2026-02-21 03:10:39 +00:00
'@typescript-eslint/scope-manager': 8.56.0
'@typescript-eslint/types': 8.56.0
'@typescript-eslint/typescript-estree': 8.56.0(typescript@5.9.3)
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
eslint: 10.0.1(jiti@2.7.0)
first commit
2026-02-21 03:10:39 +00:00
typescript: 5.9.3
transitivePeerDependencies:
- supports-color
'@typescript-eslint/visitor-keys@8.56.0':
dependencies:
'@typescript-eslint/types': 8.56.0
eslint-visitor-keys: 5.0.1
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
'@vitejs/plugin-react@5.2.0(vite@7.3.1(@types/node@25.3.0)(jiti@2.7.0)(lightningcss@1.32.0)(tsx@4.21.0)(yaml@2.9.0))':
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
dependencies:
'@babel/core': 7.29.0
'@babel/plugin-transform-react-jsx-self': 7.27.1(@babel/core@7.29.0)
'@babel/plugin-transform-react-jsx-source': 7.27.1(@babel/core@7.29.0)
'@rolldown/pluginutils': 1.0.0-rc.3
'@types/babel__core': 7.20.5
react-refresh: 0.18.0
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
vite: 7.3.1(@types/node@25.3.0)(jiti@2.7.0)(lightningcss@1.32.0)(tsx@4.21.0)(yaml@2.9.0)
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
transitivePeerDependencies:
- supports-color
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
'@vitest/coverage-v8@4.0.18(vitest@4.0.18(@opentelemetry/api@1.9.0)(@types/node@25.3.0)(jiti@2.7.0)(jsdom@28.1.0)(lightningcss@1.32.0)(tsx@4.21.0)(yaml@2.9.0))':
first commit
2026-02-21 03:10:39 +00:00
dependencies:
'@bcoe/v8-coverage': 1.0.2
'@vitest/utils': 4.0.18
ast-v8-to-istanbul: 0.3.11
istanbul-lib-coverage: 3.2.2
istanbul-lib-report: 3.0.1
istanbul-reports: 3.2.0
magicast: 0.5.2
obug: 2.1.1
std-env: 3.10.0
tinyrainbow: 3.0.3
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
vitest: 4.0.18(@opentelemetry/api@1.9.0)(@types/node@25.3.0)(jiti@2.7.0)(jsdom@28.1.0)(lightningcss@1.32.0)(tsx@4.21.0)(yaml@2.9.0)
first commit
2026-02-21 03:10:39 +00:00
'@vitest/expect@4.0.18':
dependencies:
'@standard-schema/spec': 1.1.0
'@types/chai': 5.2.3
'@vitest/spy': 4.0.18
'@vitest/utils': 4.0.18
chai: 6.2.2
tinyrainbow: 3.0.3
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
'@vitest/mocker@4.0.18(vite@7.3.1(@types/node@25.3.0)(jiti@2.7.0)(lightningcss@1.32.0)(tsx@4.21.0)(yaml@2.9.0))':
first commit
2026-02-21 03:10:39 +00:00
dependencies:
'@vitest/spy': 4.0.18
estree-walker: 3.0.3
magic-string: 0.30.21
optionalDependencies:
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
vite: 7.3.1(@types/node@25.3.0)(jiti@2.7.0)(lightningcss@1.32.0)(tsx@4.21.0)(yaml@2.9.0)
first commit
2026-02-21 03:10:39 +00:00
'@vitest/pretty-format@4.0.18':
dependencies:
tinyrainbow: 3.0.3
'@vitest/runner@4.0.18':
dependencies:
'@vitest/utils': 4.0.18
pathe: 2.0.3
'@vitest/snapshot@4.0.18':
dependencies:
'@vitest/pretty-format': 4.0.18
magic-string: 0.30.21
pathe: 2.0.3
'@vitest/spy@4.0.18': {}
'@vitest/utils@4.0.18':
dependencies:
'@vitest/pretty-format': 4.0.18
tinyrainbow: 3.0.3
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
abbrev@1.1.1: {}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
abbrev@4.0.0: {}
first commit
2026-02-21 03:10:39 +00:00
abstract-logging@2.0.1: {}
accepts@2.0.0:
dependencies:
mime-types: 3.0.2
negotiator: 1.0.0
acorn-jsx@5.3.2(acorn@8.16.0):
dependencies:
acorn: 8.16.0
acorn@8.16.0: {}
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
agent-base@6.0.2:
dependencies:
debug: 4.4.3
transitivePeerDependencies:
- supports-color
feat: add Kubernetes orchestrator for MCP server pod management mcpd can now deploy MCP server instances as Kubernetes pods instead of Docker containers. Set MCPD_ORCHESTRATOR=kubernetes to enable. - Add @kubernetes/client-node with thin wrapper (context enforcement via MCPD_K8S_CONTEXT to prevent multi-cluster mishaps) - Rewrite KubernetesOrchestrator: pod CRUD, pod IP extraction, exec via SPDY (one-shot + interactive), log streaming - Manifest generator: stdin:true for STDIO servers, args (not command) to preserve runner image entrypoint, security hardening - Orchestrator selection in main.ts via MCPD_ORCHESTRATOR env var - 25 unit tests for k8s orchestrator, all 624 tests pass Tested end-to-end on local k3s: - mcpd deployed via Pulumi, creates pods in mcpctl-servers namespace - NetworkPolicy verified: only mcpd can reach MCP server pods - Python runner (uvx) successfully runs aws-documentation-mcp-server Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-08 01:55:13 +01:00
agent-base@7.1.4: {}
first commit
2026-02-21 03:10:39 +00:00
ajv-formats@3.0.1(ajv@8.18.0):
optionalDependencies:
ajv: 8.18.0
ajv@6.14.0:
dependencies:
fast-deep-equal: 3.1.3
fast-json-stable-stringify: 2.1.0
json-schema-traverse: 0.4.1
uri-js: 4.4.1
ajv@8.18.0:
dependencies:
fast-deep-equal: 3.1.3
fast-uri: 3.1.0
json-schema-traverse: 1.0.0
require-from-string: 2.0.2
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
ansi-escapes@7.3.0:
dependencies:
environment: 1.1.0
first commit
2026-02-21 03:10:39 +00:00
ansi-regex@5.0.1: {}
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
ansi-regex@6.2.2: {}
first commit
2026-02-21 03:10:39 +00:00
ansi-styles@4.3.0:
dependencies:
color-convert: 2.0.1
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
ansi-styles@5.2.0: {}
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
ansi-styles@6.2.3: {}
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
aproba@2.1.0: {}
are-we-there-yet@2.0.0:
dependencies:
delegates: 1.0.0
readable-stream: 3.6.2
first commit
2026-02-21 03:10:39 +00:00
argparse@2.0.1: {}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
aria-query@5.3.0:
dependencies:
dequal: 2.0.3
aria-query@5.3.2: {}
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
asn1@0.2.6:
dependencies:
safer-buffer: 2.1.2
first commit
2026-02-21 03:10:39 +00:00
assertion-error@2.0.1: {}
ast-v8-to-istanbul@0.3.11:
dependencies:
'@jridgewell/trace-mapping': 0.3.31
estree-walker: 3.0.3
js-tokens: 10.0.0
feat: add Kubernetes orchestrator for MCP server pod management mcpd can now deploy MCP server instances as Kubernetes pods instead of Docker containers. Set MCPD_ORCHESTRATOR=kubernetes to enable. - Add @kubernetes/client-node with thin wrapper (context enforcement via MCPD_K8S_CONTEXT to prevent multi-cluster mishaps) - Rewrite KubernetesOrchestrator: pod CRUD, pod IP extraction, exec via SPDY (one-shot + interactive), log streaming - Manifest generator: stdin:true for STDIO servers, args (not command) to preserve runner image entrypoint, security hardening - Orchestrator selection in main.ts via MCPD_ORCHESTRATOR env var - 25 unit tests for k8s orchestrator, all 624 tests pass Tested end-to-end on local k3s: - mcpd deployed via Pulumi, creates pods in mcpctl-servers namespace - NetworkPolicy verified: only mcpd can reach MCP server pods - Python runner (uvx) successfully runs aws-documentation-mcp-server Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-08 01:55:13 +01:00
asynckit@0.4.0: {}
first commit
2026-02-21 03:10:39 +00:00
atomic-sleep@1.0.0: {}
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
auto-bind@5.0.1: {}
first commit
2026-02-21 03:10:39 +00:00
avvio@9.2.0:
dependencies:
'@fastify/error': 4.2.0
fastq: 1.20.1
feat: add Kubernetes orchestrator for MCP server pod management mcpd can now deploy MCP server instances as Kubernetes pods instead of Docker containers. Set MCPD_ORCHESTRATOR=kubernetes to enable. - Add @kubernetes/client-node with thin wrapper (context enforcement via MCPD_K8S_CONTEXT to prevent multi-cluster mishaps) - Rewrite KubernetesOrchestrator: pod CRUD, pod IP extraction, exec via SPDY (one-shot + interactive), log streaming - Manifest generator: stdin:true for STDIO servers, args (not command) to preserve runner image entrypoint, security hardening - Orchestrator selection in main.ts via MCPD_ORCHESTRATOR env var - 25 unit tests for k8s orchestrator, all 624 tests pass Tested end-to-end on local k3s: - mcpd deployed via Pulumi, creates pods in mcpctl-servers namespace - NetworkPolicy verified: only mcpd can reach MCP server pods - Python runner (uvx) successfully runs aws-documentation-mcp-server Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-08 01:55:13 +01:00
b4a@1.8.0: {}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
babel-plugin-jsx-dom-expressions@0.40.7(@babel/core@7.28.0):
dependencies:
'@babel/core': 7.28.0
'@babel/helper-module-imports': 7.18.6
'@babel/plugin-syntax-jsx': 7.29.7(@babel/core@7.28.0)
'@babel/types': 7.29.8
html-entities: 2.3.3
parse5: 7.3.0
babel-plugin-module-resolver@5.0.2:
dependencies:
find-babel-config: 2.1.2
glob: 9.3.5
pkg-up: 3.1.0
reselect: 4.1.8
resolve: 1.22.12
babel-preset-solid@1.9.12(@babel/core@7.28.0)(solid-js@1.9.12):
dependencies:
'@babel/core': 7.28.0
babel-plugin-jsx-dom-expressions: 0.40.7(@babel/core@7.28.0)
optionalDependencies:
solid-js: 1.9.12
first commit
2026-02-21 03:10:39 +00:00
balanced-match@1.0.2: {}
balanced-match@4.0.3: {}
feat: add Kubernetes orchestrator for MCP server pod management mcpd can now deploy MCP server instances as Kubernetes pods instead of Docker containers. Set MCPD_ORCHESTRATOR=kubernetes to enable. - Add @kubernetes/client-node with thin wrapper (context enforcement via MCPD_K8S_CONTEXT to prevent multi-cluster mishaps) - Rewrite KubernetesOrchestrator: pod CRUD, pod IP extraction, exec via SPDY (one-shot + interactive), log streaming - Manifest generator: stdin:true for STDIO servers, args (not command) to preserve runner image entrypoint, security hardening - Orchestrator selection in main.ts via MCPD_ORCHESTRATOR env var - 25 unit tests for k8s orchestrator, all 624 tests pass Tested end-to-end on local k3s: - mcpd deployed via Pulumi, creates pods in mcpctl-servers namespace - NetworkPolicy verified: only mcpd can reach MCP server pods - Python runner (uvx) successfully runs aws-documentation-mcp-server Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-08 01:55:13 +01:00
bare-events@2.8.2: {}
bare-fs@4.6.0:
dependencies:
bare-events: 2.8.2
bare-path: 3.0.0
bare-stream: 2.12.0(bare-events@2.8.2)
bare-url: 2.4.0
fast-fifo: 1.3.2
transitivePeerDependencies:
- bare-abort-controller
- react-native-b4a
bare-os@3.8.7: {}
bare-path@3.0.0:
dependencies:
bare-os: 3.8.7
bare-stream@2.12.0(bare-events@2.8.2):
dependencies:
streamx: 2.25.0
teex: 1.0.1
optionalDependencies:
bare-events: 2.8.2
transitivePeerDependencies:
- react-native-b4a
bare-url@2.4.0:
dependencies:
bare-path: 3.0.0
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
base64-js@1.5.1: {}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
baseline-browser-mapping@2.10.23: {}
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
bcrypt-pbkdf@1.0.2:
dependencies:
tweetnacl: 0.14.5
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
bcrypt@5.1.1:
dependencies:
'@mapbox/node-pre-gyp': 1.0.11
node-addon-api: 5.1.0
transitivePeerDependencies:
- encoding
- supports-color
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
bidi-js@1.0.3:
dependencies:
require-from-string: 2.0.2
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
bignumber.js@9.3.1: {}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
bin-links@6.0.2:
dependencies:
cmd-shim: 8.0.0
npm-normalize-package-bin: 5.0.0
proc-log: 6.1.0
read-cmd-shim: 6.0.0
write-file-atomic: 7.0.1
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
bl@4.1.0:
dependencies:
buffer: 5.7.1
inherits: 2.0.4
readable-stream: 3.6.2
first commit
2026-02-21 03:10:39 +00:00
body-parser@2.2.2:
dependencies:
bytes: 3.1.2
content-type: 1.0.5
debug: 4.4.3
http-errors: 2.0.1
iconv-lite: 0.7.2
on-finished: 2.4.1
qs: 6.15.0
raw-body: 3.0.2
type-is: 2.0.1
transitivePeerDependencies:
- supports-color
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
bowser@2.14.1: {}
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
brace-expansion@1.1.12:
dependencies:
balanced-match: 1.0.2
concat-map: 0.0.1
first commit
2026-02-21 03:10:39 +00:00
brace-expansion@2.0.2:
dependencies:
balanced-match: 1.0.2
brace-expansion@5.0.2:
dependencies:
balanced-match: 4.0.3
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
brace-expansion@5.0.9:
dependencies:
balanced-match: 4.0.3
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
browserslist@4.28.2:
dependencies:
baseline-browser-mapping: 2.10.23
caniuse-lite: 1.0.30001791
electron-to-chromium: 1.5.344
node-releases: 2.0.38
update-browserslist-db: 1.2.3(browserslist@4.28.2)
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
buffer-equal-constant-time@1.0.1: {}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
buffer-from@1.1.2: {}
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
buffer@5.7.1:
dependencies:
base64-js: 1.5.1
ieee754: 1.2.1
buildcheck@0.0.7:
optional: true
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
bun-ffi-structs@0.2.4(typescript@5.9.3):
dependencies:
typescript: 5.9.3
first commit
2026-02-21 03:10:39 +00:00
bytes@3.1.2: {}
c12@3.1.0:
dependencies:
chokidar: 4.0.3
confbox: 0.2.4
defu: 6.1.4
dotenv: 16.6.1
exsolve: 1.0.8
giget: 2.0.0
jiti: 2.6.1
ohash: 2.0.11
pathe: 2.0.3
perfect-debounce: 1.0.0
pkg-types: 2.3.0
rc9: 2.1.2
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
cacache@20.0.4:
dependencies:
'@npmcli/fs': 5.0.0
fs-minipass: 3.0.3
glob: 13.0.6
lru-cache: 11.2.6
minipass: 7.1.3
minipass-collect: 2.0.1
minipass-flush: 1.0.7
minipass-pipeline: 1.2.4
p-map: 7.0.5
ssri: 13.0.1
first commit
2026-02-21 03:10:39 +00:00
call-bind-apply-helpers@1.0.2:
dependencies:
es-errors: 1.3.0
function-bind: 1.1.2
call-bound@1.0.4:
dependencies:
call-bind-apply-helpers: 1.0.2
get-intrinsic: 1.3.0
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
caniuse-lite@1.0.30001791: {}
first commit
2026-02-21 03:10:39 +00:00
chai@6.2.2: {}
chalk@5.6.2: {}
chardet@2.1.1: {}
chokidar@4.0.3:
dependencies:
readdirp: 4.1.2
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
chownr@1.1.4: {}
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
chownr@2.0.0: {}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
chownr@3.0.0: {}
first commit
2026-02-21 03:10:39 +00:00
citty@0.1.6:
dependencies:
consola: 3.4.2
citty@0.2.1: {}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
cjs-module-lexer@2.2.0: {}
feat(web): bold redesign — Tailwind v4 + shadcn-style primitives + Skills/Proposals/Revisions UI Phase 6 of the Skills + Revisions + Proposals work. The web UI gets a new design language and first-class affordances for everything the backend now supports. ## Visual direction - Tailwind v4 with custom @theme block (oklch tokens). Dark-mode-only (internal tool — light mode doubles QA surface). - Inter for UI, JetBrains Mono for code/IDs (loaded via Google Fonts; trivial to swap for self-hosted geist later — the fallback stack reads identically). - Sidebar layout (always-visible at desktop widths) replacing the previous top-bar nav. Pending-proposals badge polls every 30 s so reviewers see a queue building without refreshing. - Lucide icons throughout. - Spacing and radii on Tailwind defaults. Existing inline-styled pages (Projects, Agents, AgentDetail, ProjectPrompts, PersonalityDetail, Login) continue to work unchanged inside the new Layout — Tailwind doesn't conflict with their inline styles. A follow-up can migrate them incrementally. ## What's added ### Build infra (src/web/) - package.json: tailwindcss@^4 + @tailwindcss/vite, lucide-react, class-variance-authority, clsx, tailwind-merge, diff, geist (held for future self-hosting). - vite.config.ts: registers the @tailwindcss/vite plugin. - src/index.css: Tailwind import + @theme tokens + @layer base. - src/main.tsx: imports index.css. - src/lib/utils.ts: shadcn-style cn() helper. ### shadcn-style primitives (src/components/ui/) Hand-written rather than generated via `npx shadcn` so the repo doesn't depend on a CLI tool that needs interactive runtime: - button.tsx — variants: primary / secondary / ghost / danger / link; sizes: sm / md / lg / icon. - card.tsx — Card + Header/Title/Description/Content/Footer subparts. - badge.tsx — variants: default / info / success / warning / danger / outline. - input.tsx — Input + Textarea + Label. - tabs.tsx — no-dep accessible Tabs (no Radix needed for our use). - separator.tsx — h/v separator with role=separator. ### Diff component (src/components/Diff.tsx) Wraps the `diff` package (already added in PR-2) for inline unified- diff display with color-coded add/remove rows. Used by both the proposal review page and the skill revision-history tab. ### New pages (src/pages/) - Dashboard.tsx — at-a-glance home. Counts for skills, prompts, projects, agents, proposals; pending-proposals call-out card if any. - Skills.tsx — list view, separated into Global vs Project/Agent- scoped sections. - SkillDetail.tsx — name + semver + description; tabs for SKILL.md / Files / Metadata / History. History tab shows revisions with click-to-diff against the live body. - Proposals.tsx — queue with Pending/Approved/Rejected tabs. Pending count is highlighted in amber. - ProposalDetail.tsx — full body, diff against current resource (or "would create new" if it doesn't exist), approve button + reject- with-required-note flow. ### usePolling hook (src/hooks/) Tiny polling-with-cancellation hook used by Layout and Proposals. ### Layout rewrite (src/components/Layout.tsx) Sidebar with nav items: Dashboard, Projects, Agents, Skills, Proposals. Lucide icons. Active-route highlighting via NavLink. Pending-proposals warning badge on the Proposals item. ### Routes (src/App.tsx) New routes: /dashboard, /skills, /skills/:name, /proposals, /proposals/:id. Default redirects to /dashboard. ### API types (src/api.ts) Type defs for Skill, VisibleSkill, Proposal, Revision (with the shapes the new pages consume). ## Tests Existing 7 web tests still pass (Login + api). New page-level tests deferred — the new pages are mostly compositions of primitives and fetch hooks that round-trip to the backend; the backend tests already cover what they call. PR-7 polish can add render-and-click tests if coverage drift surfaces. ## Verification - `pnpm --filter @mcpctl/web build` clean, no warnings. - `pnpm test:run` whole monorepo: 162 test files / 2157 tests green. - Visual smoke deferred — needs a running mcpd to populate the fixtures. Manual smoke tested locally is the next step. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-07 17:54:55 +01:00
class-variance-authority@0.7.1:
dependencies:
clsx: 2.1.1
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
cli-boxes@3.0.0: {}
cli-cursor@4.0.0:
dependencies:
restore-cursor: 4.0.0
cli-spinners@3.4.0: {}
cli-truncate@5.1.1:
dependencies:
slice-ansi: 7.1.2
string-width: 8.2.0
first commit
2026-02-21 03:10:39 +00:00
cli-width@4.1.0: {}
feat(web): bold redesign — Tailwind v4 + shadcn-style primitives + Skills/Proposals/Revisions UI Phase 6 of the Skills + Revisions + Proposals work. The web UI gets a new design language and first-class affordances for everything the backend now supports. ## Visual direction - Tailwind v4 with custom @theme block (oklch tokens). Dark-mode-only (internal tool — light mode doubles QA surface). - Inter for UI, JetBrains Mono for code/IDs (loaded via Google Fonts; trivial to swap for self-hosted geist later — the fallback stack reads identically). - Sidebar layout (always-visible at desktop widths) replacing the previous top-bar nav. Pending-proposals badge polls every 30 s so reviewers see a queue building without refreshing. - Lucide icons throughout. - Spacing and radii on Tailwind defaults. Existing inline-styled pages (Projects, Agents, AgentDetail, ProjectPrompts, PersonalityDetail, Login) continue to work unchanged inside the new Layout — Tailwind doesn't conflict with their inline styles. A follow-up can migrate them incrementally. ## What's added ### Build infra (src/web/) - package.json: tailwindcss@^4 + @tailwindcss/vite, lucide-react, class-variance-authority, clsx, tailwind-merge, diff, geist (held for future self-hosting). - vite.config.ts: registers the @tailwindcss/vite plugin. - src/index.css: Tailwind import + @theme tokens + @layer base. - src/main.tsx: imports index.css. - src/lib/utils.ts: shadcn-style cn() helper. ### shadcn-style primitives (src/components/ui/) Hand-written rather than generated via `npx shadcn` so the repo doesn't depend on a CLI tool that needs interactive runtime: - button.tsx — variants: primary / secondary / ghost / danger / link; sizes: sm / md / lg / icon. - card.tsx — Card + Header/Title/Description/Content/Footer subparts. - badge.tsx — variants: default / info / success / warning / danger / outline. - input.tsx — Input + Textarea + Label. - tabs.tsx — no-dep accessible Tabs (no Radix needed for our use). - separator.tsx — h/v separator with role=separator. ### Diff component (src/components/Diff.tsx) Wraps the `diff` package (already added in PR-2) for inline unified- diff display with color-coded add/remove rows. Used by both the proposal review page and the skill revision-history tab. ### New pages (src/pages/) - Dashboard.tsx — at-a-glance home. Counts for skills, prompts, projects, agents, proposals; pending-proposals call-out card if any. - Skills.tsx — list view, separated into Global vs Project/Agent- scoped sections. - SkillDetail.tsx — name + semver + description; tabs for SKILL.md / Files / Metadata / History. History tab shows revisions with click-to-diff against the live body. - Proposals.tsx — queue with Pending/Approved/Rejected tabs. Pending count is highlighted in amber. - ProposalDetail.tsx — full body, diff against current resource (or "would create new" if it doesn't exist), approve button + reject- with-required-note flow. ### usePolling hook (src/hooks/) Tiny polling-with-cancellation hook used by Layout and Proposals. ### Layout rewrite (src/components/Layout.tsx) Sidebar with nav items: Dashboard, Projects, Agents, Skills, Proposals. Lucide icons. Active-route highlighting via NavLink. Pending-proposals warning badge on the Proposals item. ### Routes (src/App.tsx) New routes: /dashboard, /skills, /skills/:name, /proposals, /proposals/:id. Default redirects to /dashboard. ### API types (src/api.ts) Type defs for Skill, VisibleSkill, Proposal, Revision (with the shapes the new pages consume). ## Tests Existing 7 web tests still pass (Login + api). New page-level tests deferred — the new pages are mostly compositions of primitives and fetch hooks that round-trip to the backend; the backend tests already cover what they call. PR-7 polish can add render-and-click tests if coverage drift surfaces. ## Verification - `pnpm --filter @mcpctl/web build` clean, no warnings. - `pnpm test:run` whole monorepo: 162 test files / 2157 tests green. - Visual smoke deferred — needs a running mcpd to populate the fixtures. Manual smoke tested locally is the next step. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-07 17:54:55 +01:00
client-only@0.0.1: {}
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
cliui@8.0.1:
dependencies:
string-width: 4.2.3
strip-ansi: 6.0.1
wrap-ansi: 7.0.0
feat(web): bold redesign — Tailwind v4 + shadcn-style primitives + Skills/Proposals/Revisions UI Phase 6 of the Skills + Revisions + Proposals work. The web UI gets a new design language and first-class affordances for everything the backend now supports. ## Visual direction - Tailwind v4 with custom @theme block (oklch tokens). Dark-mode-only (internal tool — light mode doubles QA surface). - Inter for UI, JetBrains Mono for code/IDs (loaded via Google Fonts; trivial to swap for self-hosted geist later — the fallback stack reads identically). - Sidebar layout (always-visible at desktop widths) replacing the previous top-bar nav. Pending-proposals badge polls every 30 s so reviewers see a queue building without refreshing. - Lucide icons throughout. - Spacing and radii on Tailwind defaults. Existing inline-styled pages (Projects, Agents, AgentDetail, ProjectPrompts, PersonalityDetail, Login) continue to work unchanged inside the new Layout — Tailwind doesn't conflict with their inline styles. A follow-up can migrate them incrementally. ## What's added ### Build infra (src/web/) - package.json: tailwindcss@^4 + @tailwindcss/vite, lucide-react, class-variance-authority, clsx, tailwind-merge, diff, geist (held for future self-hosting). - vite.config.ts: registers the @tailwindcss/vite plugin. - src/index.css: Tailwind import + @theme tokens + @layer base. - src/main.tsx: imports index.css. - src/lib/utils.ts: shadcn-style cn() helper. ### shadcn-style primitives (src/components/ui/) Hand-written rather than generated via `npx shadcn` so the repo doesn't depend on a CLI tool that needs interactive runtime: - button.tsx — variants: primary / secondary / ghost / danger / link; sizes: sm / md / lg / icon. - card.tsx — Card + Header/Title/Description/Content/Footer subparts. - badge.tsx — variants: default / info / success / warning / danger / outline. - input.tsx — Input + Textarea + Label. - tabs.tsx — no-dep accessible Tabs (no Radix needed for our use). - separator.tsx — h/v separator with role=separator. ### Diff component (src/components/Diff.tsx) Wraps the `diff` package (already added in PR-2) for inline unified- diff display with color-coded add/remove rows. Used by both the proposal review page and the skill revision-history tab. ### New pages (src/pages/) - Dashboard.tsx — at-a-glance home. Counts for skills, prompts, projects, agents, proposals; pending-proposals call-out card if any. - Skills.tsx — list view, separated into Global vs Project/Agent- scoped sections. - SkillDetail.tsx — name + semver + description; tabs for SKILL.md / Files / Metadata / History. History tab shows revisions with click-to-diff against the live body. - Proposals.tsx — queue with Pending/Approved/Rejected tabs. Pending count is highlighted in amber. - ProposalDetail.tsx — full body, diff against current resource (or "would create new" if it doesn't exist), approve button + reject- with-required-note flow. ### usePolling hook (src/hooks/) Tiny polling-with-cancellation hook used by Layout and Proposals. ### Layout rewrite (src/components/Layout.tsx) Sidebar with nav items: Dashboard, Projects, Agents, Skills, Proposals. Lucide icons. Active-route highlighting via NavLink. Pending-proposals warning badge on the Proposals item. ### Routes (src/App.tsx) New routes: /dashboard, /skills, /skills/:name, /proposals, /proposals/:id. Default redirects to /dashboard. ### API types (src/api.ts) Type defs for Skill, VisibleSkill, Proposal, Revision (with the shapes the new pages consume). ## Tests Existing 7 web tests still pass (Login + api). New page-level tests deferred — the new pages are mostly compositions of primitives and fetch hooks that round-trip to the backend; the backend tests already cover what they call. PR-7 polish can add render-and-click tests if coverage drift surfaces. ## Verification - `pnpm --filter @mcpctl/web build` clean, no warnings. - `pnpm test:run` whole monorepo: 162 test files / 2157 tests green. - Visual smoke deferred — needs a running mcpd to populate the fixtures. Manual smoke tested locally is the next step. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-07 17:54:55 +01:00
clsx@2.1.1: {}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
cmd-shim@8.0.0: {}
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
code-excerpt@4.0.0:
dependencies:
convert-to-spaces: 2.0.1
first commit
2026-02-21 03:10:39 +00:00
color-convert@2.0.1:
dependencies:
color-name: 1.1.4
color-name@1.1.4: {}
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
color-support@1.1.3: {}
feat: add Kubernetes orchestrator for MCP server pod management mcpd can now deploy MCP server instances as Kubernetes pods instead of Docker containers. Set MCPD_ORCHESTRATOR=kubernetes to enable. - Add @kubernetes/client-node with thin wrapper (context enforcement via MCPD_K8S_CONTEXT to prevent multi-cluster mishaps) - Rewrite KubernetesOrchestrator: pod CRUD, pod IP extraction, exec via SPDY (one-shot + interactive), log streaming - Manifest generator: stdin:true for STDIO servers, args (not command) to preserve runner image entrypoint, security hardening - Orchestrator selection in main.ts via MCPD_ORCHESTRATOR env var - 25 unit tests for k8s orchestrator, all 624 tests pass Tested end-to-end on local k3s: - mcpd deployed via Pulumi, creates pods in mcpctl-servers namespace - NetworkPolicy verified: only mcpd can reach MCP server pods - Python runner (uvx) successfully runs aws-documentation-mcp-server Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-08 01:55:13 +01:00
combined-stream@1.0.8:
dependencies:
delayed-stream: 1.0.0
first commit
2026-02-21 03:10:39 +00:00
commander@13.1.0: {}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
common-ancestor-path@2.0.0: {}
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
concat-map@0.0.1: {}
first commit
2026-02-21 03:10:39 +00:00
confbox@0.2.4: {}
consola@3.4.2: {}
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
console-control-strings@1.1.0: {}
feat(mcpd+deploy): serve web UI at /ui + smoke tests + docs (Stage 6) The closing stage. mcpd now hosts the Stage 5 SPA, the Docker image bundles the build artifact, a smoke test exercises the personality HTTP surface end-to-end, and the user-facing docs spell out the mental model. mcpd: - Add @fastify/static dep. - New routes/web-ui.ts: registers /ui/* against a static bundle. Looks for the bundle at $MCPD_WEB_ROOT, then /usr/share/mcpd/web (the Docker image path), then a dev-tree fallback. Logs and skips cleanly if missing — API-only deploys keep working. - SPA fallback: any /ui/<path> that doesn't match a file falls through to index.html so direct hits to react-router URLs work. - /ui/* falls through to `kind: skip` in mapUrlToPermission, so the static assets are served unauthenticated. Each API call from the SPA still carries the bearer token. Deploy: - Dockerfile.mcpd builds the @mcpctl/web bundle in the same builder stage and copies dist/ to /usr/share/mcpd/web in the runtime image. Smoke (personality.smoke.test.ts): - Live mcpd flow: create secret/llm/agent/personality, attach an agent-direct prompt, verify the binding listing, reject double- attach (409) + foreign-agent prompt (400), set defaultPersonality by name, detach + delete cleanup. Docs: - New docs/personalities.md: VLAN-on-ethernet model, system-block ordering table, three prompt scopes, CLI walkthrough, web UI walkthrough, full API surface, RBAC notes. - agents.md and chat.md cross-link. - README's Agents section gains a Personalities subsection. Test count after Stage 6: mcpd: 801/801 cli: 430/430 web: 7/7 db: 58/62 (4 pre-existing) Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:48:43 +01:00
content-disposition@0.5.4:
dependencies:
safe-buffer: 5.2.1
first commit
2026-02-21 03:10:39 +00:00
content-disposition@1.0.1: {}
content-type@1.0.5: {}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
convert-source-map@2.0.0: {}
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
convert-to-spaces@2.0.1: {}
first commit
2026-02-21 03:10:39 +00:00
cookie-signature@1.2.2: {}
cookie@0.7.2: {}
cookie@1.1.1: {}
cors@2.8.6:
dependencies:
object-assign: 4.1.1
vary: 1.1.2
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
cpu-features@0.0.10:
dependencies:
buildcheck: 0.0.7
nan: 2.25.0
optional: true
first commit
2026-02-21 03:10:39 +00:00
cross-spawn@7.0.6:
dependencies:
path-key: 3.1.1
shebang-command: 2.0.0
which: 2.0.2
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
css-tree@3.2.1:
dependencies:
mdn-data: 2.27.1
source-map-js: 1.2.1
css.escape@1.5.1: {}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
cssesc@3.0.0: {}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
cssstyle@6.2.0:
dependencies:
'@asamuzakjp/css-color': 5.1.11
'@csstools/css-syntax-patches-for-csstree': 1.1.3(css-tree@3.2.1)
css-tree: 3.2.1
lru-cache: 11.2.6
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
csstype@3.2.3: {}
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
data-uri-to-buffer@4.0.1: {}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
data-urls@7.0.0:
dependencies:
whatwg-mimetype: 5.0.0
whatwg-url: 16.0.1
transitivePeerDependencies:
- '@noble/hashes'
first commit
2026-02-21 03:10:39 +00:00
debug@4.4.3:
dependencies:
ms: 2.1.3
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
decimal.js@10.6.0: {}
first commit
2026-02-21 03:10:39 +00:00
deep-is@0.1.4: {}
deepmerge-ts@7.1.5: {}
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
deepmerge@4.3.1: {}
first commit
2026-02-21 03:10:39 +00:00
defu@6.1.4: {}
feat: add Kubernetes orchestrator for MCP server pod management mcpd can now deploy MCP server instances as Kubernetes pods instead of Docker containers. Set MCPD_ORCHESTRATOR=kubernetes to enable. - Add @kubernetes/client-node with thin wrapper (context enforcement via MCPD_K8S_CONTEXT to prevent multi-cluster mishaps) - Rewrite KubernetesOrchestrator: pod CRUD, pod IP extraction, exec via SPDY (one-shot + interactive), log streaming - Manifest generator: stdin:true for STDIO servers, args (not command) to preserve runner image entrypoint, security hardening - Orchestrator selection in main.ts via MCPD_ORCHESTRATOR env var - 25 unit tests for k8s orchestrator, all 624 tests pass Tested end-to-end on local k3s: - mcpd deployed via Pulumi, creates pods in mcpctl-servers namespace - NetworkPolicy verified: only mcpd can reach MCP server pods - Python runner (uvx) successfully runs aws-documentation-mcp-server Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-08 01:55:13 +01:00
delayed-stream@1.0.0: {}
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
delegates@1.0.0: {}
first commit
2026-02-21 03:10:39 +00:00
depd@2.0.0: {}
dequal@2.0.3: {}
destr@2.0.5: {}
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
detect-libc@2.1.2: {}
feat(mcpd): ResourceRevision + ResourceProposal services + Prompt revision integration Phase 2 of the Skills + Revisions + Proposals work. Stands up the generic revision/proposal layer and wires Prompt into it. Skills will plug into the same infrastructure in PR-3 with no further service changes required. This PR is intentionally additive: PromptRequest table and routes are unchanged. The /api/v1/proposals API runs side-by-side with the legacy /api/v1/promptrequests API. The PromptRequest cutover (rename + backfill + mcplocal rewire) is deferred to a later PR so this one stays reviewable. ## What's added ### Repositories (src/mcpd/src/repositories/) - resource-revision.repository.ts — append-only revision log keyed by (resourceType, resourceId). Soft FK; no relations declared. Supports history listing, semver lookup, and contentHash cross-resource search. - resource-proposal.repository.ts — generic propose queue. Status lifecycle pending → approved | rejected. Mirrors Prompt's `?? ''` workaround for nullable-FK compound lookups. ### Services (src/mcpd/src/services/) - resource-revision.service.ts — record() inserts a revision with a stable sha256 contentHash computed from canonicalised JSON (key-sorted at every level so reordered objects produce the same hash). Caller passes a pre-computed semver; service does NOT decide bump policy. - resource-proposal.service.ts — propose / approve / reject / list, with a per-resourceType handler registry. PromptService registers the 'prompt' handler at construction; the SkillService will register 'skill' in PR-3. approve() runs in a Prisma $transaction so the resource update + revision insert + proposal status flip are atomic. ### Pure utility (src/mcpd/src/utils/semver.ts) - bumpSemver(current, kind) for major / minor / patch - compareSemver(a, b) — numeric, not lex (10 > 9) - isValidSemver(s) - Invalid input falls back to '0.1.0' rather than throwing — keeps the audit-write path from blowing up the prompt update if a row's semver ever drifts out of MAJOR.MINOR.PATCH shape. ### Routes (src/mcpd/src/routes/) - revisions.ts — GET /api/v1/revisions?resourceType=&resourceId=, GET /api/v1/revisions/:id, GET /api/v1/revisions/:id/diff?against=<id|live> (unified-format diff via the `diff` package), and POST /api/v1/prompts/:id/restore-revision { revisionId, note? }. - proposals.ts — GET / POST /api/v1/proposals, GET /api/v1/proposals/:id, PUT for body updates, POST .../approve and POST .../reject, plus DELETE. ## What's changed - PromptService.create / update now record a ResourceRevision when the revision service is wired. Update auto-bumps patch on content change; authors can override via `--bump major|minor|patch` or `--semver X.Y.Z` on the CLI (forwarded into the PUT body). Best-effort: revision write failures are swallowed so the prompt save still succeeds (revision is audit, not source of truth). - PromptService.setProposalService registers a 'prompt' approval handler with the proposal service. Approval runs in a Prisma transaction: upsert prompt → record revision → update currentRevisionId → flip proposal status. semver bumps to 0.1.0 on first approval, patch thereafter. - New CLI flags on `mcpctl edit prompt`: --bump, --semver, --note. They're prompt-only (validated client-side); other resources reject them. - Aliases in shared.ts: `proposal`/`prop` → proposals, `revision`/`rev` → revisions. - diff dependency added to mcpd. ## Tests - src/mcpd/tests/utils/semver.test.ts — covers bump/compare/validate including numeric (not lex) semver compare and invalid-input fallback. - prompt-service.test.ts updated: makePrompt fixture now sets semver + agentId + currentRevisionId; updatePrompt assertion expects the auto-bumped patch in the same update call. - prompt-routes.test.ts updated symmetrically. ## RBAC `proposals` and `revisions` URL segments map to the existing `prompts` permission for now. PR-7 may split if a "reviewer" role becomes useful. ## Verification Full suite: 158 test files / 2127 tests green. `pnpm build` clean across all 6 workspace packages. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-07 00:38:35 +01:00
diff@5.2.2: {}
feat: eager vLLM warmup and smart page titles in paginate stage - Add warmup() to LlmProvider interface for eager subprocess startup - ManagedVllmProvider.warmup() starts vLLM in background on project load - ProviderRegistry.warmupAll() triggers all managed providers - NamedProvider proxies warmup() to inner provider - paginate stage generates LLM-powered descriptive page titles when available, cached by content hash, falls back to generic "Page N" - project-mcp-endpoint calls warmupAll() on router creation so vLLM is loading while the session initializes Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-03-03 19:07:39 +00:00
diff@8.0.3: {}
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
diff@8.0.4: {}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
diff@9.0.0: {}
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
docker-modem@5.0.6:
dependencies:
debug: 4.4.3
readable-stream: 3.6.2
split-ca: 1.0.1
ssh2: 1.17.0
transitivePeerDependencies:
- supports-color
dockerode@4.0.9:
dependencies:
'@balena/dockerignore': 1.0.2
'@grpc/grpc-js': 1.14.3
'@grpc/proto-loader': 0.7.15
docker-modem: 5.0.6
protobufjs: 7.5.4
tar-fs: 2.1.4
uuid: 10.0.0
transitivePeerDependencies:
- supports-color
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
dom-accessibility-api@0.5.16: {}
dom-accessibility-api@0.6.3: {}
dompurify@3.2.7:
optionalDependencies:
'@types/trusted-types': 2.0.7
first commit
2026-02-21 03:10:39 +00:00
dotenv@16.6.1: {}
dunder-proto@1.0.1:
dependencies:
call-bind-apply-helpers: 1.0.2
es-errors: 1.3.0
gopd: 1.2.0
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
ecdsa-sig-formatter@1.0.11:
dependencies:
safe-buffer: 5.2.1
first commit
2026-02-21 03:10:39 +00:00
ee-first@1.1.1: {}
effect@3.18.4:
dependencies:
'@standard-schema/spec': 1.1.0
fast-check: 3.23.2
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
effect@4.0.0-beta.83:
dependencies:
'@standard-schema/spec': 1.1.0
fast-check: 4.9.0
find-my-way-ts: 0.1.6
ini: 7.0.0
kubernetes-types: 1.30.0
msgpackr: 2.0.5
multipasta: 0.2.8
toml: 4.3.0
uuid: 14.0.1
yaml: 2.9.0
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
electron-to-chromium@1.5.344: {}
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
emoji-regex@10.6.0: {}
first commit
2026-02-21 03:10:39 +00:00
emoji-regex@8.0.0: {}
empathic@2.0.0: {}
encodeurl@2.0.0: {}
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
end-of-stream@1.4.5:
dependencies:
once: 1.4.0
feat(web): bold redesign — Tailwind v4 + shadcn-style primitives + Skills/Proposals/Revisions UI Phase 6 of the Skills + Revisions + Proposals work. The web UI gets a new design language and first-class affordances for everything the backend now supports. ## Visual direction - Tailwind v4 with custom @theme block (oklch tokens). Dark-mode-only (internal tool — light mode doubles QA surface). - Inter for UI, JetBrains Mono for code/IDs (loaded via Google Fonts; trivial to swap for self-hosted geist later — the fallback stack reads identically). - Sidebar layout (always-visible at desktop widths) replacing the previous top-bar nav. Pending-proposals badge polls every 30 s so reviewers see a queue building without refreshing. - Lucide icons throughout. - Spacing and radii on Tailwind defaults. Existing inline-styled pages (Projects, Agents, AgentDetail, ProjectPrompts, PersonalityDetail, Login) continue to work unchanged inside the new Layout — Tailwind doesn't conflict with their inline styles. A follow-up can migrate them incrementally. ## What's added ### Build infra (src/web/) - package.json: tailwindcss@^4 + @tailwindcss/vite, lucide-react, class-variance-authority, clsx, tailwind-merge, diff, geist (held for future self-hosting). - vite.config.ts: registers the @tailwindcss/vite plugin. - src/index.css: Tailwind import + @theme tokens + @layer base. - src/main.tsx: imports index.css. - src/lib/utils.ts: shadcn-style cn() helper. ### shadcn-style primitives (src/components/ui/) Hand-written rather than generated via `npx shadcn` so the repo doesn't depend on a CLI tool that needs interactive runtime: - button.tsx — variants: primary / secondary / ghost / danger / link; sizes: sm / md / lg / icon. - card.tsx — Card + Header/Title/Description/Content/Footer subparts. - badge.tsx — variants: default / info / success / warning / danger / outline. - input.tsx — Input + Textarea + Label. - tabs.tsx — no-dep accessible Tabs (no Radix needed for our use). - separator.tsx — h/v separator with role=separator. ### Diff component (src/components/Diff.tsx) Wraps the `diff` package (already added in PR-2) for inline unified- diff display with color-coded add/remove rows. Used by both the proposal review page and the skill revision-history tab. ### New pages (src/pages/) - Dashboard.tsx — at-a-glance home. Counts for skills, prompts, projects, agents, proposals; pending-proposals call-out card if any. - Skills.tsx — list view, separated into Global vs Project/Agent- scoped sections. - SkillDetail.tsx — name + semver + description; tabs for SKILL.md / Files / Metadata / History. History tab shows revisions with click-to-diff against the live body. - Proposals.tsx — queue with Pending/Approved/Rejected tabs. Pending count is highlighted in amber. - ProposalDetail.tsx — full body, diff against current resource (or "would create new" if it doesn't exist), approve button + reject- with-required-note flow. ### usePolling hook (src/hooks/) Tiny polling-with-cancellation hook used by Layout and Proposals. ### Layout rewrite (src/components/Layout.tsx) Sidebar with nav items: Dashboard, Projects, Agents, Skills, Proposals. Lucide icons. Active-route highlighting via NavLink. Pending-proposals warning badge on the Proposals item. ### Routes (src/App.tsx) New routes: /dashboard, /skills, /skills/:name, /proposals, /proposals/:id. Default redirects to /dashboard. ### API types (src/api.ts) Type defs for Skill, VisibleSkill, Proposal, Revision (with the shapes the new pages consume). ## Tests Existing 7 web tests still pass (Login + api). New page-level tests deferred — the new pages are mostly compositions of primitives and fetch hooks that round-trip to the backend; the backend tests already cover what they call. PR-7 polish can add render-and-click tests if coverage drift surfaces. ## Verification - `pnpm --filter @mcpctl/web build` clean, no warnings. - `pnpm test:run` whole monorepo: 162 test files / 2157 tests green. - Visual smoke deferred — needs a running mcpd to populate the fixtures. Manual smoke tested locally is the next step. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-07 17:54:55 +01:00
enhanced-resolve@5.21.0:
dependencies:
graceful-fs: 4.2.11
tapable: 2.3.3
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
entities@6.0.1: {}
entities@7.0.1: {}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
entities@8.0.0: {}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
env-paths@2.2.1: {}
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
environment@1.1.0: {}
first commit
2026-02-21 03:10:39 +00:00
es-define-property@1.0.1: {}
es-errors@1.3.0: {}
es-module-lexer@1.7.0: {}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
es-module-lexer@2.3.1: {}
first commit
2026-02-21 03:10:39 +00:00
es-object-atoms@1.1.1:
dependencies:
es-errors: 1.3.0
feat: add Kubernetes orchestrator for MCP server pod management mcpd can now deploy MCP server instances as Kubernetes pods instead of Docker containers. Set MCPD_ORCHESTRATOR=kubernetes to enable. - Add @kubernetes/client-node with thin wrapper (context enforcement via MCPD_K8S_CONTEXT to prevent multi-cluster mishaps) - Rewrite KubernetesOrchestrator: pod CRUD, pod IP extraction, exec via SPDY (one-shot + interactive), log streaming - Manifest generator: stdin:true for STDIO servers, args (not command) to preserve runner image entrypoint, security hardening - Orchestrator selection in main.ts via MCPD_ORCHESTRATOR env var - 25 unit tests for k8s orchestrator, all 624 tests pass Tested end-to-end on local k3s: - mcpd deployed via Pulumi, creates pods in mcpctl-servers namespace - NetworkPolicy verified: only mcpd can reach MCP server pods - Python runner (uvx) successfully runs aws-documentation-mcp-server Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-08 01:55:13 +01:00
es-set-tostringtag@2.1.0:
dependencies:
es-errors: 1.3.0
get-intrinsic: 1.3.0
has-tostringtag: 1.0.2
hasown: 2.0.2
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
es-toolkit@1.44.0: {}
first commit
2026-02-21 03:10:39 +00:00
esbuild@0.27.3:
optionalDependencies:
'@esbuild/aix-ppc64': 0.27.3
'@esbuild/android-arm': 0.27.3
'@esbuild/android-arm64': 0.27.3
'@esbuild/android-x64': 0.27.3
'@esbuild/darwin-arm64': 0.27.3
'@esbuild/darwin-x64': 0.27.3
'@esbuild/freebsd-arm64': 0.27.3
'@esbuild/freebsd-x64': 0.27.3
'@esbuild/linux-arm': 0.27.3
'@esbuild/linux-arm64': 0.27.3
'@esbuild/linux-ia32': 0.27.3
'@esbuild/linux-loong64': 0.27.3
'@esbuild/linux-mips64el': 0.27.3
'@esbuild/linux-ppc64': 0.27.3
'@esbuild/linux-riscv64': 0.27.3
'@esbuild/linux-s390x': 0.27.3
'@esbuild/linux-x64': 0.27.3
'@esbuild/netbsd-arm64': 0.27.3
'@esbuild/netbsd-x64': 0.27.3
'@esbuild/openbsd-arm64': 0.27.3
'@esbuild/openbsd-x64': 0.27.3
'@esbuild/openharmony-arm64': 0.27.3
'@esbuild/sunos-x64': 0.27.3
'@esbuild/win32-arm64': 0.27.3
'@esbuild/win32-ia32': 0.27.3
'@esbuild/win32-x64': 0.27.3
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
escalade@3.2.0: {}
first commit
2026-02-21 03:10:39 +00:00
escape-html@1.0.3: {}
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
escape-string-regexp@2.0.0: {}
first commit
2026-02-21 03:10:39 +00:00
escape-string-regexp@4.0.0: {}
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
eslint-config-prettier@10.1.8(eslint@10.0.1(jiti@2.7.0)):
first commit
2026-02-21 03:10:39 +00:00
dependencies:
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
eslint: 10.0.1(jiti@2.7.0)
first commit
2026-02-21 03:10:39 +00:00
eslint-scope@9.1.1:
dependencies:
'@types/esrecurse': 4.3.1
'@types/estree': 1.0.8
esrecurse: 4.3.0
estraverse: 5.3.0
eslint-visitor-keys@3.4.3: {}
eslint-visitor-keys@5.0.1: {}
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
eslint@10.0.1(jiti@2.7.0):
first commit
2026-02-21 03:10:39 +00:00
dependencies:
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
'@eslint-community/eslint-utils': 4.9.1(eslint@10.0.1(jiti@2.7.0))
first commit
2026-02-21 03:10:39 +00:00
'@eslint-community/regexpp': 4.12.2
'@eslint/config-array': 0.23.2
'@eslint/config-helpers': 0.5.2
'@eslint/core': 1.1.0
'@eslint/plugin-kit': 0.6.0
'@humanfs/node': 0.16.7
'@humanwhocodes/module-importer': 1.0.1
'@humanwhocodes/retry': 0.4.3
'@types/estree': 1.0.8
ajv: 6.14.0
cross-spawn: 7.0.6
debug: 4.4.3
escape-string-regexp: 4.0.0
eslint-scope: 9.1.1
eslint-visitor-keys: 5.0.1
espree: 11.1.1
esquery: 1.7.0
esutils: 2.0.3
fast-deep-equal: 3.1.3
file-entry-cache: 8.0.0
find-up: 5.0.0
glob-parent: 6.0.2
ignore: 5.3.2
imurmurhash: 0.1.4
is-glob: 4.0.3
json-stable-stringify-without-jsonify: 1.0.1
minimatch: 10.2.2
natural-compare: 1.4.0
optionator: 0.9.4
optionalDependencies:
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
jiti: 2.7.0
first commit
2026-02-21 03:10:39 +00:00
transitivePeerDependencies:
- supports-color
espree@11.1.1:
dependencies:
acorn: 8.16.0
acorn-jsx: 5.3.2(acorn@8.16.0)
eslint-visitor-keys: 5.0.1
esquery@1.7.0:
dependencies:
estraverse: 5.3.0
esrecurse@4.3.0:
dependencies:
estraverse: 5.3.0
estraverse@5.3.0: {}
estree-walker@3.0.3:
dependencies:
'@types/estree': 1.0.8
esutils@2.0.3: {}
etag@1.8.1: {}
feat: add Kubernetes orchestrator for MCP server pod management mcpd can now deploy MCP server instances as Kubernetes pods instead of Docker containers. Set MCPD_ORCHESTRATOR=kubernetes to enable. - Add @kubernetes/client-node with thin wrapper (context enforcement via MCPD_K8S_CONTEXT to prevent multi-cluster mishaps) - Rewrite KubernetesOrchestrator: pod CRUD, pod IP extraction, exec via SPDY (one-shot + interactive), log streaming - Manifest generator: stdin:true for STDIO servers, args (not command) to preserve runner image entrypoint, security hardening - Orchestrator selection in main.ts via MCPD_ORCHESTRATOR env var - 25 unit tests for k8s orchestrator, all 624 tests pass Tested end-to-end on local k3s: - mcpd deployed via Pulumi, creates pods in mcpctl-servers namespace - NetworkPolicy verified: only mcpd can reach MCP server pods - Python runner (uvx) successfully runs aws-documentation-mcp-server Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-08 01:55:13 +01:00
events-universal@1.0.1:
dependencies:
bare-events: 2.8.2
transitivePeerDependencies:
- bare-abort-controller
first commit
2026-02-21 03:10:39 +00:00
eventsource-parser@3.0.6: {}
eventsource@3.0.7:
dependencies:
eventsource-parser: 3.0.6
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
execa@5.1.1:
dependencies:
cross-spawn: 7.0.6
get-stream: 6.0.1
human-signals: 2.1.0
is-stream: 2.0.1
merge-stream: 2.0.0
npm-run-path: 4.0.1
onetime: 5.1.2
signal-exit: 3.0.7
strip-final-newline: 2.0.0
first commit
2026-02-21 03:10:39 +00:00
expect-type@1.3.0: {}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
exponential-backoff@3.1.3: {}
first commit
2026-02-21 03:10:39 +00:00
express-rate-limit@8.2.1(express@5.2.1):
dependencies:
express: 5.2.1
ip-address: 10.0.1
express@5.2.1:
dependencies:
accepts: 2.0.0
body-parser: 2.2.2
content-disposition: 1.0.1
content-type: 1.0.5
cookie: 0.7.2
cookie-signature: 1.2.2
debug: 4.4.3
depd: 2.0.0
encodeurl: 2.0.0
escape-html: 1.0.3
etag: 1.8.1
finalhandler: 2.1.1
fresh: 2.0.0
http-errors: 2.0.1
merge-descriptors: 2.0.0
mime-types: 3.0.2
on-finished: 2.4.1
once: 1.4.0
parseurl: 1.3.3
proxy-addr: 2.0.7
qs: 6.15.0
range-parser: 1.2.1
router: 2.2.0
send: 1.2.1
serve-static: 2.2.1
statuses: 2.0.2
type-is: 2.0.1
vary: 1.1.2
transitivePeerDependencies:
- supports-color
exsolve@1.0.8: {}
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
extend@3.0.2: {}
first commit
2026-02-21 03:10:39 +00:00
fast-check@3.23.2:
dependencies:
pure-rand: 6.1.0
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
fast-check@4.9.0:
dependencies:
pure-rand: 8.4.2
first commit
2026-02-21 03:10:39 +00:00
fast-decode-uri-component@1.0.1: {}
fast-deep-equal@3.1.3: {}
feat: add Kubernetes orchestrator for MCP server pod management mcpd can now deploy MCP server instances as Kubernetes pods instead of Docker containers. Set MCPD_ORCHESTRATOR=kubernetes to enable. - Add @kubernetes/client-node with thin wrapper (context enforcement via MCPD_K8S_CONTEXT to prevent multi-cluster mishaps) - Rewrite KubernetesOrchestrator: pod CRUD, pod IP extraction, exec via SPDY (one-shot + interactive), log streaming - Manifest generator: stdin:true for STDIO servers, args (not command) to preserve runner image entrypoint, security hardening - Orchestrator selection in main.ts via MCPD_ORCHESTRATOR env var - 25 unit tests for k8s orchestrator, all 624 tests pass Tested end-to-end on local k3s: - mcpd deployed via Pulumi, creates pods in mcpctl-servers namespace - NetworkPolicy verified: only mcpd can reach MCP server pods - Python runner (uvx) successfully runs aws-documentation-mcp-server Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-08 01:55:13 +01:00
fast-fifo@1.3.2: {}
first commit
2026-02-21 03:10:39 +00:00
fast-json-stable-stringify@2.1.0: {}
fast-json-stringify@6.3.0:
dependencies:
'@fastify/merge-json-schemas': 0.2.1
ajv: 8.18.0
ajv-formats: 3.0.1(ajv@8.18.0)
fast-uri: 3.1.0
json-schema-ref-resolver: 3.0.0
rfdc: 1.4.1
fast-levenshtein@2.0.6: {}
fast-querystring@1.1.2:
dependencies:
fast-decode-uri-component: 1.0.1
fast-uri@3.1.0: {}
fastify-plugin@5.1.0: {}
fastify@5.7.4:
dependencies:
'@fastify/ajv-compiler': 4.0.5
'@fastify/error': 4.2.0
'@fastify/fast-json-stringify-compiler': 5.0.3
'@fastify/proxy-addr': 5.1.0
abstract-logging: 2.0.1
avvio: 9.2.0
fast-json-stringify: 6.3.0
find-my-way: 9.4.0
light-my-request: 6.6.0
pino: 10.3.1
process-warning: 5.0.0
rfdc: 1.4.1
secure-json-parse: 4.1.0
semver: 7.7.4
toad-cache: 3.7.0
fastq@1.20.1:
dependencies:
reusify: 1.1.0
fdir@6.5.0(picomatch@4.0.3):
optionalDependencies:
picomatch: 4.0.3
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
fetch-blob@3.2.0:
dependencies:
node-domexception: 1.0.0
web-streams-polyfill: 3.3.3
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
figures@6.1.0:
dependencies:
is-unicode-supported: 2.1.0
first commit
2026-02-21 03:10:39 +00:00
file-entry-cache@8.0.0:
dependencies:
flat-cache: 4.0.1
finalhandler@2.1.1:
dependencies:
debug: 4.4.3
encodeurl: 2.0.0
escape-html: 1.0.3
on-finished: 2.4.1
parseurl: 1.3.3
statuses: 2.0.2
transitivePeerDependencies:
- supports-color
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
find-babel-config@2.1.2:
dependencies:
json5: 2.2.3
find-my-way-ts@0.1.6: {}
first commit
2026-02-21 03:10:39 +00:00
find-my-way@9.4.0:
dependencies:
fast-deep-equal: 3.1.3
fast-querystring: 1.1.2
safe-regex2: 5.0.0
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
find-up@3.0.0:
dependencies:
locate-path: 3.0.0
first commit
2026-02-21 03:10:39 +00:00
find-up@5.0.0:
dependencies:
locate-path: 6.0.0
path-exists: 4.0.0
flat-cache@4.0.1:
dependencies:
flatted: 3.3.3
keyv: 4.5.4
flatted@3.3.3: {}
feat(mcpd+deploy): serve web UI at /ui + smoke tests + docs (Stage 6) The closing stage. mcpd now hosts the Stage 5 SPA, the Docker image bundles the build artifact, a smoke test exercises the personality HTTP surface end-to-end, and the user-facing docs spell out the mental model. mcpd: - Add @fastify/static dep. - New routes/web-ui.ts: registers /ui/* against a static bundle. Looks for the bundle at $MCPD_WEB_ROOT, then /usr/share/mcpd/web (the Docker image path), then a dev-tree fallback. Logs and skips cleanly if missing — API-only deploys keep working. - SPA fallback: any /ui/<path> that doesn't match a file falls through to index.html so direct hits to react-router URLs work. - /ui/* falls through to `kind: skip` in mapUrlToPermission, so the static assets are served unauthenticated. Each API call from the SPA still carries the bearer token. Deploy: - Dockerfile.mcpd builds the @mcpctl/web bundle in the same builder stage and copies dist/ to /usr/share/mcpd/web in the runtime image. Smoke (personality.smoke.test.ts): - Live mcpd flow: create secret/llm/agent/personality, attach an agent-direct prompt, verify the binding listing, reject double- attach (409) + foreign-agent prompt (400), set defaultPersonality by name, detach + delete cleanup. Docs: - New docs/personalities.md: VLAN-on-ethernet model, system-block ordering table, three prompt scopes, CLI walkthrough, web UI walkthrough, full API surface, RBAC notes. - agents.md and chat.md cross-link. - README's Agents section gains a Personalities subsection. Test count after Stage 6: mcpd: 801/801 cli: 430/430 web: 7/7 db: 58/62 (4 pre-existing) Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:48:43 +01:00
foreground-child@3.3.1:
dependencies:
cross-spawn: 7.0.6
signal-exit: 4.1.0
feat: add Kubernetes orchestrator for MCP server pod management mcpd can now deploy MCP server instances as Kubernetes pods instead of Docker containers. Set MCPD_ORCHESTRATOR=kubernetes to enable. - Add @kubernetes/client-node with thin wrapper (context enforcement via MCPD_K8S_CONTEXT to prevent multi-cluster mishaps) - Rewrite KubernetesOrchestrator: pod CRUD, pod IP extraction, exec via SPDY (one-shot + interactive), log streaming - Manifest generator: stdin:true for STDIO servers, args (not command) to preserve runner image entrypoint, security hardening - Orchestrator selection in main.ts via MCPD_ORCHESTRATOR env var - 25 unit tests for k8s orchestrator, all 624 tests pass Tested end-to-end on local k3s: - mcpd deployed via Pulumi, creates pods in mcpctl-servers namespace - NetworkPolicy verified: only mcpd can reach MCP server pods - Python runner (uvx) successfully runs aws-documentation-mcp-server Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-08 01:55:13 +01:00
form-data@4.0.5:
dependencies:
asynckit: 0.4.0
combined-stream: 1.0.8
es-set-tostringtag: 2.1.0
hasown: 2.0.2
mime-types: 2.1.35
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
formdata-polyfill@4.0.10:
dependencies:
fetch-blob: 3.2.0
first commit
2026-02-21 03:10:39 +00:00
forwarded@0.2.0: {}
fresh@2.0.0: {}
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
fs-constants@1.0.0: {}
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
fs-minipass@2.1.0:
dependencies:
minipass: 3.3.6
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
fs-minipass@3.0.3:
dependencies:
minipass: 7.1.3
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
fs.realpath@1.0.0: {}
first commit
2026-02-21 03:10:39 +00:00
fsevents@2.3.3:
optional: true
function-bind@1.1.2: {}
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
gauge@3.0.2:
dependencies:
aproba: 2.1.0
color-support: 1.1.3
console-control-strings: 1.1.0
has-unicode: 2.0.1
object-assign: 4.1.1
signal-exit: 3.0.7
string-width: 4.2.3
strip-ansi: 6.0.1
wide-align: 1.1.5
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
gaxios@7.3.0:
dependencies:
extend: 3.0.2
https-proxy-agent: 7.0.6
node-fetch: 3.3.2
transitivePeerDependencies:
- supports-color
gcp-metadata@8.1.2:
dependencies:
gaxios: 7.3.0
google-logging-utils: 1.1.3
json-bigint: 1.0.0
transitivePeerDependencies:
- supports-color
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
geist@1.7.0(next@16.2.5(@babel/core@7.29.0)(@opentelemetry/api@1.9.1)(react-dom@19.2.5(react@19.2.5))(react@19.2.5)):
feat(web): bold redesign — Tailwind v4 + shadcn-style primitives + Skills/Proposals/Revisions UI Phase 6 of the Skills + Revisions + Proposals work. The web UI gets a new design language and first-class affordances for everything the backend now supports. ## Visual direction - Tailwind v4 with custom @theme block (oklch tokens). Dark-mode-only (internal tool — light mode doubles QA surface). - Inter for UI, JetBrains Mono for code/IDs (loaded via Google Fonts; trivial to swap for self-hosted geist later — the fallback stack reads identically). - Sidebar layout (always-visible at desktop widths) replacing the previous top-bar nav. Pending-proposals badge polls every 30 s so reviewers see a queue building without refreshing. - Lucide icons throughout. - Spacing and radii on Tailwind defaults. Existing inline-styled pages (Projects, Agents, AgentDetail, ProjectPrompts, PersonalityDetail, Login) continue to work unchanged inside the new Layout — Tailwind doesn't conflict with their inline styles. A follow-up can migrate them incrementally. ## What's added ### Build infra (src/web/) - package.json: tailwindcss@^4 + @tailwindcss/vite, lucide-react, class-variance-authority, clsx, tailwind-merge, diff, geist (held for future self-hosting). - vite.config.ts: registers the @tailwindcss/vite plugin. - src/index.css: Tailwind import + @theme tokens + @layer base. - src/main.tsx: imports index.css. - src/lib/utils.ts: shadcn-style cn() helper. ### shadcn-style primitives (src/components/ui/) Hand-written rather than generated via `npx shadcn` so the repo doesn't depend on a CLI tool that needs interactive runtime: - button.tsx — variants: primary / secondary / ghost / danger / link; sizes: sm / md / lg / icon. - card.tsx — Card + Header/Title/Description/Content/Footer subparts. - badge.tsx — variants: default / info / success / warning / danger / outline. - input.tsx — Input + Textarea + Label. - tabs.tsx — no-dep accessible Tabs (no Radix needed for our use). - separator.tsx — h/v separator with role=separator. ### Diff component (src/components/Diff.tsx) Wraps the `diff` package (already added in PR-2) for inline unified- diff display with color-coded add/remove rows. Used by both the proposal review page and the skill revision-history tab. ### New pages (src/pages/) - Dashboard.tsx — at-a-glance home. Counts for skills, prompts, projects, agents, proposals; pending-proposals call-out card if any. - Skills.tsx — list view, separated into Global vs Project/Agent- scoped sections. - SkillDetail.tsx — name + semver + description; tabs for SKILL.md / Files / Metadata / History. History tab shows revisions with click-to-diff against the live body. - Proposals.tsx — queue with Pending/Approved/Rejected tabs. Pending count is highlighted in amber. - ProposalDetail.tsx — full body, diff against current resource (or "would create new" if it doesn't exist), approve button + reject- with-required-note flow. ### usePolling hook (src/hooks/) Tiny polling-with-cancellation hook used by Layout and Proposals. ### Layout rewrite (src/components/Layout.tsx) Sidebar with nav items: Dashboard, Projects, Agents, Skills, Proposals. Lucide icons. Active-route highlighting via NavLink. Pending-proposals warning badge on the Proposals item. ### Routes (src/App.tsx) New routes: /dashboard, /skills, /skills/:name, /proposals, /proposals/:id. Default redirects to /dashboard. ### API types (src/api.ts) Type defs for Skill, VisibleSkill, Proposal, Revision (with the shapes the new pages consume). ## Tests Existing 7 web tests still pass (Login + api). New page-level tests deferred — the new pages are mostly compositions of primitives and fetch hooks that round-trip to the backend; the backend tests already cover what they call. PR-7 polish can add render-and-click tests if coverage drift surfaces. ## Verification - `pnpm --filter @mcpctl/web build` clean, no warnings. - `pnpm test:run` whole monorepo: 162 test files / 2157 tests green. - Visual smoke deferred — needs a running mcpd to populate the fixtures. Manual smoke tested locally is the next step. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-07 17:54:55 +01:00
dependencies:
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
next: 16.2.5(@babel/core@7.29.0)(@opentelemetry/api@1.9.1)(react-dom@19.2.5(react@19.2.5))(react@19.2.5)
feat(web): bold redesign — Tailwind v4 + shadcn-style primitives + Skills/Proposals/Revisions UI Phase 6 of the Skills + Revisions + Proposals work. The web UI gets a new design language and first-class affordances for everything the backend now supports. ## Visual direction - Tailwind v4 with custom @theme block (oklch tokens). Dark-mode-only (internal tool — light mode doubles QA surface). - Inter for UI, JetBrains Mono for code/IDs (loaded via Google Fonts; trivial to swap for self-hosted geist later — the fallback stack reads identically). - Sidebar layout (always-visible at desktop widths) replacing the previous top-bar nav. Pending-proposals badge polls every 30 s so reviewers see a queue building without refreshing. - Lucide icons throughout. - Spacing and radii on Tailwind defaults. Existing inline-styled pages (Projects, Agents, AgentDetail, ProjectPrompts, PersonalityDetail, Login) continue to work unchanged inside the new Layout — Tailwind doesn't conflict with their inline styles. A follow-up can migrate them incrementally. ## What's added ### Build infra (src/web/) - package.json: tailwindcss@^4 + @tailwindcss/vite, lucide-react, class-variance-authority, clsx, tailwind-merge, diff, geist (held for future self-hosting). - vite.config.ts: registers the @tailwindcss/vite plugin. - src/index.css: Tailwind import + @theme tokens + @layer base. - src/main.tsx: imports index.css. - src/lib/utils.ts: shadcn-style cn() helper. ### shadcn-style primitives (src/components/ui/) Hand-written rather than generated via `npx shadcn` so the repo doesn't depend on a CLI tool that needs interactive runtime: - button.tsx — variants: primary / secondary / ghost / danger / link; sizes: sm / md / lg / icon. - card.tsx — Card + Header/Title/Description/Content/Footer subparts. - badge.tsx — variants: default / info / success / warning / danger / outline. - input.tsx — Input + Textarea + Label. - tabs.tsx — no-dep accessible Tabs (no Radix needed for our use). - separator.tsx — h/v separator with role=separator. ### Diff component (src/components/Diff.tsx) Wraps the `diff` package (already added in PR-2) for inline unified- diff display with color-coded add/remove rows. Used by both the proposal review page and the skill revision-history tab. ### New pages (src/pages/) - Dashboard.tsx — at-a-glance home. Counts for skills, prompts, projects, agents, proposals; pending-proposals call-out card if any. - Skills.tsx — list view, separated into Global vs Project/Agent- scoped sections. - SkillDetail.tsx — name + semver + description; tabs for SKILL.md / Files / Metadata / History. History tab shows revisions with click-to-diff against the live body. - Proposals.tsx — queue with Pending/Approved/Rejected tabs. Pending count is highlighted in amber. - ProposalDetail.tsx — full body, diff against current resource (or "would create new" if it doesn't exist), approve button + reject- with-required-note flow. ### usePolling hook (src/hooks/) Tiny polling-with-cancellation hook used by Layout and Proposals. ### Layout rewrite (src/components/Layout.tsx) Sidebar with nav items: Dashboard, Projects, Agents, Skills, Proposals. Lucide icons. Active-route highlighting via NavLink. Pending-proposals warning badge on the Proposals item. ### Routes (src/App.tsx) New routes: /dashboard, /skills, /skills/:name, /proposals, /proposals/:id. Default redirects to /dashboard. ### API types (src/api.ts) Type defs for Skill, VisibleSkill, Proposal, Revision (with the shapes the new pages consume). ## Tests Existing 7 web tests still pass (Login + api). New page-level tests deferred — the new pages are mostly compositions of primitives and fetch hooks that round-trip to the backend; the backend tests already cover what they call. PR-7 polish can add render-and-click tests if coverage drift surfaces. ## Verification - `pnpm --filter @mcpctl/web build` clean, no warnings. - `pnpm test:run` whole monorepo: 162 test files / 2157 tests green. - Visual smoke deferred — needs a running mcpd to populate the fixtures. Manual smoke tested locally is the next step. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-07 17:54:55 +01:00
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
gensync@1.0.0-beta.2: {}
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
get-caller-file@2.0.5: {}
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
get-east-asian-width@1.5.0: {}
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
get-east-asian-width@1.6.0: {}
first commit
2026-02-21 03:10:39 +00:00
get-intrinsic@1.3.0:
dependencies:
call-bind-apply-helpers: 1.0.2
es-define-property: 1.0.1
es-errors: 1.3.0
es-object-atoms: 1.1.1
function-bind: 1.1.2
get-proto: 1.0.1
gopd: 1.2.0
has-symbols: 1.1.0
hasown: 2.0.2
math-intrinsics: 1.1.0
get-proto@1.0.1:
dependencies:
dunder-proto: 1.0.1
es-object-atoms: 1.1.1
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
get-stream@6.0.1: {}
first commit
2026-02-21 03:10:39 +00:00
get-tsconfig@4.13.6:
dependencies:
resolve-pkg-maps: 1.0.0
giget@2.0.0:
dependencies:
citty: 0.1.6
consola: 3.4.2
defu: 6.1.4
node-fetch-native: 1.6.7
nypm: 0.6.5
pathe: 2.0.3
glob-parent@6.0.2:
dependencies:
is-glob: 4.0.3
feat(mcpd+deploy): serve web UI at /ui + smoke tests + docs (Stage 6) The closing stage. mcpd now hosts the Stage 5 SPA, the Docker image bundles the build artifact, a smoke test exercises the personality HTTP surface end-to-end, and the user-facing docs spell out the mental model. mcpd: - Add @fastify/static dep. - New routes/web-ui.ts: registers /ui/* against a static bundle. Looks for the bundle at $MCPD_WEB_ROOT, then /usr/share/mcpd/web (the Docker image path), then a dev-tree fallback. Logs and skips cleanly if missing — API-only deploys keep working. - SPA fallback: any /ui/<path> that doesn't match a file falls through to index.html so direct hits to react-router URLs work. - /ui/* falls through to `kind: skip` in mapUrlToPermission, so the static assets are served unauthenticated. Each API call from the SPA still carries the bearer token. Deploy: - Dockerfile.mcpd builds the @mcpctl/web bundle in the same builder stage and copies dist/ to /usr/share/mcpd/web in the runtime image. Smoke (personality.smoke.test.ts): - Live mcpd flow: create secret/llm/agent/personality, attach an agent-direct prompt, verify the binding listing, reject double- attach (409) + foreign-agent prompt (400), set defaultPersonality by name, detach + delete cleanup. Docs: - New docs/personalities.md: VLAN-on-ethernet model, system-block ordering table, three prompt scopes, CLI walkthrough, web UI walkthrough, full API surface, RBAC notes. - agents.md and chat.md cross-link. - README's Agents section gains a Personalities subsection. Test count after Stage 6: mcpd: 801/801 cli: 430/430 web: 7/7 db: 58/62 (4 pre-existing) Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:48:43 +01:00
glob@11.1.0:
dependencies:
foreground-child: 3.3.1
jackspeak: 4.2.3
minimatch: 10.2.2
minipass: 7.1.3
package-json-from-dist: 1.0.1
path-scurry: 2.0.2
first commit
2026-02-21 03:10:39 +00:00
glob@13.0.6:
dependencies:
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
minimatch: 10.2.5
first commit
2026-02-21 03:10:39 +00:00
minipass: 7.1.3
path-scurry: 2.0.2
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
glob@7.2.3:
dependencies:
fs.realpath: 1.0.0
inflight: 1.0.6
inherits: 2.0.4
minimatch: 3.1.2
once: 1.4.0
path-is-absolute: 1.0.1
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
glob@9.3.5:
dependencies:
fs.realpath: 1.0.0
minimatch: 8.0.7
minipass: 4.2.8
path-scurry: 1.11.1
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
google-auth-library@10.9.1:
dependencies:
base64-js: 1.5.1
ecdsa-sig-formatter: 1.0.11
gaxios: 7.3.0
gcp-metadata: 8.1.2
google-logging-utils: 1.1.3
jws: 4.0.1
transitivePeerDependencies:
- supports-color
google-logging-utils@1.1.3: {}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
google-protobuf@3.21.4: {}
first commit
2026-02-21 03:10:39 +00:00
gopd@1.2.0: {}
feat(web): bold redesign — Tailwind v4 + shadcn-style primitives + Skills/Proposals/Revisions UI Phase 6 of the Skills + Revisions + Proposals work. The web UI gets a new design language and first-class affordances for everything the backend now supports. ## Visual direction - Tailwind v4 with custom @theme block (oklch tokens). Dark-mode-only (internal tool — light mode doubles QA surface). - Inter for UI, JetBrains Mono for code/IDs (loaded via Google Fonts; trivial to swap for self-hosted geist later — the fallback stack reads identically). - Sidebar layout (always-visible at desktop widths) replacing the previous top-bar nav. Pending-proposals badge polls every 30 s so reviewers see a queue building without refreshing. - Lucide icons throughout. - Spacing and radii on Tailwind defaults. Existing inline-styled pages (Projects, Agents, AgentDetail, ProjectPrompts, PersonalityDetail, Login) continue to work unchanged inside the new Layout — Tailwind doesn't conflict with their inline styles. A follow-up can migrate them incrementally. ## What's added ### Build infra (src/web/) - package.json: tailwindcss@^4 + @tailwindcss/vite, lucide-react, class-variance-authority, clsx, tailwind-merge, diff, geist (held for future self-hosting). - vite.config.ts: registers the @tailwindcss/vite plugin. - src/index.css: Tailwind import + @theme tokens + @layer base. - src/main.tsx: imports index.css. - src/lib/utils.ts: shadcn-style cn() helper. ### shadcn-style primitives (src/components/ui/) Hand-written rather than generated via `npx shadcn` so the repo doesn't depend on a CLI tool that needs interactive runtime: - button.tsx — variants: primary / secondary / ghost / danger / link; sizes: sm / md / lg / icon. - card.tsx — Card + Header/Title/Description/Content/Footer subparts. - badge.tsx — variants: default / info / success / warning / danger / outline. - input.tsx — Input + Textarea + Label. - tabs.tsx — no-dep accessible Tabs (no Radix needed for our use). - separator.tsx — h/v separator with role=separator. ### Diff component (src/components/Diff.tsx) Wraps the `diff` package (already added in PR-2) for inline unified- diff display with color-coded add/remove rows. Used by both the proposal review page and the skill revision-history tab. ### New pages (src/pages/) - Dashboard.tsx — at-a-glance home. Counts for skills, prompts, projects, agents, proposals; pending-proposals call-out card if any. - Skills.tsx — list view, separated into Global vs Project/Agent- scoped sections. - SkillDetail.tsx — name + semver + description; tabs for SKILL.md / Files / Metadata / History. History tab shows revisions with click-to-diff against the live body. - Proposals.tsx — queue with Pending/Approved/Rejected tabs. Pending count is highlighted in amber. - ProposalDetail.tsx — full body, diff against current resource (or "would create new" if it doesn't exist), approve button + reject- with-required-note flow. ### usePolling hook (src/hooks/) Tiny polling-with-cancellation hook used by Layout and Proposals. ### Layout rewrite (src/components/Layout.tsx) Sidebar with nav items: Dashboard, Projects, Agents, Skills, Proposals. Lucide icons. Active-route highlighting via NavLink. Pending-proposals warning badge on the Proposals item. ### Routes (src/App.tsx) New routes: /dashboard, /skills, /skills/:name, /proposals, /proposals/:id. Default redirects to /dashboard. ### API types (src/api.ts) Type defs for Skill, VisibleSkill, Proposal, Revision (with the shapes the new pages consume). ## Tests Existing 7 web tests still pass (Login + api). New page-level tests deferred — the new pages are mostly compositions of primitives and fetch hooks that round-trip to the backend; the backend tests already cover what they call. PR-7 polish can add render-and-click tests if coverage drift surfaces. ## Verification - `pnpm --filter @mcpctl/web build` clean, no warnings. - `pnpm test:run` whole monorepo: 162 test files / 2157 tests green. - Visual smoke deferred — needs a running mcpd to populate the fixtures. Manual smoke tested locally is the next step. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-07 17:54:55 +01:00
graceful-fs@4.2.11: {}
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
grok-mermaid@0.2.2: {}
first commit
2026-02-21 03:10:39 +00:00
has-flag@4.0.0: {}
has-symbols@1.1.0: {}
feat: add Kubernetes orchestrator for MCP server pod management mcpd can now deploy MCP server instances as Kubernetes pods instead of Docker containers. Set MCPD_ORCHESTRATOR=kubernetes to enable. - Add @kubernetes/client-node with thin wrapper (context enforcement via MCPD_K8S_CONTEXT to prevent multi-cluster mishaps) - Rewrite KubernetesOrchestrator: pod CRUD, pod IP extraction, exec via SPDY (one-shot + interactive), log streaming - Manifest generator: stdin:true for STDIO servers, args (not command) to preserve runner image entrypoint, security hardening - Orchestrator selection in main.ts via MCPD_ORCHESTRATOR env var - 25 unit tests for k8s orchestrator, all 624 tests pass Tested end-to-end on local k3s: - mcpd deployed via Pulumi, creates pods in mcpctl-servers namespace - NetworkPolicy verified: only mcpd can reach MCP server pods - Python runner (uvx) successfully runs aws-documentation-mcp-server Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-08 01:55:13 +01:00
has-tostringtag@1.0.2:
dependencies:
has-symbols: 1.1.0
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
has-unicode@2.0.1: {}
first commit
2026-02-21 03:10:39 +00:00
hasown@2.0.2:
dependencies:
function-bind: 1.1.2
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
hasown@2.0.4:
dependencies:
function-bind: 1.1.2
first commit
2026-02-21 03:10:39 +00:00
helmet@7.2.0: {}
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
highlight.js@10.7.3: {}
first commit
2026-02-21 03:10:39 +00:00
hono@4.12.0: {}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
hosted-git-info@7.0.2:
dependencies:
lru-cache: 10.4.3
hosted-git-info@9.0.3:
dependencies:
lru-cache: 11.2.6
feat: add Kubernetes orchestrator for MCP server pod management mcpd can now deploy MCP server instances as Kubernetes pods instead of Docker containers. Set MCPD_ORCHESTRATOR=kubernetes to enable. - Add @kubernetes/client-node with thin wrapper (context enforcement via MCPD_K8S_CONTEXT to prevent multi-cluster mishaps) - Rewrite KubernetesOrchestrator: pod CRUD, pod IP extraction, exec via SPDY (one-shot + interactive), log streaming - Manifest generator: stdin:true for STDIO servers, args (not command) to preserve runner image entrypoint, security hardening - Orchestrator selection in main.ts via MCPD_ORCHESTRATOR env var - 25 unit tests for k8s orchestrator, all 624 tests pass Tested end-to-end on local k3s: - mcpd deployed via Pulumi, creates pods in mcpctl-servers namespace - NetworkPolicy verified: only mcpd can reach MCP server pods - Python runner (uvx) successfully runs aws-documentation-mcp-server Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-08 01:55:13 +01:00
hpagent@1.2.0: {}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
html-encoding-sniffer@6.0.0:
dependencies:
'@exodus/bytes': 1.15.0
transitivePeerDependencies:
- '@noble/hashes'
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
html-entities@2.3.3: {}
first commit
2026-02-21 03:10:39 +00:00
html-escaper@2.0.2: {}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
http-cache-semantics@4.2.0: {}
first commit
2026-02-21 03:10:39 +00:00
http-errors@2.0.1:
dependencies:
depd: 2.0.0
inherits: 2.0.4
setprototypeof: 1.2.0
statuses: 2.0.2
toidentifier: 1.0.1
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
http-proxy-agent@7.0.2:
dependencies:
agent-base: 7.1.4
debug: 4.4.3
transitivePeerDependencies:
- supports-color
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
https-proxy-agent@5.0.1:
dependencies:
agent-base: 6.0.2
debug: 4.4.3
transitivePeerDependencies:
- supports-color
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
https-proxy-agent@7.0.6:
dependencies:
agent-base: 7.1.4
debug: 4.4.3
transitivePeerDependencies:
- supports-color
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
human-signals@2.1.0: {}
first commit
2026-02-21 03:10:39 +00:00
iconv-lite@0.7.2:
dependencies:
safer-buffer: 2.1.2
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
ieee754@1.2.1: {}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
ignore-walk@8.0.0:
dependencies:
minimatch: 10.2.2
first commit
2026-02-21 03:10:39 +00:00
ignore@5.3.2: {}
ignore@7.0.5: {}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
import-in-the-middle@3.3.1:
dependencies:
cjs-module-lexer: 2.2.0
es-module-lexer: 2.3.1
module-details-from-path: 1.0.4
first commit
2026-02-21 03:10:39 +00:00
imurmurhash@0.1.4: {}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
indent-string@4.0.0: {}
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
indent-string@5.0.0: {}
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
inflight@1.0.6:
dependencies:
once: 1.4.0
wrappy: 1.0.2
first commit
2026-02-21 03:10:39 +00:00
inherits@2.0.4: {}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
ini@2.0.0: {}
ini@6.0.0: {}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
ini@7.0.0: {}
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
ink@6.8.0(@types/react@19.2.14)(react@19.2.4):
dependencies:
'@alcalzone/ansi-tokenize': 0.2.5
ansi-escapes: 7.3.0
ansi-styles: 6.2.3
auto-bind: 5.0.1
chalk: 5.6.2
cli-boxes: 3.0.0
cli-cursor: 4.0.0
cli-truncate: 5.1.1
code-excerpt: 4.0.0
es-toolkit: 1.44.0
indent-string: 5.0.0
is-in-ci: 2.0.0
patch-console: 2.0.0
react: 19.2.4
react-reconciler: 0.33.0(react@19.2.4)
scheduler: 0.27.0
signal-exit: 3.0.7
slice-ansi: 8.0.0
stack-utils: 2.0.6
string-width: 8.2.0
terminal-size: 4.0.1
type-fest: 5.4.4
widest-line: 6.0.0
wrap-ansi: 9.0.2
ws: 8.19.0
yoga-layout: 3.2.1
optionalDependencies:
'@types/react': 19.2.14
transitivePeerDependencies:
- bufferutil
- utf-8-validate
feat: implement MCP registry client with multi-source search Add registry client that queries Official, Glama, and Smithery MCP registries with caching, request deduplication, retry logic, and result ranking/dedup. Includes 53 tests covering all components. Also fix null priority values in cancelled tasks (19-21) that broke Task Master, and add new tasks 25-27 for registry completion and CLI discover/install commands. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 03:46:14 +00:00
inquirer@12.11.1(@types/node@25.3.0):
first commit
2026-02-21 03:10:39 +00:00
dependencies:
'@inquirer/ansi': 1.0.2
feat: implement MCP registry client with multi-source search Add registry client that queries Official, Glama, and Smithery MCP registries with caching, request deduplication, retry logic, and result ranking/dedup. Includes 53 tests covering all components. Also fix null priority values in cancelled tasks (19-21) that broke Task Master, and add new tasks 25-27 for registry completion and CLI discover/install commands. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 03:46:14 +00:00
'@inquirer/core': 10.3.2(@types/node@25.3.0)
'@inquirer/prompts': 7.10.1(@types/node@25.3.0)
'@inquirer/type': 3.0.10(@types/node@25.3.0)
first commit
2026-02-21 03:10:39 +00:00
mute-stream: 2.0.0
run-async: 4.0.6
rxjs: 7.8.2
feat: implement MCP registry client with multi-source search Add registry client that queries Official, Glama, and Smithery MCP registries with caching, request deduplication, retry logic, and result ranking/dedup. Includes 53 tests covering all components. Also fix null priority values in cancelled tasks (19-21) that broke Task Master, and add new tasks 25-27 for registry completion and CLI discover/install commands. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 03:46:14 +00:00
optionalDependencies:
'@types/node': 25.3.0
first commit
2026-02-21 03:10:39 +00:00
ip-address@10.0.1: {}
ipaddr.js@1.9.1: {}
ipaddr.js@2.3.0: {}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
is-core-module@2.16.2:
dependencies:
hasown: 2.0.4
first commit
2026-02-21 03:10:39 +00:00
is-extglob@2.1.1: {}
is-fullwidth-code-point@3.0.0: {}
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
is-fullwidth-code-point@5.1.0:
dependencies:
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
get-east-asian-width: 1.6.0
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
first commit
2026-02-21 03:10:39 +00:00
is-glob@4.0.3:
dependencies:
is-extglob: 2.1.1
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
is-in-ci@2.0.0: {}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
is-potential-custom-element-name@1.0.1: {}
first commit
2026-02-21 03:10:39 +00:00
is-promise@4.0.0: {}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
is-stream@2.0.1: {}
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
is-unicode-supported@2.1.0: {}
first commit
2026-02-21 03:10:39 +00:00
isexe@2.0.0: {}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
isexe@4.0.0: {}
feat: add Kubernetes orchestrator for MCP server pod management mcpd can now deploy MCP server instances as Kubernetes pods instead of Docker containers. Set MCPD_ORCHESTRATOR=kubernetes to enable. - Add @kubernetes/client-node with thin wrapper (context enforcement via MCPD_K8S_CONTEXT to prevent multi-cluster mishaps) - Rewrite KubernetesOrchestrator: pod CRUD, pod IP extraction, exec via SPDY (one-shot + interactive), log streaming - Manifest generator: stdin:true for STDIO servers, args (not command) to preserve runner image entrypoint, security hardening - Orchestrator selection in main.ts via MCPD_ORCHESTRATOR env var - 25 unit tests for k8s orchestrator, all 624 tests pass Tested end-to-end on local k3s: - mcpd deployed via Pulumi, creates pods in mcpctl-servers namespace - NetworkPolicy verified: only mcpd can reach MCP server pods - Python runner (uvx) successfully runs aws-documentation-mcp-server Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-08 01:55:13 +01:00
isomorphic-ws@5.0.0(ws@8.19.0):
dependencies:
ws: 8.19.0
first commit
2026-02-21 03:10:39 +00:00
istanbul-lib-coverage@3.2.2: {}
istanbul-lib-report@3.0.1:
dependencies:
istanbul-lib-coverage: 3.2.2
make-dir: 4.0.0
supports-color: 7.2.0
istanbul-reports@3.2.0:
dependencies:
html-escaper: 2.0.2
istanbul-lib-report: 3.0.1
feat(mcpd+deploy): serve web UI at /ui + smoke tests + docs (Stage 6) The closing stage. mcpd now hosts the Stage 5 SPA, the Docker image bundles the build artifact, a smoke test exercises the personality HTTP surface end-to-end, and the user-facing docs spell out the mental model. mcpd: - Add @fastify/static dep. - New routes/web-ui.ts: registers /ui/* against a static bundle. Looks for the bundle at $MCPD_WEB_ROOT, then /usr/share/mcpd/web (the Docker image path), then a dev-tree fallback. Logs and skips cleanly if missing — API-only deploys keep working. - SPA fallback: any /ui/<path> that doesn't match a file falls through to index.html so direct hits to react-router URLs work. - /ui/* falls through to `kind: skip` in mapUrlToPermission, so the static assets are served unauthenticated. Each API call from the SPA still carries the bearer token. Deploy: - Dockerfile.mcpd builds the @mcpctl/web bundle in the same builder stage and copies dist/ to /usr/share/mcpd/web in the runtime image. Smoke (personality.smoke.test.ts): - Live mcpd flow: create secret/llm/agent/personality, attach an agent-direct prompt, verify the binding listing, reject double- attach (409) + foreign-agent prompt (400), set defaultPersonality by name, detach + delete cleanup. Docs: - New docs/personalities.md: VLAN-on-ethernet model, system-block ordering table, three prompt scopes, CLI walkthrough, web UI walkthrough, full API surface, RBAC notes. - agents.md and chat.md cross-link. - README's Agents section gains a Personalities subsection. Test count after Stage 6: mcpd: 801/801 cli: 430/430 web: 7/7 db: 58/62 (4 pre-existing) Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:48:43 +01:00
jackspeak@4.2.3:
dependencies:
'@isaacs/cliui': 9.0.0
first commit
2026-02-21 03:10:39 +00:00
jiti@2.6.1: {}
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
jiti@2.7.0: {}
first commit
2026-02-21 03:10:39 +00:00
jose@6.1.3: {}
js-tokens@10.0.0: {}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
js-tokens@4.0.0: {}
first commit
2026-02-21 03:10:39 +00:00
js-yaml@4.1.1:
dependencies:
argparse: 2.0.1
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
jsdom@28.1.0:
dependencies:
'@acemir/cssom': 0.9.31
'@asamuzakjp/dom-selector': 6.8.1
'@bramus/specificity': 2.4.2
'@exodus/bytes': 1.15.0
cssstyle: 6.2.0
data-urls: 7.0.0
decimal.js: 10.6.0
html-encoding-sniffer: 6.0.0
http-proxy-agent: 7.0.2
https-proxy-agent: 7.0.6
is-potential-custom-element-name: 1.0.1
parse5: 8.0.1
saxes: 6.0.0
symbol-tree: 3.2.4
tough-cookie: 6.0.1
undici: 7.25.0
w3c-xmlserializer: 5.0.0
webidl-conversions: 8.0.1
whatwg-mimetype: 5.0.0
whatwg-url: 16.0.1
xml-name-validator: 5.0.0
transitivePeerDependencies:
- '@noble/hashes'
- supports-color
feat: add Kubernetes orchestrator for MCP server pod management mcpd can now deploy MCP server instances as Kubernetes pods instead of Docker containers. Set MCPD_ORCHESTRATOR=kubernetes to enable. - Add @kubernetes/client-node with thin wrapper (context enforcement via MCPD_K8S_CONTEXT to prevent multi-cluster mishaps) - Rewrite KubernetesOrchestrator: pod CRUD, pod IP extraction, exec via SPDY (one-shot + interactive), log streaming - Manifest generator: stdin:true for STDIO servers, args (not command) to preserve runner image entrypoint, security hardening - Orchestrator selection in main.ts via MCPD_ORCHESTRATOR env var - 25 unit tests for k8s orchestrator, all 624 tests pass Tested end-to-end on local k3s: - mcpd deployed via Pulumi, creates pods in mcpctl-servers namespace - NetworkPolicy verified: only mcpd can reach MCP server pods - Python runner (uvx) successfully runs aws-documentation-mcp-server Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-08 01:55:13 +01:00
jsep@1.4.0: {}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
jsesc@3.1.0: {}
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
json-bigint@1.0.0:
dependencies:
bignumber.js: 9.3.1
first commit
2026-02-21 03:10:39 +00:00
json-buffer@3.0.1: {}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
json-parse-even-better-errors@5.0.0: {}
first commit
2026-02-21 03:10:39 +00:00
json-schema-ref-resolver@3.0.0:
dependencies:
dequal: 2.0.3
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
json-schema-to-ts@3.1.1:
dependencies:
'@babel/runtime': 7.29.2
ts-algebra: 2.0.0
first commit
2026-02-21 03:10:39 +00:00
json-schema-traverse@0.4.1: {}
json-schema-traverse@1.0.0: {}
json-schema-typed@8.0.2: {}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
json-schema@0.4.0: {}
first commit
2026-02-21 03:10:39 +00:00
json-stable-stringify-without-jsonify@1.0.1: {}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
json-stringify-nice@1.1.4: {}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
json5@2.2.3: {}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
jsonparse@1.3.1: {}
feat: add Kubernetes orchestrator for MCP server pod management mcpd can now deploy MCP server instances as Kubernetes pods instead of Docker containers. Set MCPD_ORCHESTRATOR=kubernetes to enable. - Add @kubernetes/client-node with thin wrapper (context enforcement via MCPD_K8S_CONTEXT to prevent multi-cluster mishaps) - Rewrite KubernetesOrchestrator: pod CRUD, pod IP extraction, exec via SPDY (one-shot + interactive), log streaming - Manifest generator: stdin:true for STDIO servers, args (not command) to preserve runner image entrypoint, security hardening - Orchestrator selection in main.ts via MCPD_ORCHESTRATOR env var - 25 unit tests for k8s orchestrator, all 624 tests pass Tested end-to-end on local k3s: - mcpd deployed via Pulumi, creates pods in mcpctl-servers namespace - NetworkPolicy verified: only mcpd can reach MCP server pods - Python runner (uvx) successfully runs aws-documentation-mcp-server Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-08 01:55:13 +01:00
jsonpath-plus@10.4.0:
dependencies:
'@jsep-plugin/assignment': 1.3.0(jsep@1.4.0)
'@jsep-plugin/regex': 1.0.4(jsep@1.4.0)
jsep: 1.4.0
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
just-diff-apply@5.5.0: {}
just-diff@6.0.2: {}
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
jwa@2.0.1:
dependencies:
buffer-equal-constant-time: 1.0.1
ecdsa-sig-formatter: 1.0.11
safe-buffer: 5.2.1
jws@4.0.1:
dependencies:
jwa: 2.0.1
safe-buffer: 5.2.1
first commit
2026-02-21 03:10:39 +00:00
keyv@4.5.4:
dependencies:
json-buffer: 3.0.1
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
kubernetes-types@1.30.0: {}
first commit
2026-02-21 03:10:39 +00:00
levn@0.4.1:
dependencies:
prelude-ls: 1.2.1
type-check: 0.4.0
light-my-request@6.6.0:
dependencies:
cookie: 1.1.1
process-warning: 4.0.1
set-cookie-parser: 2.7.2
feat(web): bold redesign — Tailwind v4 + shadcn-style primitives + Skills/Proposals/Revisions UI Phase 6 of the Skills + Revisions + Proposals work. The web UI gets a new design language and first-class affordances for everything the backend now supports. ## Visual direction - Tailwind v4 with custom @theme block (oklch tokens). Dark-mode-only (internal tool — light mode doubles QA surface). - Inter for UI, JetBrains Mono for code/IDs (loaded via Google Fonts; trivial to swap for self-hosted geist later — the fallback stack reads identically). - Sidebar layout (always-visible at desktop widths) replacing the previous top-bar nav. Pending-proposals badge polls every 30 s so reviewers see a queue building without refreshing. - Lucide icons throughout. - Spacing and radii on Tailwind defaults. Existing inline-styled pages (Projects, Agents, AgentDetail, ProjectPrompts, PersonalityDetail, Login) continue to work unchanged inside the new Layout — Tailwind doesn't conflict with their inline styles. A follow-up can migrate them incrementally. ## What's added ### Build infra (src/web/) - package.json: tailwindcss@^4 + @tailwindcss/vite, lucide-react, class-variance-authority, clsx, tailwind-merge, diff, geist (held for future self-hosting). - vite.config.ts: registers the @tailwindcss/vite plugin. - src/index.css: Tailwind import + @theme tokens + @layer base. - src/main.tsx: imports index.css. - src/lib/utils.ts: shadcn-style cn() helper. ### shadcn-style primitives (src/components/ui/) Hand-written rather than generated via `npx shadcn` so the repo doesn't depend on a CLI tool that needs interactive runtime: - button.tsx — variants: primary / secondary / ghost / danger / link; sizes: sm / md / lg / icon. - card.tsx — Card + Header/Title/Description/Content/Footer subparts. - badge.tsx — variants: default / info / success / warning / danger / outline. - input.tsx — Input + Textarea + Label. - tabs.tsx — no-dep accessible Tabs (no Radix needed for our use). - separator.tsx — h/v separator with role=separator. ### Diff component (src/components/Diff.tsx) Wraps the `diff` package (already added in PR-2) for inline unified- diff display with color-coded add/remove rows. Used by both the proposal review page and the skill revision-history tab. ### New pages (src/pages/) - Dashboard.tsx — at-a-glance home. Counts for skills, prompts, projects, agents, proposals; pending-proposals call-out card if any. - Skills.tsx — list view, separated into Global vs Project/Agent- scoped sections. - SkillDetail.tsx — name + semver + description; tabs for SKILL.md / Files / Metadata / History. History tab shows revisions with click-to-diff against the live body. - Proposals.tsx — queue with Pending/Approved/Rejected tabs. Pending count is highlighted in amber. - ProposalDetail.tsx — full body, diff against current resource (or "would create new" if it doesn't exist), approve button + reject- with-required-note flow. ### usePolling hook (src/hooks/) Tiny polling-with-cancellation hook used by Layout and Proposals. ### Layout rewrite (src/components/Layout.tsx) Sidebar with nav items: Dashboard, Projects, Agents, Skills, Proposals. Lucide icons. Active-route highlighting via NavLink. Pending-proposals warning badge on the Proposals item. ### Routes (src/App.tsx) New routes: /dashboard, /skills, /skills/:name, /proposals, /proposals/:id. Default redirects to /dashboard. ### API types (src/api.ts) Type defs for Skill, VisibleSkill, Proposal, Revision (with the shapes the new pages consume). ## Tests Existing 7 web tests still pass (Login + api). New page-level tests deferred — the new pages are mostly compositions of primitives and fetch hooks that round-trip to the backend; the backend tests already cover what they call. PR-7 polish can add render-and-click tests if coverage drift surfaces. ## Verification - `pnpm --filter @mcpctl/web build` clean, no warnings. - `pnpm test:run` whole monorepo: 162 test files / 2157 tests green. - Visual smoke deferred — needs a running mcpd to populate the fixtures. Manual smoke tested locally is the next step. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-07 17:54:55 +01:00
lightningcss-android-arm64@1.32.0:
optional: true
lightningcss-darwin-arm64@1.32.0:
optional: true
lightningcss-darwin-x64@1.32.0:
optional: true
lightningcss-freebsd-x64@1.32.0:
optional: true
lightningcss-linux-arm-gnueabihf@1.32.0:
optional: true
lightningcss-linux-arm64-gnu@1.32.0:
optional: true
lightningcss-linux-arm64-musl@1.32.0:
optional: true
lightningcss-linux-x64-gnu@1.32.0:
optional: true
lightningcss-linux-x64-musl@1.32.0:
optional: true
lightningcss-win32-arm64-msvc@1.32.0:
optional: true
lightningcss-win32-x64-msvc@1.32.0:
optional: true
lightningcss@1.32.0:
dependencies:
detect-libc: 2.1.2
optionalDependencies:
lightningcss-android-arm64: 1.32.0
lightningcss-darwin-arm64: 1.32.0
lightningcss-darwin-x64: 1.32.0
lightningcss-freebsd-x64: 1.32.0
lightningcss-linux-arm-gnueabihf: 1.32.0
lightningcss-linux-arm64-gnu: 1.32.0
lightningcss-linux-arm64-musl: 1.32.0
lightningcss-linux-x64-gnu: 1.32.0
lightningcss-linux-x64-musl: 1.32.0
lightningcss-win32-arm64-msvc: 1.32.0
lightningcss-win32-x64-msvc: 1.32.0
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
locate-path@3.0.0:
dependencies:
p-locate: 3.0.0
path-exists: 3.0.0
first commit
2026-02-21 03:10:39 +00:00
locate-path@6.0.0:
dependencies:
p-locate: 5.0.0
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
lodash.camelcase@4.3.0: {}
long@5.3.2: {}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
lru-cache@10.4.3: {}
first commit
2026-02-21 03:10:39 +00:00
lru-cache@11.2.6: {}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
lru-cache@5.1.1:
dependencies:
yallist: 3.1.1
feat(web): bold redesign — Tailwind v4 + shadcn-style primitives + Skills/Proposals/Revisions UI Phase 6 of the Skills + Revisions + Proposals work. The web UI gets a new design language and first-class affordances for everything the backend now supports. ## Visual direction - Tailwind v4 with custom @theme block (oklch tokens). Dark-mode-only (internal tool — light mode doubles QA surface). - Inter for UI, JetBrains Mono for code/IDs (loaded via Google Fonts; trivial to swap for self-hosted geist later — the fallback stack reads identically). - Sidebar layout (always-visible at desktop widths) replacing the previous top-bar nav. Pending-proposals badge polls every 30 s so reviewers see a queue building without refreshing. - Lucide icons throughout. - Spacing and radii on Tailwind defaults. Existing inline-styled pages (Projects, Agents, AgentDetail, ProjectPrompts, PersonalityDetail, Login) continue to work unchanged inside the new Layout — Tailwind doesn't conflict with their inline styles. A follow-up can migrate them incrementally. ## What's added ### Build infra (src/web/) - package.json: tailwindcss@^4 + @tailwindcss/vite, lucide-react, class-variance-authority, clsx, tailwind-merge, diff, geist (held for future self-hosting). - vite.config.ts: registers the @tailwindcss/vite plugin. - src/index.css: Tailwind import + @theme tokens + @layer base. - src/main.tsx: imports index.css. - src/lib/utils.ts: shadcn-style cn() helper. ### shadcn-style primitives (src/components/ui/) Hand-written rather than generated via `npx shadcn` so the repo doesn't depend on a CLI tool that needs interactive runtime: - button.tsx — variants: primary / secondary / ghost / danger / link; sizes: sm / md / lg / icon. - card.tsx — Card + Header/Title/Description/Content/Footer subparts. - badge.tsx — variants: default / info / success / warning / danger / outline. - input.tsx — Input + Textarea + Label. - tabs.tsx — no-dep accessible Tabs (no Radix needed for our use). - separator.tsx — h/v separator with role=separator. ### Diff component (src/components/Diff.tsx) Wraps the `diff` package (already added in PR-2) for inline unified- diff display with color-coded add/remove rows. Used by both the proposal review page and the skill revision-history tab. ### New pages (src/pages/) - Dashboard.tsx — at-a-glance home. Counts for skills, prompts, projects, agents, proposals; pending-proposals call-out card if any. - Skills.tsx — list view, separated into Global vs Project/Agent- scoped sections. - SkillDetail.tsx — name + semver + description; tabs for SKILL.md / Files / Metadata / History. History tab shows revisions with click-to-diff against the live body. - Proposals.tsx — queue with Pending/Approved/Rejected tabs. Pending count is highlighted in amber. - ProposalDetail.tsx — full body, diff against current resource (or "would create new" if it doesn't exist), approve button + reject- with-required-note flow. ### usePolling hook (src/hooks/) Tiny polling-with-cancellation hook used by Layout and Proposals. ### Layout rewrite (src/components/Layout.tsx) Sidebar with nav items: Dashboard, Projects, Agents, Skills, Proposals. Lucide icons. Active-route highlighting via NavLink. Pending-proposals warning badge on the Proposals item. ### Routes (src/App.tsx) New routes: /dashboard, /skills, /skills/:name, /proposals, /proposals/:id. Default redirects to /dashboard. ### API types (src/api.ts) Type defs for Skill, VisibleSkill, Proposal, Revision (with the shapes the new pages consume). ## Tests Existing 7 web tests still pass (Login + api). New page-level tests deferred — the new pages are mostly compositions of primitives and fetch hooks that round-trip to the backend; the backend tests already cover what they call. PR-7 polish can add render-and-click tests if coverage drift surfaces. ## Verification - `pnpm --filter @mcpctl/web build` clean, no warnings. - `pnpm test:run` whole monorepo: 162 test files / 2157 tests green. - Visual smoke deferred — needs a running mcpd to populate the fixtures. Manual smoke tested locally is the next step. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-07 17:54:55 +01:00
lucide-react@0.487.0(react@19.2.5):
dependencies:
react: 19.2.5
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
lz-string@1.5.0: {}
first commit
2026-02-21 03:10:39 +00:00
magic-string@0.30.21:
dependencies:
'@jridgewell/sourcemap-codec': 1.5.5
magicast@0.5.2:
dependencies:
'@babel/parser': 7.29.0
'@babel/types': 7.29.0
source-map-js: 1.2.1
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
make-dir@3.1.0:
dependencies:
semver: 6.3.1
first commit
2026-02-21 03:10:39 +00:00
make-dir@4.0.0:
dependencies:
semver: 7.7.4
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
make-fetch-happen@15.0.6:
dependencies:
'@gar/promise-retry': 1.0.3
'@npmcli/agent': 4.0.2
'@npmcli/redact': 4.0.0
cacache: 20.0.4
http-cache-semantics: 4.2.0
minipass: 7.1.3
minipass-fetch: 5.0.2
minipass-flush: 1.0.7
minipass-pipeline: 1.2.4
negotiator: 1.0.0
proc-log: 6.1.0
ssri: 13.0.1
transitivePeerDependencies:
- supports-color
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
marked@14.0.0: {}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
marked@17.0.1: {}
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
marked@18.0.5: {}
first commit
2026-02-21 03:10:39 +00:00
math-intrinsics@1.1.0: {}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
mdn-data@2.27.1: {}
first commit
2026-02-21 03:10:39 +00:00
media-typer@1.1.0: {}
merge-descriptors@2.0.0: {}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
merge-stream@2.0.0: {}
feat: add Kubernetes orchestrator for MCP server pod management mcpd can now deploy MCP server instances as Kubernetes pods instead of Docker containers. Set MCPD_ORCHESTRATOR=kubernetes to enable. - Add @kubernetes/client-node with thin wrapper (context enforcement via MCPD_K8S_CONTEXT to prevent multi-cluster mishaps) - Rewrite KubernetesOrchestrator: pod CRUD, pod IP extraction, exec via SPDY (one-shot + interactive), log streaming - Manifest generator: stdin:true for STDIO servers, args (not command) to preserve runner image entrypoint, security hardening - Orchestrator selection in main.ts via MCPD_ORCHESTRATOR env var - 25 unit tests for k8s orchestrator, all 624 tests pass Tested end-to-end on local k3s: - mcpd deployed via Pulumi, creates pods in mcpctl-servers namespace - NetworkPolicy verified: only mcpd can reach MCP server pods - Python runner (uvx) successfully runs aws-documentation-mcp-server Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-08 01:55:13 +01:00
mime-db@1.52.0: {}
first commit
2026-02-21 03:10:39 +00:00
mime-db@1.54.0: {}
feat: add Kubernetes orchestrator for MCP server pod management mcpd can now deploy MCP server instances as Kubernetes pods instead of Docker containers. Set MCPD_ORCHESTRATOR=kubernetes to enable. - Add @kubernetes/client-node with thin wrapper (context enforcement via MCPD_K8S_CONTEXT to prevent multi-cluster mishaps) - Rewrite KubernetesOrchestrator: pod CRUD, pod IP extraction, exec via SPDY (one-shot + interactive), log streaming - Manifest generator: stdin:true for STDIO servers, args (not command) to preserve runner image entrypoint, security hardening - Orchestrator selection in main.ts via MCPD_ORCHESTRATOR env var - 25 unit tests for k8s orchestrator, all 624 tests pass Tested end-to-end on local k3s: - mcpd deployed via Pulumi, creates pods in mcpctl-servers namespace - NetworkPolicy verified: only mcpd can reach MCP server pods - Python runner (uvx) successfully runs aws-documentation-mcp-server Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-08 01:55:13 +01:00
mime-types@2.1.35:
dependencies:
mime-db: 1.52.0
first commit
2026-02-21 03:10:39 +00:00
mime-types@3.0.2:
dependencies:
mime-db: 1.54.0
feat(mcpd+deploy): serve web UI at /ui + smoke tests + docs (Stage 6) The closing stage. mcpd now hosts the Stage 5 SPA, the Docker image bundles the build artifact, a smoke test exercises the personality HTTP surface end-to-end, and the user-facing docs spell out the mental model. mcpd: - Add @fastify/static dep. - New routes/web-ui.ts: registers /ui/* against a static bundle. Looks for the bundle at $MCPD_WEB_ROOT, then /usr/share/mcpd/web (the Docker image path), then a dev-tree fallback. Logs and skips cleanly if missing — API-only deploys keep working. - SPA fallback: any /ui/<path> that doesn't match a file falls through to index.html so direct hits to react-router URLs work. - /ui/* falls through to `kind: skip` in mapUrlToPermission, so the static assets are served unauthenticated. Each API call from the SPA still carries the bearer token. Deploy: - Dockerfile.mcpd builds the @mcpctl/web bundle in the same builder stage and copies dist/ to /usr/share/mcpd/web in the runtime image. Smoke (personality.smoke.test.ts): - Live mcpd flow: create secret/llm/agent/personality, attach an agent-direct prompt, verify the binding listing, reject double- attach (409) + foreign-agent prompt (400), set defaultPersonality by name, detach + delete cleanup. Docs: - New docs/personalities.md: VLAN-on-ethernet model, system-block ordering table, three prompt scopes, CLI walkthrough, web UI walkthrough, full API surface, RBAC notes. - agents.md and chat.md cross-link. - README's Agents section gains a Personalities subsection. Test count after Stage 6: mcpd: 801/801 cli: 430/430 web: 7/7 db: 58/62 (4 pre-existing) Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:48:43 +01:00
mime@3.0.0: {}
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
mimic-fn@2.1.0: {}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
min-indent@1.0.1: {}
first commit
2026-02-21 03:10:39 +00:00
minimatch@10.2.2:
dependencies:
brace-expansion: 5.0.2
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
minimatch@10.2.5:
dependencies:
brace-expansion: 5.0.9
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
minimatch@3.1.2:
dependencies:
brace-expansion: 1.1.12
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
minimatch@8.0.7:
dependencies:
brace-expansion: 2.0.2
first commit
2026-02-21 03:10:39 +00:00
minimatch@9.0.5:
dependencies:
brace-expansion: 2.0.2
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
minimist@1.2.8: {}
minipass-collect@2.0.1:
dependencies:
minipass: 7.1.3
minipass-fetch@5.0.2:
dependencies:
minipass: 7.1.3
minipass-sized: 2.0.0
minizlib: 3.1.0
optionalDependencies:
iconv-lite: 0.7.2
minipass-flush@1.0.7:
dependencies:
minipass: 3.3.6
minipass-pipeline@1.2.4:
dependencies:
minipass: 3.3.6
minipass-sized@2.0.0:
dependencies:
minipass: 7.1.3
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
minipass@3.3.6:
dependencies:
yallist: 4.0.0
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
minipass@4.2.8: {}
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
minipass@5.0.0: {}
first commit
2026-02-21 03:10:39 +00:00
minipass@7.1.3: {}
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
minizlib@2.1.2:
dependencies:
minipass: 3.3.6
yallist: 4.0.0
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
minizlib@3.1.0:
dependencies:
minipass: 7.1.3
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
mkdirp-classic@0.5.3: {}
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
mkdirp@1.0.4: {}
first commit
2026-02-21 03:10:39 +00:00
mnemonist@0.40.0:
dependencies:
obliterator: 2.0.5
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
module-details-from-path@1.0.4: {}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
monaco-editor@0.55.1:
dependencies:
dompurify: 3.2.7
marked: 14.0.0
first commit
2026-02-21 03:10:39 +00:00
ms@2.1.3: {}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
msgpackr-extract@3.0.4:
dependencies:
node-gyp-build-optional-packages: 5.2.2
optionalDependencies:
'@msgpackr-extract/msgpackr-extract-darwin-arm64': 3.0.4
'@msgpackr-extract/msgpackr-extract-darwin-x64': 3.0.4
'@msgpackr-extract/msgpackr-extract-linux-arm': 3.0.4
'@msgpackr-extract/msgpackr-extract-linux-arm64': 3.0.4
'@msgpackr-extract/msgpackr-extract-linux-x64': 3.0.4
'@msgpackr-extract/msgpackr-extract-win32-x64': 3.0.4
optional: true
msgpackr@2.0.5:
optionalDependencies:
msgpackr-extract: 3.0.4
multipasta@0.2.8: {}
first commit
2026-02-21 03:10:39 +00:00
mute-stream@2.0.0: {}
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
nan@2.25.0:
optional: true
first commit
2026-02-21 03:10:39 +00:00
nanoid@3.3.11: {}
natural-compare@1.4.0: {}
negotiator@1.0.0: {}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
next@16.2.5(@babel/core@7.29.0)(@opentelemetry/api@1.9.1)(react-dom@19.2.5(react@19.2.5))(react@19.2.5):
feat(web): bold redesign — Tailwind v4 + shadcn-style primitives + Skills/Proposals/Revisions UI Phase 6 of the Skills + Revisions + Proposals work. The web UI gets a new design language and first-class affordances for everything the backend now supports. ## Visual direction - Tailwind v4 with custom @theme block (oklch tokens). Dark-mode-only (internal tool — light mode doubles QA surface). - Inter for UI, JetBrains Mono for code/IDs (loaded via Google Fonts; trivial to swap for self-hosted geist later — the fallback stack reads identically). - Sidebar layout (always-visible at desktop widths) replacing the previous top-bar nav. Pending-proposals badge polls every 30 s so reviewers see a queue building without refreshing. - Lucide icons throughout. - Spacing and radii on Tailwind defaults. Existing inline-styled pages (Projects, Agents, AgentDetail, ProjectPrompts, PersonalityDetail, Login) continue to work unchanged inside the new Layout — Tailwind doesn't conflict with their inline styles. A follow-up can migrate them incrementally. ## What's added ### Build infra (src/web/) - package.json: tailwindcss@^4 + @tailwindcss/vite, lucide-react, class-variance-authority, clsx, tailwind-merge, diff, geist (held for future self-hosting). - vite.config.ts: registers the @tailwindcss/vite plugin. - src/index.css: Tailwind import + @theme tokens + @layer base. - src/main.tsx: imports index.css. - src/lib/utils.ts: shadcn-style cn() helper. ### shadcn-style primitives (src/components/ui/) Hand-written rather than generated via `npx shadcn` so the repo doesn't depend on a CLI tool that needs interactive runtime: - button.tsx — variants: primary / secondary / ghost / danger / link; sizes: sm / md / lg / icon. - card.tsx — Card + Header/Title/Description/Content/Footer subparts. - badge.tsx — variants: default / info / success / warning / danger / outline. - input.tsx — Input + Textarea + Label. - tabs.tsx — no-dep accessible Tabs (no Radix needed for our use). - separator.tsx — h/v separator with role=separator. ### Diff component (src/components/Diff.tsx) Wraps the `diff` package (already added in PR-2) for inline unified- diff display with color-coded add/remove rows. Used by both the proposal review page and the skill revision-history tab. ### New pages (src/pages/) - Dashboard.tsx — at-a-glance home. Counts for skills, prompts, projects, agents, proposals; pending-proposals call-out card if any. - Skills.tsx — list view, separated into Global vs Project/Agent- scoped sections. - SkillDetail.tsx — name + semver + description; tabs for SKILL.md / Files / Metadata / History. History tab shows revisions with click-to-diff against the live body. - Proposals.tsx — queue with Pending/Approved/Rejected tabs. Pending count is highlighted in amber. - ProposalDetail.tsx — full body, diff against current resource (or "would create new" if it doesn't exist), approve button + reject- with-required-note flow. ### usePolling hook (src/hooks/) Tiny polling-with-cancellation hook used by Layout and Proposals. ### Layout rewrite (src/components/Layout.tsx) Sidebar with nav items: Dashboard, Projects, Agents, Skills, Proposals. Lucide icons. Active-route highlighting via NavLink. Pending-proposals warning badge on the Proposals item. ### Routes (src/App.tsx) New routes: /dashboard, /skills, /skills/:name, /proposals, /proposals/:id. Default redirects to /dashboard. ### API types (src/api.ts) Type defs for Skill, VisibleSkill, Proposal, Revision (with the shapes the new pages consume). ## Tests Existing 7 web tests still pass (Login + api). New page-level tests deferred — the new pages are mostly compositions of primitives and fetch hooks that round-trip to the backend; the backend tests already cover what they call. PR-7 polish can add render-and-click tests if coverage drift surfaces. ## Verification - `pnpm --filter @mcpctl/web build` clean, no warnings. - `pnpm test:run` whole monorepo: 162 test files / 2157 tests green. - Visual smoke deferred — needs a running mcpd to populate the fixtures. Manual smoke tested locally is the next step. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-07 17:54:55 +01:00
dependencies:
'@next/env': 16.2.5
'@swc/helpers': 0.5.15
baseline-browser-mapping: 2.10.23
caniuse-lite: 1.0.30001791
postcss: 8.4.31
react: 19.2.5
react-dom: 19.2.5(react@19.2.5)
styled-jsx: 5.1.6(@babel/core@7.29.0)(react@19.2.5)
optionalDependencies:
'@next/swc-darwin-arm64': 16.2.5
'@next/swc-darwin-x64': 16.2.5
'@next/swc-linux-arm64-gnu': 16.2.5
'@next/swc-linux-arm64-musl': 16.2.5
'@next/swc-linux-x64-gnu': 16.2.5
'@next/swc-linux-x64-musl': 16.2.5
'@next/swc-win32-arm64-msvc': 16.2.5
'@next/swc-win32-x64-msvc': 16.2.5
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
'@opentelemetry/api': 1.9.1
feat(web): bold redesign — Tailwind v4 + shadcn-style primitives + Skills/Proposals/Revisions UI Phase 6 of the Skills + Revisions + Proposals work. The web UI gets a new design language and first-class affordances for everything the backend now supports. ## Visual direction - Tailwind v4 with custom @theme block (oklch tokens). Dark-mode-only (internal tool — light mode doubles QA surface). - Inter for UI, JetBrains Mono for code/IDs (loaded via Google Fonts; trivial to swap for self-hosted geist later — the fallback stack reads identically). - Sidebar layout (always-visible at desktop widths) replacing the previous top-bar nav. Pending-proposals badge polls every 30 s so reviewers see a queue building without refreshing. - Lucide icons throughout. - Spacing and radii on Tailwind defaults. Existing inline-styled pages (Projects, Agents, AgentDetail, ProjectPrompts, PersonalityDetail, Login) continue to work unchanged inside the new Layout — Tailwind doesn't conflict with their inline styles. A follow-up can migrate them incrementally. ## What's added ### Build infra (src/web/) - package.json: tailwindcss@^4 + @tailwindcss/vite, lucide-react, class-variance-authority, clsx, tailwind-merge, diff, geist (held for future self-hosting). - vite.config.ts: registers the @tailwindcss/vite plugin. - src/index.css: Tailwind import + @theme tokens + @layer base. - src/main.tsx: imports index.css. - src/lib/utils.ts: shadcn-style cn() helper. ### shadcn-style primitives (src/components/ui/) Hand-written rather than generated via `npx shadcn` so the repo doesn't depend on a CLI tool that needs interactive runtime: - button.tsx — variants: primary / secondary / ghost / danger / link; sizes: sm / md / lg / icon. - card.tsx — Card + Header/Title/Description/Content/Footer subparts. - badge.tsx — variants: default / info / success / warning / danger / outline. - input.tsx — Input + Textarea + Label. - tabs.tsx — no-dep accessible Tabs (no Radix needed for our use). - separator.tsx — h/v separator with role=separator. ### Diff component (src/components/Diff.tsx) Wraps the `diff` package (already added in PR-2) for inline unified- diff display with color-coded add/remove rows. Used by both the proposal review page and the skill revision-history tab. ### New pages (src/pages/) - Dashboard.tsx — at-a-glance home. Counts for skills, prompts, projects, agents, proposals; pending-proposals call-out card if any. - Skills.tsx — list view, separated into Global vs Project/Agent- scoped sections. - SkillDetail.tsx — name + semver + description; tabs for SKILL.md / Files / Metadata / History. History tab shows revisions with click-to-diff against the live body. - Proposals.tsx — queue with Pending/Approved/Rejected tabs. Pending count is highlighted in amber. - ProposalDetail.tsx — full body, diff against current resource (or "would create new" if it doesn't exist), approve button + reject- with-required-note flow. ### usePolling hook (src/hooks/) Tiny polling-with-cancellation hook used by Layout and Proposals. ### Layout rewrite (src/components/Layout.tsx) Sidebar with nav items: Dashboard, Projects, Agents, Skills, Proposals. Lucide icons. Active-route highlighting via NavLink. Pending-proposals warning badge on the Proposals item. ### Routes (src/App.tsx) New routes: /dashboard, /skills, /skills/:name, /proposals, /proposals/:id. Default redirects to /dashboard. ### API types (src/api.ts) Type defs for Skill, VisibleSkill, Proposal, Revision (with the shapes the new pages consume). ## Tests Existing 7 web tests still pass (Login + api). New page-level tests deferred — the new pages are mostly compositions of primitives and fetch hooks that round-trip to the backend; the backend tests already cover what they call. PR-7 polish can add render-and-click tests if coverage drift surfaces. ## Verification - `pnpm --filter @mcpctl/web build` clean, no warnings. - `pnpm test:run` whole monorepo: 162 test files / 2157 tests green. - Visual smoke deferred — needs a running mcpd to populate the fixtures. Manual smoke tested locally is the next step. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-07 17:54:55 +01:00
sharp: 0.34.5
transitivePeerDependencies:
- '@babel/core'
- babel-plugin-macros
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
node-addon-api@5.1.0: {}
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
node-domexception@1.0.0: {}
first commit
2026-02-21 03:10:39 +00:00
node-fetch-native@1.6.7: {}
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
node-fetch@2.7.0:
dependencies:
whatwg-url: 5.0.0
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
node-fetch@3.3.2:
dependencies:
data-uri-to-buffer: 4.0.1
fetch-blob: 3.2.0
formdata-polyfill: 4.0.10
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
node-gyp-build-optional-packages@5.2.2:
dependencies:
detect-libc: 2.1.2
optional: true
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
node-gyp@12.4.0:
dependencies:
env-paths: 2.2.1
exponential-backoff: 3.1.3
graceful-fs: 4.2.11
nopt: 9.0.0
proc-log: 6.1.0
semver: 7.7.4
tar: 7.5.20
tinyglobby: 0.2.15
undici: 6.27.0
which: 6.0.1
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
node-releases@2.0.38: {}
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
nopt@5.0.0:
dependencies:
abbrev: 1.1.1
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
nopt@9.0.0:
dependencies:
abbrev: 4.0.0
normalize-package-data@6.0.2:
dependencies:
hosted-git-info: 7.0.2
semver: 7.7.4
validate-npm-package-license: 3.0.4
npm-bundled@5.0.0:
dependencies:
npm-normalize-package-bin: 5.0.0
npm-install-checks@8.0.0:
dependencies:
semver: 7.7.4
npm-normalize-package-bin@5.0.0: {}
npm-package-arg@13.0.2:
dependencies:
hosted-git-info: 9.0.3
proc-log: 6.1.0
semver: 7.7.4
validate-npm-package-name: 7.0.2
npm-packlist@10.0.4:
dependencies:
ignore-walk: 8.0.0
proc-log: 6.1.0
npm-pick-manifest@11.0.3:
dependencies:
npm-install-checks: 8.0.0
npm-normalize-package-bin: 5.0.0
npm-package-arg: 13.0.2
semver: 7.7.4
npm-registry-fetch@19.1.1:
dependencies:
'@npmcli/redact': 4.0.0
jsonparse: 1.3.1
make-fetch-happen: 15.0.6
minipass: 7.1.3
minipass-fetch: 5.0.2
minizlib: 3.1.0
npm-package-arg: 13.0.2
proc-log: 6.1.0
transitivePeerDependencies:
- supports-color
npm-run-path@4.0.1:
dependencies:
path-key: 3.1.1
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
npmlog@5.0.1:
dependencies:
are-we-there-yet: 2.0.0
console-control-strings: 1.1.0
gauge: 3.0.2
set-blocking: 2.0.0
first commit
2026-02-21 03:10:39 +00:00
nypm@0.6.5:
dependencies:
citty: 0.2.1
pathe: 2.0.3
tinyexec: 1.0.2
feat: add Kubernetes orchestrator for MCP server pod management mcpd can now deploy MCP server instances as Kubernetes pods instead of Docker containers. Set MCPD_ORCHESTRATOR=kubernetes to enable. - Add @kubernetes/client-node with thin wrapper (context enforcement via MCPD_K8S_CONTEXT to prevent multi-cluster mishaps) - Rewrite KubernetesOrchestrator: pod CRUD, pod IP extraction, exec via SPDY (one-shot + interactive), log streaming - Manifest generator: stdin:true for STDIO servers, args (not command) to preserve runner image entrypoint, security hardening - Orchestrator selection in main.ts via MCPD_ORCHESTRATOR env var - 25 unit tests for k8s orchestrator, all 624 tests pass Tested end-to-end on local k3s: - mcpd deployed via Pulumi, creates pods in mcpctl-servers namespace - NetworkPolicy verified: only mcpd can reach MCP server pods - Python runner (uvx) successfully runs aws-documentation-mcp-server Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-08 01:55:13 +01:00
oauth4webapi@3.8.5: {}
first commit
2026-02-21 03:10:39 +00:00
object-assign@4.1.1: {}
object-inspect@1.13.4: {}
obliterator@2.0.5: {}
obug@2.1.1: {}
ohash@2.0.11: {}
on-exit-leak-free@2.1.2: {}
on-finished@2.4.1:
dependencies:
ee-first: 1.1.1
once@1.4.0:
dependencies:
wrappy: 1.0.2
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
onetime@5.1.2:
dependencies:
mimic-fn: 2.1.0
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
openai@6.26.0(ws@8.19.0)(zod@3.25.76):
optionalDependencies:
ws: 8.19.0
zod: 3.25.76
feat: add Kubernetes orchestrator for MCP server pod management mcpd can now deploy MCP server instances as Kubernetes pods instead of Docker containers. Set MCPD_ORCHESTRATOR=kubernetes to enable. - Add @kubernetes/client-node with thin wrapper (context enforcement via MCPD_K8S_CONTEXT to prevent multi-cluster mishaps) - Rewrite KubernetesOrchestrator: pod CRUD, pod IP extraction, exec via SPDY (one-shot + interactive), log streaming - Manifest generator: stdin:true for STDIO servers, args (not command) to preserve runner image entrypoint, security hardening - Orchestrator selection in main.ts via MCPD_ORCHESTRATOR env var - 25 unit tests for k8s orchestrator, all 624 tests pass Tested end-to-end on local k3s: - mcpd deployed via Pulumi, creates pods in mcpctl-servers namespace - NetworkPolicy verified: only mcpd can reach MCP server pods - Python runner (uvx) successfully runs aws-documentation-mcp-server Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-08 01:55:13 +01:00
openid-client@6.8.2:
dependencies:
jose: 6.1.3
oauth4webapi: 3.8.5
first commit
2026-02-21 03:10:39 +00:00
optionator@0.9.4:
dependencies:
deep-is: 0.1.4
fast-levenshtein: 2.0.6
levn: 0.4.1
prelude-ls: 1.2.1
type-check: 0.4.0
word-wrap: 1.2.5
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
p-limit@2.3.0:
dependencies:
p-try: 2.2.0
first commit
2026-02-21 03:10:39 +00:00
p-limit@3.1.0:
dependencies:
yocto-queue: 0.1.0
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
p-locate@3.0.0:
dependencies:
p-limit: 2.3.0
first commit
2026-02-21 03:10:39 +00:00
p-locate@5.0.0:
dependencies:
p-limit: 3.1.0
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
p-map@7.0.5: {}
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
p-retry@4.6.2:
dependencies:
'@types/retry': 0.12.0
retry: 0.13.1
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
p-try@2.2.0: {}
first commit
2026-02-21 03:10:39 +00:00
package-json-from-dist@1.0.1: {}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
pacote@21.5.1:
dependencies:
'@gar/promise-retry': 1.0.3
'@npmcli/git': 7.0.2
'@npmcli/installed-package-contents': 4.0.0
'@npmcli/package-json': 7.0.5
'@npmcli/promise-spawn': 9.0.1
'@npmcli/run-script': 10.0.4
cacache: 20.0.4
fs-minipass: 3.0.3
minipass: 7.1.3
npm-package-arg: 13.0.2
npm-packlist: 10.0.4
npm-pick-manifest: 11.0.3
npm-registry-fetch: 19.1.1
proc-log: 6.1.0
sigstore: 4.1.1
ssri: 13.0.1
tar: 7.5.20
transitivePeerDependencies:
- supports-color
parse-conflict-json@5.0.1:
dependencies:
json-parse-even-better-errors: 5.0.0
just-diff: 6.0.2
just-diff-apply: 5.5.0
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
parse5@7.3.0:
dependencies:
entities: 6.0.1
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
parse5@8.0.1:
dependencies:
entities: 8.0.0
first commit
2026-02-21 03:10:39 +00:00
parseurl@1.3.3: {}
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
partial-json@0.1.7: {}
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
patch-console@2.0.0: {}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
path-exists@3.0.0: {}
first commit
2026-02-21 03:10:39 +00:00
path-exists@4.0.0: {}
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
path-is-absolute@1.0.1: {}
first commit
2026-02-21 03:10:39 +00:00
path-key@3.1.1: {}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
path-parse@1.0.7: {}
path-scurry@1.11.1:
dependencies:
lru-cache: 10.4.3
minipass: 7.1.3
first commit
2026-02-21 03:10:39 +00:00
path-scurry@2.0.2:
dependencies:
lru-cache: 11.2.6
minipass: 7.1.3
path-to-regexp@8.3.0: {}
pathe@2.0.3: {}
perfect-debounce@1.0.0: {}
picocolors@1.1.1: {}
picomatch@4.0.3: {}
pino-abstract-transport@3.0.0:
dependencies:
split2: 4.2.0
pino-std-serializers@7.1.0: {}
pino@10.3.1:
dependencies:
'@pinojs/redact': 0.4.0
atomic-sleep: 1.0.0
on-exit-leak-free: 2.1.2
pino-abstract-transport: 3.0.0
pino-std-serializers: 7.1.0
process-warning: 5.0.0
quick-format-unescaped: 4.0.4
real-require: 0.2.0
safe-stable-stringify: 2.5.0
sonic-boom: 4.2.1
thread-stream: 4.0.0
pkce-challenge@5.0.1: {}
pkg-types@2.3.0:
dependencies:
confbox: 0.2.4
exsolve: 1.0.8
pathe: 2.0.3
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
pkg-up@3.1.0:
dependencies:
find-up: 3.0.0
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
postcss-selector-parser@7.1.4:
dependencies:
cssesc: 3.0.0
util-deprecate: 1.0.2
feat(web): bold redesign — Tailwind v4 + shadcn-style primitives + Skills/Proposals/Revisions UI Phase 6 of the Skills + Revisions + Proposals work. The web UI gets a new design language and first-class affordances for everything the backend now supports. ## Visual direction - Tailwind v4 with custom @theme block (oklch tokens). Dark-mode-only (internal tool — light mode doubles QA surface). - Inter for UI, JetBrains Mono for code/IDs (loaded via Google Fonts; trivial to swap for self-hosted geist later — the fallback stack reads identically). - Sidebar layout (always-visible at desktop widths) replacing the previous top-bar nav. Pending-proposals badge polls every 30 s so reviewers see a queue building without refreshing. - Lucide icons throughout. - Spacing and radii on Tailwind defaults. Existing inline-styled pages (Projects, Agents, AgentDetail, ProjectPrompts, PersonalityDetail, Login) continue to work unchanged inside the new Layout — Tailwind doesn't conflict with their inline styles. A follow-up can migrate them incrementally. ## What's added ### Build infra (src/web/) - package.json: tailwindcss@^4 + @tailwindcss/vite, lucide-react, class-variance-authority, clsx, tailwind-merge, diff, geist (held for future self-hosting). - vite.config.ts: registers the @tailwindcss/vite plugin. - src/index.css: Tailwind import + @theme tokens + @layer base. - src/main.tsx: imports index.css. - src/lib/utils.ts: shadcn-style cn() helper. ### shadcn-style primitives (src/components/ui/) Hand-written rather than generated via `npx shadcn` so the repo doesn't depend on a CLI tool that needs interactive runtime: - button.tsx — variants: primary / secondary / ghost / danger / link; sizes: sm / md / lg / icon. - card.tsx — Card + Header/Title/Description/Content/Footer subparts. - badge.tsx — variants: default / info / success / warning / danger / outline. - input.tsx — Input + Textarea + Label. - tabs.tsx — no-dep accessible Tabs (no Radix needed for our use). - separator.tsx — h/v separator with role=separator. ### Diff component (src/components/Diff.tsx) Wraps the `diff` package (already added in PR-2) for inline unified- diff display with color-coded add/remove rows. Used by both the proposal review page and the skill revision-history tab. ### New pages (src/pages/) - Dashboard.tsx — at-a-glance home. Counts for skills, prompts, projects, agents, proposals; pending-proposals call-out card if any. - Skills.tsx — list view, separated into Global vs Project/Agent- scoped sections. - SkillDetail.tsx — name + semver + description; tabs for SKILL.md / Files / Metadata / History. History tab shows revisions with click-to-diff against the live body. - Proposals.tsx — queue with Pending/Approved/Rejected tabs. Pending count is highlighted in amber. - ProposalDetail.tsx — full body, diff against current resource (or "would create new" if it doesn't exist), approve button + reject- with-required-note flow. ### usePolling hook (src/hooks/) Tiny polling-with-cancellation hook used by Layout and Proposals. ### Layout rewrite (src/components/Layout.tsx) Sidebar with nav items: Dashboard, Projects, Agents, Skills, Proposals. Lucide icons. Active-route highlighting via NavLink. Pending-proposals warning badge on the Proposals item. ### Routes (src/App.tsx) New routes: /dashboard, /skills, /skills/:name, /proposals, /proposals/:id. Default redirects to /dashboard. ### API types (src/api.ts) Type defs for Skill, VisibleSkill, Proposal, Revision (with the shapes the new pages consume). ## Tests Existing 7 web tests still pass (Login + api). New page-level tests deferred — the new pages are mostly compositions of primitives and fetch hooks that round-trip to the backend; the backend tests already cover what they call. PR-7 polish can add render-and-click tests if coverage drift surfaces. ## Verification - `pnpm --filter @mcpctl/web build` clean, no warnings. - `pnpm test:run` whole monorepo: 162 test files / 2157 tests green. - Visual smoke deferred — needs a running mcpd to populate the fixtures. Manual smoke tested locally is the next step. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-07 17:54:55 +01:00
postcss@8.4.31:
dependencies:
nanoid: 3.3.11
picocolors: 1.1.1
source-map-js: 1.2.1
first commit
2026-02-21 03:10:39 +00:00
postcss@8.5.6:
dependencies:
nanoid: 3.3.11
picocolors: 1.1.1
source-map-js: 1.2.1
prelude-ls@1.2.1: {}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
pretty-format@27.5.1:
dependencies:
ansi-regex: 5.0.1
ansi-styles: 5.2.0
react-is: 17.0.2
first commit
2026-02-21 03:10:39 +00:00
prisma@6.19.2(typescript@5.9.3):
dependencies:
'@prisma/config': 6.19.2
'@prisma/engines': 6.19.2
optionalDependencies:
typescript: 5.9.3
transitivePeerDependencies:
- magicast
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
proc-log@6.1.0: {}
first commit
2026-02-21 03:10:39 +00:00
process-warning@4.0.1: {}
process-warning@5.0.0: {}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
proggy@4.0.0: {}
promise-all-reject-late@1.0.1: {}
promise-call-limit@3.0.2: {}
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
proper-lockfile@4.1.2:
dependencies:
graceful-fs: 4.2.11
retry: 0.12.0
signal-exit: 3.0.7
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
protobufjs@7.5.4:
dependencies:
'@protobufjs/aspromise': 1.1.2
'@protobufjs/base64': 1.1.2
'@protobufjs/codegen': 2.0.4
'@protobufjs/eventemitter': 1.1.0
'@protobufjs/fetch': 1.1.0
'@protobufjs/float': 1.0.2
'@protobufjs/inquire': 1.1.0
'@protobufjs/path': 1.1.2
'@protobufjs/pool': 1.1.0
'@protobufjs/utf8': 1.1.0
'@types/node': 25.3.0
long: 5.3.2
first commit
2026-02-21 03:10:39 +00:00
proxy-addr@2.0.7:
dependencies:
forwarded: 0.2.0
ipaddr.js: 1.9.1
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
pump@3.0.3:
dependencies:
end-of-stream: 1.4.5
once: 1.4.0
first commit
2026-02-21 03:10:39 +00:00
punycode@2.3.1: {}
pure-rand@6.1.0: {}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
pure-rand@8.4.2: {}
first commit
2026-02-21 03:10:39 +00:00
qs@6.15.0:
dependencies:
side-channel: 1.1.0
quick-format-unescaped@4.0.4: {}
range-parser@1.2.1: {}
raw-body@3.0.2:
dependencies:
bytes: 3.1.2
http-errors: 2.0.1
iconv-lite: 0.7.2
unpipe: 1.0.0
rc9@2.1.2:
dependencies:
defu: 6.1.4
destr: 2.0.5
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
react-dom@19.2.5(react@19.2.5):
dependencies:
react: 19.2.5
scheduler: 0.27.0
react-is@17.0.2: {}
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
react-reconciler@0.33.0(react@19.2.4):
dependencies:
react: 19.2.4
scheduler: 0.27.0
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
react-refresh@0.18.0: {}
react-router-dom@7.14.2(react-dom@19.2.5(react@19.2.5))(react@19.2.5):
dependencies:
react: 19.2.5
react-dom: 19.2.5(react@19.2.5)
react-router: 7.14.2(react-dom@19.2.5(react@19.2.5))(react@19.2.5)
react-router@7.14.2(react-dom@19.2.5(react@19.2.5))(react@19.2.5):
dependencies:
cookie: 1.1.1
react: 19.2.5
set-cookie-parser: 2.7.2
optionalDependencies:
react-dom: 19.2.5(react@19.2.5)
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
react@19.2.4: {}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
react@19.2.5: {}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
read-cmd-shim@6.0.0: {}
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
readable-stream@3.6.2:
dependencies:
inherits: 2.0.4
string_decoder: 1.3.0
util-deprecate: 1.0.2
first commit
2026-02-21 03:10:39 +00:00
readdirp@4.1.2: {}
real-require@0.2.0: {}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
redent@3.0.0:
dependencies:
indent-string: 4.0.0
strip-indent: 3.0.0
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
require-directory@2.1.1: {}
first commit
2026-02-21 03:10:39 +00:00
require-from-string@2.0.2: {}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
require-in-the-middle@8.0.1:
dependencies:
debug: 4.4.3
module-details-from-path: 1.0.4
transitivePeerDependencies:
- supports-color
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
reselect@4.1.8: {}
first commit
2026-02-21 03:10:39 +00:00
resolve-pkg-maps@1.0.0: {}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
resolve@1.22.12:
dependencies:
es-errors: 1.3.0
is-core-module: 2.16.2
path-parse: 1.0.7
supports-preserve-symlinks-flag: 1.0.0
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
restore-cursor@4.0.0:
dependencies:
onetime: 5.1.2
signal-exit: 3.0.7
first commit
2026-02-21 03:10:39 +00:00
ret@0.5.0: {}
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
retry@0.12.0: {}
retry@0.13.1: {}
first commit
2026-02-21 03:10:39 +00:00
reusify@1.1.0: {}
feat: add Kubernetes orchestrator for MCP server pod management mcpd can now deploy MCP server instances as Kubernetes pods instead of Docker containers. Set MCPD_ORCHESTRATOR=kubernetes to enable. - Add @kubernetes/client-node with thin wrapper (context enforcement via MCPD_K8S_CONTEXT to prevent multi-cluster mishaps) - Rewrite KubernetesOrchestrator: pod CRUD, pod IP extraction, exec via SPDY (one-shot + interactive), log streaming - Manifest generator: stdin:true for STDIO servers, args (not command) to preserve runner image entrypoint, security hardening - Orchestrator selection in main.ts via MCPD_ORCHESTRATOR env var - 25 unit tests for k8s orchestrator, all 624 tests pass Tested end-to-end on local k3s: - mcpd deployed via Pulumi, creates pods in mcpctl-servers namespace - NetworkPolicy verified: only mcpd can reach MCP server pods - Python runner (uvx) successfully runs aws-documentation-mcp-server Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-08 01:55:13 +01:00
rfc4648@1.5.4: {}
first commit
2026-02-21 03:10:39 +00:00
rfdc@1.4.1: {}
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
rimraf@3.0.2:
dependencies:
glob: 7.2.3
first commit
2026-02-21 03:10:39 +00:00
rimraf@6.1.3:
dependencies:
glob: 13.0.6
package-json-from-dist: 1.0.1
rollup@4.58.0:
dependencies:
'@types/estree': 1.0.8
optionalDependencies:
'@rollup/rollup-android-arm-eabi': 4.58.0
'@rollup/rollup-android-arm64': 4.58.0
'@rollup/rollup-darwin-arm64': 4.58.0
'@rollup/rollup-darwin-x64': 4.58.0
'@rollup/rollup-freebsd-arm64': 4.58.0
'@rollup/rollup-freebsd-x64': 4.58.0
'@rollup/rollup-linux-arm-gnueabihf': 4.58.0
'@rollup/rollup-linux-arm-musleabihf': 4.58.0
'@rollup/rollup-linux-arm64-gnu': 4.58.0
'@rollup/rollup-linux-arm64-musl': 4.58.0
'@rollup/rollup-linux-loong64-gnu': 4.58.0
'@rollup/rollup-linux-loong64-musl': 4.58.0
'@rollup/rollup-linux-ppc64-gnu': 4.58.0
'@rollup/rollup-linux-ppc64-musl': 4.58.0
'@rollup/rollup-linux-riscv64-gnu': 4.58.0
'@rollup/rollup-linux-riscv64-musl': 4.58.0
'@rollup/rollup-linux-s390x-gnu': 4.58.0
'@rollup/rollup-linux-x64-gnu': 4.58.0
'@rollup/rollup-linux-x64-musl': 4.58.0
'@rollup/rollup-openbsd-x64': 4.58.0
'@rollup/rollup-openharmony-arm64': 4.58.0
'@rollup/rollup-win32-arm64-msvc': 4.58.0
'@rollup/rollup-win32-ia32-msvc': 4.58.0
'@rollup/rollup-win32-x64-gnu': 4.58.0
'@rollup/rollup-win32-x64-msvc': 4.58.0
fsevents: 2.3.3
router@2.2.0:
dependencies:
debug: 4.4.3
depd: 2.0.0
is-promise: 4.0.0
parseurl: 1.3.3
path-to-regexp: 8.3.0
transitivePeerDependencies:
- supports-color
run-async@4.0.6: {}
rxjs@7.8.2:
dependencies:
tslib: 2.8.1
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
s-js@0.4.9: {}
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
safe-buffer@5.2.1: {}
first commit
2026-02-21 03:10:39 +00:00
safe-regex2@5.0.0:
dependencies:
ret: 0.5.0
safe-stable-stringify@2.5.0: {}
safer-buffer@2.1.2: {}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
saxes@6.0.0:
dependencies:
xmlchars: 2.2.0
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
scheduler@0.27.0: {}
first commit
2026-02-21 03:10:39 +00:00
secure-json-parse@4.1.0: {}
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
semver@6.3.1: {}
first commit
2026-02-21 03:10:39 +00:00
semver@7.7.4: {}
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
semver@7.8.0: {}
first commit
2026-02-21 03:10:39 +00:00
send@1.2.1:
dependencies:
debug: 4.4.3
encodeurl: 2.0.0
escape-html: 1.0.3
etag: 1.8.1
fresh: 2.0.0
http-errors: 2.0.1
mime-types: 3.0.2
ms: 2.1.3
on-finished: 2.4.1
range-parser: 1.2.1
statuses: 2.0.2
transitivePeerDependencies:
- supports-color
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
seroval-plugins@1.5.6(seroval@1.5.6):
dependencies:
seroval: 1.5.6
seroval@1.5.6: {}
first commit
2026-02-21 03:10:39 +00:00
serve-static@2.2.1:
dependencies:
encodeurl: 2.0.0
escape-html: 1.0.3
parseurl: 1.3.3
send: 1.2.1
transitivePeerDependencies:
- supports-color
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
set-blocking@2.0.0: {}
first commit
2026-02-21 03:10:39 +00:00
set-cookie-parser@2.7.2: {}
setprototypeof@1.2.0: {}
feat(web): bold redesign — Tailwind v4 + shadcn-style primitives + Skills/Proposals/Revisions UI Phase 6 of the Skills + Revisions + Proposals work. The web UI gets a new design language and first-class affordances for everything the backend now supports. ## Visual direction - Tailwind v4 with custom @theme block (oklch tokens). Dark-mode-only (internal tool — light mode doubles QA surface). - Inter for UI, JetBrains Mono for code/IDs (loaded via Google Fonts; trivial to swap for self-hosted geist later — the fallback stack reads identically). - Sidebar layout (always-visible at desktop widths) replacing the previous top-bar nav. Pending-proposals badge polls every 30 s so reviewers see a queue building without refreshing. - Lucide icons throughout. - Spacing and radii on Tailwind defaults. Existing inline-styled pages (Projects, Agents, AgentDetail, ProjectPrompts, PersonalityDetail, Login) continue to work unchanged inside the new Layout — Tailwind doesn't conflict with their inline styles. A follow-up can migrate them incrementally. ## What's added ### Build infra (src/web/) - package.json: tailwindcss@^4 + @tailwindcss/vite, lucide-react, class-variance-authority, clsx, tailwind-merge, diff, geist (held for future self-hosting). - vite.config.ts: registers the @tailwindcss/vite plugin. - src/index.css: Tailwind import + @theme tokens + @layer base. - src/main.tsx: imports index.css. - src/lib/utils.ts: shadcn-style cn() helper. ### shadcn-style primitives (src/components/ui/) Hand-written rather than generated via `npx shadcn` so the repo doesn't depend on a CLI tool that needs interactive runtime: - button.tsx — variants: primary / secondary / ghost / danger / link; sizes: sm / md / lg / icon. - card.tsx — Card + Header/Title/Description/Content/Footer subparts. - badge.tsx — variants: default / info / success / warning / danger / outline. - input.tsx — Input + Textarea + Label. - tabs.tsx — no-dep accessible Tabs (no Radix needed for our use). - separator.tsx — h/v separator with role=separator. ### Diff component (src/components/Diff.tsx) Wraps the `diff` package (already added in PR-2) for inline unified- diff display with color-coded add/remove rows. Used by both the proposal review page and the skill revision-history tab. ### New pages (src/pages/) - Dashboard.tsx — at-a-glance home. Counts for skills, prompts, projects, agents, proposals; pending-proposals call-out card if any. - Skills.tsx — list view, separated into Global vs Project/Agent- scoped sections. - SkillDetail.tsx — name + semver + description; tabs for SKILL.md / Files / Metadata / History. History tab shows revisions with click-to-diff against the live body. - Proposals.tsx — queue with Pending/Approved/Rejected tabs. Pending count is highlighted in amber. - ProposalDetail.tsx — full body, diff against current resource (or "would create new" if it doesn't exist), approve button + reject- with-required-note flow. ### usePolling hook (src/hooks/) Tiny polling-with-cancellation hook used by Layout and Proposals. ### Layout rewrite (src/components/Layout.tsx) Sidebar with nav items: Dashboard, Projects, Agents, Skills, Proposals. Lucide icons. Active-route highlighting via NavLink. Pending-proposals warning badge on the Proposals item. ### Routes (src/App.tsx) New routes: /dashboard, /skills, /skills/:name, /proposals, /proposals/:id. Default redirects to /dashboard. ### API types (src/api.ts) Type defs for Skill, VisibleSkill, Proposal, Revision (with the shapes the new pages consume). ## Tests Existing 7 web tests still pass (Login + api). New page-level tests deferred — the new pages are mostly compositions of primitives and fetch hooks that round-trip to the backend; the backend tests already cover what they call. PR-7 polish can add render-and-click tests if coverage drift surfaces. ## Verification - `pnpm --filter @mcpctl/web build` clean, no warnings. - `pnpm test:run` whole monorepo: 162 test files / 2157 tests green. - Visual smoke deferred — needs a running mcpd to populate the fixtures. Manual smoke tested locally is the next step. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-07 17:54:55 +01:00
sharp@0.34.5:
dependencies:
'@img/colour': 1.1.0
detect-libc: 2.1.2
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
semver: 7.8.0
feat(web): bold redesign — Tailwind v4 + shadcn-style primitives + Skills/Proposals/Revisions UI Phase 6 of the Skills + Revisions + Proposals work. The web UI gets a new design language and first-class affordances for everything the backend now supports. ## Visual direction - Tailwind v4 with custom @theme block (oklch tokens). Dark-mode-only (internal tool — light mode doubles QA surface). - Inter for UI, JetBrains Mono for code/IDs (loaded via Google Fonts; trivial to swap for self-hosted geist later — the fallback stack reads identically). - Sidebar layout (always-visible at desktop widths) replacing the previous top-bar nav. Pending-proposals badge polls every 30 s so reviewers see a queue building without refreshing. - Lucide icons throughout. - Spacing and radii on Tailwind defaults. Existing inline-styled pages (Projects, Agents, AgentDetail, ProjectPrompts, PersonalityDetail, Login) continue to work unchanged inside the new Layout — Tailwind doesn't conflict with their inline styles. A follow-up can migrate them incrementally. ## What's added ### Build infra (src/web/) - package.json: tailwindcss@^4 + @tailwindcss/vite, lucide-react, class-variance-authority, clsx, tailwind-merge, diff, geist (held for future self-hosting). - vite.config.ts: registers the @tailwindcss/vite plugin. - src/index.css: Tailwind import + @theme tokens + @layer base. - src/main.tsx: imports index.css. - src/lib/utils.ts: shadcn-style cn() helper. ### shadcn-style primitives (src/components/ui/) Hand-written rather than generated via `npx shadcn` so the repo doesn't depend on a CLI tool that needs interactive runtime: - button.tsx — variants: primary / secondary / ghost / danger / link; sizes: sm / md / lg / icon. - card.tsx — Card + Header/Title/Description/Content/Footer subparts. - badge.tsx — variants: default / info / success / warning / danger / outline. - input.tsx — Input + Textarea + Label. - tabs.tsx — no-dep accessible Tabs (no Radix needed for our use). - separator.tsx — h/v separator with role=separator. ### Diff component (src/components/Diff.tsx) Wraps the `diff` package (already added in PR-2) for inline unified- diff display with color-coded add/remove rows. Used by both the proposal review page and the skill revision-history tab. ### New pages (src/pages/) - Dashboard.tsx — at-a-glance home. Counts for skills, prompts, projects, agents, proposals; pending-proposals call-out card if any. - Skills.tsx — list view, separated into Global vs Project/Agent- scoped sections. - SkillDetail.tsx — name + semver + description; tabs for SKILL.md / Files / Metadata / History. History tab shows revisions with click-to-diff against the live body. - Proposals.tsx — queue with Pending/Approved/Rejected tabs. Pending count is highlighted in amber. - ProposalDetail.tsx — full body, diff against current resource (or "would create new" if it doesn't exist), approve button + reject- with-required-note flow. ### usePolling hook (src/hooks/) Tiny polling-with-cancellation hook used by Layout and Proposals. ### Layout rewrite (src/components/Layout.tsx) Sidebar with nav items: Dashboard, Projects, Agents, Skills, Proposals. Lucide icons. Active-route highlighting via NavLink. Pending-proposals warning badge on the Proposals item. ### Routes (src/App.tsx) New routes: /dashboard, /skills, /skills/:name, /proposals, /proposals/:id. Default redirects to /dashboard. ### API types (src/api.ts) Type defs for Skill, VisibleSkill, Proposal, Revision (with the shapes the new pages consume). ## Tests Existing 7 web tests still pass (Login + api). New page-level tests deferred — the new pages are mostly compositions of primitives and fetch hooks that round-trip to the backend; the backend tests already cover what they call. PR-7 polish can add render-and-click tests if coverage drift surfaces. ## Verification - `pnpm --filter @mcpctl/web build` clean, no warnings. - `pnpm test:run` whole monorepo: 162 test files / 2157 tests green. - Visual smoke deferred — needs a running mcpd to populate the fixtures. Manual smoke tested locally is the next step. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-07 17:54:55 +01:00
optionalDependencies:
'@img/sharp-darwin-arm64': 0.34.5
'@img/sharp-darwin-x64': 0.34.5
'@img/sharp-libvips-darwin-arm64': 1.2.4
'@img/sharp-libvips-darwin-x64': 1.2.4
'@img/sharp-libvips-linux-arm': 1.2.4
'@img/sharp-libvips-linux-arm64': 1.2.4
'@img/sharp-libvips-linux-ppc64': 1.2.4
'@img/sharp-libvips-linux-riscv64': 1.2.4
'@img/sharp-libvips-linux-s390x': 1.2.4
'@img/sharp-libvips-linux-x64': 1.2.4
'@img/sharp-libvips-linuxmusl-arm64': 1.2.4
'@img/sharp-libvips-linuxmusl-x64': 1.2.4
'@img/sharp-linux-arm': 0.34.5
'@img/sharp-linux-arm64': 0.34.5
'@img/sharp-linux-ppc64': 0.34.5
'@img/sharp-linux-riscv64': 0.34.5
'@img/sharp-linux-s390x': 0.34.5
'@img/sharp-linux-x64': 0.34.5
'@img/sharp-linuxmusl-arm64': 0.34.5
'@img/sharp-linuxmusl-x64': 0.34.5
'@img/sharp-wasm32': 0.34.5
'@img/sharp-win32-arm64': 0.34.5
'@img/sharp-win32-ia32': 0.34.5
'@img/sharp-win32-x64': 0.34.5
optional: true
first commit
2026-02-21 03:10:39 +00:00
shebang-command@2.0.0:
dependencies:
shebang-regex: 3.0.0
shebang-regex@3.0.0: {}
side-channel-list@1.0.0:
dependencies:
es-errors: 1.3.0
object-inspect: 1.13.4
side-channel-map@1.0.1:
dependencies:
call-bound: 1.0.4
es-errors: 1.3.0
get-intrinsic: 1.3.0
object-inspect: 1.13.4
side-channel-weakmap@1.0.2:
dependencies:
call-bound: 1.0.4
es-errors: 1.3.0
get-intrinsic: 1.3.0
object-inspect: 1.13.4
side-channel-map: 1.0.1
side-channel@1.1.0:
dependencies:
es-errors: 1.3.0
object-inspect: 1.13.4
side-channel-list: 1.0.0
side-channel-map: 1.0.1
side-channel-weakmap: 1.0.2
siginfo@2.0.0: {}
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
signal-exit@3.0.7: {}
first commit
2026-02-21 03:10:39 +00:00
signal-exit@4.1.0: {}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
sigstore@4.1.1:
dependencies:
'@sigstore/bundle': 4.0.0
'@sigstore/core': 3.2.1
'@sigstore/protobuf-specs': 0.5.1
'@sigstore/sign': 4.1.1
'@sigstore/tuf': 4.0.2
'@sigstore/verify': 3.1.1
transitivePeerDependencies:
- supports-color
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
slice-ansi@7.1.2:
dependencies:
ansi-styles: 6.2.3
is-fullwidth-code-point: 5.1.0
slice-ansi@8.0.0:
dependencies:
ansi-styles: 6.2.3
is-fullwidth-code-point: 5.1.0
feat: add Kubernetes orchestrator for MCP server pod management mcpd can now deploy MCP server instances as Kubernetes pods instead of Docker containers. Set MCPD_ORCHESTRATOR=kubernetes to enable. - Add @kubernetes/client-node with thin wrapper (context enforcement via MCPD_K8S_CONTEXT to prevent multi-cluster mishaps) - Rewrite KubernetesOrchestrator: pod CRUD, pod IP extraction, exec via SPDY (one-shot + interactive), log streaming - Manifest generator: stdin:true for STDIO servers, args (not command) to preserve runner image entrypoint, security hardening - Orchestrator selection in main.ts via MCPD_ORCHESTRATOR env var - 25 unit tests for k8s orchestrator, all 624 tests pass Tested end-to-end on local k3s: - mcpd deployed via Pulumi, creates pods in mcpctl-servers namespace - NetworkPolicy verified: only mcpd can reach MCP server pods - Python runner (uvx) successfully runs aws-documentation-mcp-server Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-08 01:55:13 +01:00
smart-buffer@4.2.0: {}
socks-proxy-agent@8.0.5:
dependencies:
agent-base: 7.1.4
debug: 4.4.3
socks: 2.8.7
transitivePeerDependencies:
- supports-color
socks@2.8.7:
dependencies:
ip-address: 10.0.1
smart-buffer: 4.2.0
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
solid-js@1.9.12:
dependencies:
csstype: 3.2.3
seroval: 1.5.6
seroval-plugins: 1.5.6(seroval@1.5.6)
first commit
2026-02-21 03:10:39 +00:00
sonic-boom@4.2.1:
dependencies:
atomic-sleep: 1.0.0
source-map-js@1.2.1: {}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
source-map-support@0.5.21:
dependencies:
buffer-from: 1.1.2
source-map: 0.6.1
source-map@0.6.1: {}
spdx-correct@3.2.0:
dependencies:
spdx-expression-parse: 3.0.1
spdx-license-ids: 3.0.23
spdx-exceptions@2.5.0: {}
spdx-expression-parse@3.0.1:
dependencies:
spdx-exceptions: 2.5.0
spdx-license-ids: 3.0.23
spdx-expression-parse@4.0.0:
dependencies:
spdx-exceptions: 2.5.0
spdx-license-ids: 3.0.23
spdx-license-ids@3.0.23: {}
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
split-ca@1.0.1: {}
first commit
2026-02-21 03:10:39 +00:00
split2@4.2.0: {}
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
ssh2@1.17.0:
dependencies:
asn1: 0.2.6
bcrypt-pbkdf: 1.0.2
optionalDependencies:
cpu-features: 0.0.10
nan: 2.25.0
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
ssri@13.0.1:
dependencies:
minipass: 7.1.3
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
stack-utils@2.0.6:
dependencies:
escape-string-regexp: 2.0.0
first commit
2026-02-21 03:10:39 +00:00
stackback@0.0.2: {}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
state-local@1.0.7: {}
first commit
2026-02-21 03:10:39 +00:00
statuses@2.0.2: {}
std-env@3.10.0: {}
feat: add Kubernetes orchestrator for MCP server pod management mcpd can now deploy MCP server instances as Kubernetes pods instead of Docker containers. Set MCPD_ORCHESTRATOR=kubernetes to enable. - Add @kubernetes/client-node with thin wrapper (context enforcement via MCPD_K8S_CONTEXT to prevent multi-cluster mishaps) - Rewrite KubernetesOrchestrator: pod CRUD, pod IP extraction, exec via SPDY (one-shot + interactive), log streaming - Manifest generator: stdin:true for STDIO servers, args (not command) to preserve runner image entrypoint, security hardening - Orchestrator selection in main.ts via MCPD_ORCHESTRATOR env var - 25 unit tests for k8s orchestrator, all 624 tests pass Tested end-to-end on local k3s: - mcpd deployed via Pulumi, creates pods in mcpctl-servers namespace - NetworkPolicy verified: only mcpd can reach MCP server pods - Python runner (uvx) successfully runs aws-documentation-mcp-server Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-08 01:55:13 +01:00
stream-buffers@3.0.3: {}
streamx@2.25.0:
dependencies:
events-universal: 1.0.1
fast-fifo: 1.3.2
text-decoder: 1.2.7
transitivePeerDependencies:
- bare-abort-controller
- react-native-b4a
first commit
2026-02-21 03:10:39 +00:00
string-width@4.2.3:
dependencies:
emoji-regex: 8.0.0
is-fullwidth-code-point: 3.0.0
strip-ansi: 6.0.1
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
string-width@7.2.0:
dependencies:
emoji-regex: 10.6.0
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
get-east-asian-width: 1.6.0
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
strip-ansi: 7.1.2
string-width@8.2.0:
dependencies:
get-east-asian-width: 1.5.0
strip-ansi: 7.1.2
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
string_decoder@1.3.0:
dependencies:
safe-buffer: 5.2.1
first commit
2026-02-21 03:10:39 +00:00
strip-ansi@6.0.1:
dependencies:
ansi-regex: 5.0.1
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
strip-ansi@7.1.2:
dependencies:
ansi-regex: 6.2.2
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
strip-final-newline@2.0.0: {}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
strip-indent@3.0.0:
dependencies:
min-indent: 1.0.1
feat(web): bold redesign — Tailwind v4 + shadcn-style primitives + Skills/Proposals/Revisions UI Phase 6 of the Skills + Revisions + Proposals work. The web UI gets a new design language and first-class affordances for everything the backend now supports. ## Visual direction - Tailwind v4 with custom @theme block (oklch tokens). Dark-mode-only (internal tool — light mode doubles QA surface). - Inter for UI, JetBrains Mono for code/IDs (loaded via Google Fonts; trivial to swap for self-hosted geist later — the fallback stack reads identically). - Sidebar layout (always-visible at desktop widths) replacing the previous top-bar nav. Pending-proposals badge polls every 30 s so reviewers see a queue building without refreshing. - Lucide icons throughout. - Spacing and radii on Tailwind defaults. Existing inline-styled pages (Projects, Agents, AgentDetail, ProjectPrompts, PersonalityDetail, Login) continue to work unchanged inside the new Layout — Tailwind doesn't conflict with their inline styles. A follow-up can migrate them incrementally. ## What's added ### Build infra (src/web/) - package.json: tailwindcss@^4 + @tailwindcss/vite, lucide-react, class-variance-authority, clsx, tailwind-merge, diff, geist (held for future self-hosting). - vite.config.ts: registers the @tailwindcss/vite plugin. - src/index.css: Tailwind import + @theme tokens + @layer base. - src/main.tsx: imports index.css. - src/lib/utils.ts: shadcn-style cn() helper. ### shadcn-style primitives (src/components/ui/) Hand-written rather than generated via `npx shadcn` so the repo doesn't depend on a CLI tool that needs interactive runtime: - button.tsx — variants: primary / secondary / ghost / danger / link; sizes: sm / md / lg / icon. - card.tsx — Card + Header/Title/Description/Content/Footer subparts. - badge.tsx — variants: default / info / success / warning / danger / outline. - input.tsx — Input + Textarea + Label. - tabs.tsx — no-dep accessible Tabs (no Radix needed for our use). - separator.tsx — h/v separator with role=separator. ### Diff component (src/components/Diff.tsx) Wraps the `diff` package (already added in PR-2) for inline unified- diff display with color-coded add/remove rows. Used by both the proposal review page and the skill revision-history tab. ### New pages (src/pages/) - Dashboard.tsx — at-a-glance home. Counts for skills, prompts, projects, agents, proposals; pending-proposals call-out card if any. - Skills.tsx — list view, separated into Global vs Project/Agent- scoped sections. - SkillDetail.tsx — name + semver + description; tabs for SKILL.md / Files / Metadata / History. History tab shows revisions with click-to-diff against the live body. - Proposals.tsx — queue with Pending/Approved/Rejected tabs. Pending count is highlighted in amber. - ProposalDetail.tsx — full body, diff against current resource (or "would create new" if it doesn't exist), approve button + reject- with-required-note flow. ### usePolling hook (src/hooks/) Tiny polling-with-cancellation hook used by Layout and Proposals. ### Layout rewrite (src/components/Layout.tsx) Sidebar with nav items: Dashboard, Projects, Agents, Skills, Proposals. Lucide icons. Active-route highlighting via NavLink. Pending-proposals warning badge on the Proposals item. ### Routes (src/App.tsx) New routes: /dashboard, /skills, /skills/:name, /proposals, /proposals/:id. Default redirects to /dashboard. ### API types (src/api.ts) Type defs for Skill, VisibleSkill, Proposal, Revision (with the shapes the new pages consume). ## Tests Existing 7 web tests still pass (Login + api). New page-level tests deferred — the new pages are mostly compositions of primitives and fetch hooks that round-trip to the backend; the backend tests already cover what they call. PR-7 polish can add render-and-click tests if coverage drift surfaces. ## Verification - `pnpm --filter @mcpctl/web build` clean, no warnings. - `pnpm test:run` whole monorepo: 162 test files / 2157 tests green. - Visual smoke deferred — needs a running mcpd to populate the fixtures. Manual smoke tested locally is the next step. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-07 17:54:55 +01:00
styled-jsx@5.1.6(@babel/core@7.29.0)(react@19.2.5):
dependencies:
client-only: 0.0.1
react: 19.2.5
optionalDependencies:
'@babel/core': 7.29.0
first commit
2026-02-21 03:10:39 +00:00
supports-color@7.2.0:
dependencies:
has-flag: 4.0.0
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
supports-preserve-symlinks-flag@1.0.0: {}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
symbol-tree@3.2.4: {}
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
tagged-tag@1.0.0: {}
feat(web): bold redesign — Tailwind v4 + shadcn-style primitives + Skills/Proposals/Revisions UI Phase 6 of the Skills + Revisions + Proposals work. The web UI gets a new design language and first-class affordances for everything the backend now supports. ## Visual direction - Tailwind v4 with custom @theme block (oklch tokens). Dark-mode-only (internal tool — light mode doubles QA surface). - Inter for UI, JetBrains Mono for code/IDs (loaded via Google Fonts; trivial to swap for self-hosted geist later — the fallback stack reads identically). - Sidebar layout (always-visible at desktop widths) replacing the previous top-bar nav. Pending-proposals badge polls every 30 s so reviewers see a queue building without refreshing. - Lucide icons throughout. - Spacing and radii on Tailwind defaults. Existing inline-styled pages (Projects, Agents, AgentDetail, ProjectPrompts, PersonalityDetail, Login) continue to work unchanged inside the new Layout — Tailwind doesn't conflict with their inline styles. A follow-up can migrate them incrementally. ## What's added ### Build infra (src/web/) - package.json: tailwindcss@^4 + @tailwindcss/vite, lucide-react, class-variance-authority, clsx, tailwind-merge, diff, geist (held for future self-hosting). - vite.config.ts: registers the @tailwindcss/vite plugin. - src/index.css: Tailwind import + @theme tokens + @layer base. - src/main.tsx: imports index.css. - src/lib/utils.ts: shadcn-style cn() helper. ### shadcn-style primitives (src/components/ui/) Hand-written rather than generated via `npx shadcn` so the repo doesn't depend on a CLI tool that needs interactive runtime: - button.tsx — variants: primary / secondary / ghost / danger / link; sizes: sm / md / lg / icon. - card.tsx — Card + Header/Title/Description/Content/Footer subparts. - badge.tsx — variants: default / info / success / warning / danger / outline. - input.tsx — Input + Textarea + Label. - tabs.tsx — no-dep accessible Tabs (no Radix needed for our use). - separator.tsx — h/v separator with role=separator. ### Diff component (src/components/Diff.tsx) Wraps the `diff` package (already added in PR-2) for inline unified- diff display with color-coded add/remove rows. Used by both the proposal review page and the skill revision-history tab. ### New pages (src/pages/) - Dashboard.tsx — at-a-glance home. Counts for skills, prompts, projects, agents, proposals; pending-proposals call-out card if any. - Skills.tsx — list view, separated into Global vs Project/Agent- scoped sections. - SkillDetail.tsx — name + semver + description; tabs for SKILL.md / Files / Metadata / History. History tab shows revisions with click-to-diff against the live body. - Proposals.tsx — queue with Pending/Approved/Rejected tabs. Pending count is highlighted in amber. - ProposalDetail.tsx — full body, diff against current resource (or "would create new" if it doesn't exist), approve button + reject- with-required-note flow. ### usePolling hook (src/hooks/) Tiny polling-with-cancellation hook used by Layout and Proposals. ### Layout rewrite (src/components/Layout.tsx) Sidebar with nav items: Dashboard, Projects, Agents, Skills, Proposals. Lucide icons. Active-route highlighting via NavLink. Pending-proposals warning badge on the Proposals item. ### Routes (src/App.tsx) New routes: /dashboard, /skills, /skills/:name, /proposals, /proposals/:id. Default redirects to /dashboard. ### API types (src/api.ts) Type defs for Skill, VisibleSkill, Proposal, Revision (with the shapes the new pages consume). ## Tests Existing 7 web tests still pass (Login + api). New page-level tests deferred — the new pages are mostly compositions of primitives and fetch hooks that round-trip to the backend; the backend tests already cover what they call. PR-7 polish can add render-and-click tests if coverage drift surfaces. ## Verification - `pnpm --filter @mcpctl/web build` clean, no warnings. - `pnpm test:run` whole monorepo: 162 test files / 2157 tests green. - Visual smoke deferred — needs a running mcpd to populate the fixtures. Manual smoke tested locally is the next step. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-07 17:54:55 +01:00
tailwind-merge@3.5.0: {}
tailwindcss@4.2.4: {}
tapable@2.3.3: {}
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
tar-fs@2.1.4:
dependencies:
chownr: 1.1.4
mkdirp-classic: 0.5.3
pump: 3.0.3
tar-stream: 2.2.0
feat: add Kubernetes orchestrator for MCP server pod management mcpd can now deploy MCP server instances as Kubernetes pods instead of Docker containers. Set MCPD_ORCHESTRATOR=kubernetes to enable. - Add @kubernetes/client-node with thin wrapper (context enforcement via MCPD_K8S_CONTEXT to prevent multi-cluster mishaps) - Rewrite KubernetesOrchestrator: pod CRUD, pod IP extraction, exec via SPDY (one-shot + interactive), log streaming - Manifest generator: stdin:true for STDIO servers, args (not command) to preserve runner image entrypoint, security hardening - Orchestrator selection in main.ts via MCPD_ORCHESTRATOR env var - 25 unit tests for k8s orchestrator, all 624 tests pass Tested end-to-end on local k3s: - mcpd deployed via Pulumi, creates pods in mcpctl-servers namespace - NetworkPolicy verified: only mcpd can reach MCP server pods - Python runner (uvx) successfully runs aws-documentation-mcp-server Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-08 01:55:13 +01:00
tar-fs@3.1.2:
dependencies:
pump: 3.0.3
tar-stream: 3.1.8
optionalDependencies:
bare-fs: 4.6.0
bare-path: 3.0.0
transitivePeerDependencies:
- bare-abort-controller
- bare-buffer
- react-native-b4a
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
tar-stream@2.2.0:
dependencies:
bl: 4.1.0
end-of-stream: 1.4.5
fs-constants: 1.0.0
inherits: 2.0.4
readable-stream: 3.6.2
feat: add Kubernetes orchestrator for MCP server pod management mcpd can now deploy MCP server instances as Kubernetes pods instead of Docker containers. Set MCPD_ORCHESTRATOR=kubernetes to enable. - Add @kubernetes/client-node with thin wrapper (context enforcement via MCPD_K8S_CONTEXT to prevent multi-cluster mishaps) - Rewrite KubernetesOrchestrator: pod CRUD, pod IP extraction, exec via SPDY (one-shot + interactive), log streaming - Manifest generator: stdin:true for STDIO servers, args (not command) to preserve runner image entrypoint, security hardening - Orchestrator selection in main.ts via MCPD_ORCHESTRATOR env var - 25 unit tests for k8s orchestrator, all 624 tests pass Tested end-to-end on local k3s: - mcpd deployed via Pulumi, creates pods in mcpctl-servers namespace - NetworkPolicy verified: only mcpd can reach MCP server pods - Python runner (uvx) successfully runs aws-documentation-mcp-server Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-08 01:55:13 +01:00
tar-stream@3.1.8:
dependencies:
b4a: 1.8.0
bare-fs: 4.6.0
fast-fifo: 1.3.2
streamx: 2.25.0
transitivePeerDependencies:
- bare-abort-controller
- bare-buffer
- react-native-b4a
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
tar@6.2.1:
dependencies:
chownr: 2.0.0
fs-minipass: 2.1.0
minipass: 5.0.0
minizlib: 2.1.2
mkdirp: 1.0.4
yallist: 4.0.0
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
tar@7.5.20:
dependencies:
'@isaacs/fs-minipass': 4.0.1
chownr: 3.0.0
minipass: 7.1.3
minizlib: 3.1.0
yallist: 5.0.0
feat: add Kubernetes orchestrator for MCP server pod management mcpd can now deploy MCP server instances as Kubernetes pods instead of Docker containers. Set MCPD_ORCHESTRATOR=kubernetes to enable. - Add @kubernetes/client-node with thin wrapper (context enforcement via MCPD_K8S_CONTEXT to prevent multi-cluster mishaps) - Rewrite KubernetesOrchestrator: pod CRUD, pod IP extraction, exec via SPDY (one-shot + interactive), log streaming - Manifest generator: stdin:true for STDIO servers, args (not command) to preserve runner image entrypoint, security hardening - Orchestrator selection in main.ts via MCPD_ORCHESTRATOR env var - 25 unit tests for k8s orchestrator, all 624 tests pass Tested end-to-end on local k3s: - mcpd deployed via Pulumi, creates pods in mcpctl-servers namespace - NetworkPolicy verified: only mcpd can reach MCP server pods - Python runner (uvx) successfully runs aws-documentation-mcp-server Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-08 01:55:13 +01:00
teex@1.0.1:
dependencies:
streamx: 2.25.0
transitivePeerDependencies:
- bare-abort-controller
- react-native-b4a
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
terminal-size@4.0.1: {}
feat: add Kubernetes orchestrator for MCP server pod management mcpd can now deploy MCP server instances as Kubernetes pods instead of Docker containers. Set MCPD_ORCHESTRATOR=kubernetes to enable. - Add @kubernetes/client-node with thin wrapper (context enforcement via MCPD_K8S_CONTEXT to prevent multi-cluster mishaps) - Rewrite KubernetesOrchestrator: pod CRUD, pod IP extraction, exec via SPDY (one-shot + interactive), log streaming - Manifest generator: stdin:true for STDIO servers, args (not command) to preserve runner image entrypoint, security hardening - Orchestrator selection in main.ts via MCPD_ORCHESTRATOR env var - 25 unit tests for k8s orchestrator, all 624 tests pass Tested end-to-end on local k3s: - mcpd deployed via Pulumi, creates pods in mcpctl-servers namespace - NetworkPolicy verified: only mcpd can reach MCP server pods - Python runner (uvx) successfully runs aws-documentation-mcp-server Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-08 01:55:13 +01:00
text-decoder@1.2.7:
dependencies:
b4a: 1.8.0
transitivePeerDependencies:
- react-native-b4a
first commit
2026-02-21 03:10:39 +00:00
thread-stream@4.0.0:
dependencies:
real-require: 0.2.0
tinybench@2.9.0: {}
tinyexec@1.0.2: {}
tinyglobby@0.2.15:
dependencies:
fdir: 6.5.0(picomatch@4.0.3)
picomatch: 4.0.3
tinyrainbow@3.0.3: {}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
tldts-core@7.0.28: {}
tldts@7.0.28:
dependencies:
tldts-core: 7.0.28
first commit
2026-02-21 03:10:39 +00:00
toad-cache@3.7.0: {}
toidentifier@1.0.1: {}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
toml@4.3.0: {}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
tough-cookie@6.0.1:
dependencies:
tldts: 7.0.28
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
tr46@0.0.3: {}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
tr46@6.0.0:
dependencies:
punycode: 2.3.1
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
treeverse@3.0.0: {}
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
ts-algebra@2.0.0: {}
first commit
2026-02-21 03:10:39 +00:00
ts-api-utils@2.4.0(typescript@5.9.3):
dependencies:
typescript: 5.9.3
tslib@2.8.1: {}
tsx@4.21.0:
dependencies:
esbuild: 0.27.3
get-tsconfig: 4.13.6
optionalDependencies:
fsevents: 2.3.3
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
tuf-js@4.1.0:
dependencies:
'@tufjs/models': 4.1.0
debug: 4.4.3
make-fetch-happen: 15.0.6
transitivePeerDependencies:
- supports-color
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
tweetnacl@0.14.5: {}
first commit
2026-02-21 03:10:39 +00:00
type-check@0.4.0:
dependencies:
prelude-ls: 1.2.1
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
type-fest@5.4.4:
dependencies:
tagged-tag: 1.0.0
first commit
2026-02-21 03:10:39 +00:00
type-is@2.0.1:
dependencies:
content-type: 1.0.5
media-typer: 1.1.0
mime-types: 3.0.2
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
typebox@1.3.11: {}
typebox@1.3.7: {}
first commit
2026-02-21 03:10:39 +00:00
typescript@5.9.3: {}
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
undici-types@5.26.5: {}
feat: add Kubernetes orchestrator for MCP server pod management mcpd can now deploy MCP server instances as Kubernetes pods instead of Docker containers. Set MCPD_ORCHESTRATOR=kubernetes to enable. - Add @kubernetes/client-node with thin wrapper (context enforcement via MCPD_K8S_CONTEXT to prevent multi-cluster mishaps) - Rewrite KubernetesOrchestrator: pod CRUD, pod IP extraction, exec via SPDY (one-shot + interactive), log streaming - Manifest generator: stdin:true for STDIO servers, args (not command) to preserve runner image entrypoint, security hardening - Orchestrator selection in main.ts via MCPD_ORCHESTRATOR env var - 25 unit tests for k8s orchestrator, all 624 tests pass Tested end-to-end on local k3s: - mcpd deployed via Pulumi, creates pods in mcpctl-servers namespace - NetworkPolicy verified: only mcpd can reach MCP server pods - Python runner (uvx) successfully runs aws-documentation-mcp-server Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-08 01:55:13 +01:00
undici-types@7.16.0: {}
feat: implement MCP registry client with multi-source search Add registry client that queries Official, Glama, and Smithery MCP registries with caching, request deduplication, retry logic, and result ranking/dedup. Includes 53 tests covering all components. Also fix null priority values in cancelled tasks (19-21) that broke Task Master, and add new tasks 25-27 for registry completion and CLI discover/install commands. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 03:46:14 +00:00
undici-types@7.18.2: {}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
undici@6.27.0: {}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
undici@7.25.0: {}
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
undici@8.9.0: {}
first commit
2026-02-21 03:10:39 +00:00
unpipe@1.0.0: {}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
upath@1.2.0: {}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
update-browserslist-db@1.2.3(browserslist@4.28.2):
dependencies:
browserslist: 4.28.2
escalade: 3.2.0
picocolors: 1.1.1
first commit
2026-02-21 03:10:39 +00:00
uri-js@4.4.1:
dependencies:
punycode: 2.3.1
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
util-deprecate@1.0.2: {}
uuid@10.0.0: {}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
uuid@14.0.1: {}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
validate-npm-package-license@3.0.4:
dependencies:
spdx-correct: 3.2.0
spdx-expression-parse: 3.0.1
validate-npm-package-name@7.0.2: {}
first commit
2026-02-21 03:10:39 +00:00
vary@1.1.2: {}
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
vite@7.3.1(@types/node@25.3.0)(jiti@2.7.0)(lightningcss@1.32.0)(tsx@4.21.0)(yaml@2.9.0):
first commit
2026-02-21 03:10:39 +00:00
dependencies:
esbuild: 0.27.3
fdir: 6.5.0(picomatch@4.0.3)
picomatch: 4.0.3
postcss: 8.5.6
rollup: 4.58.0
tinyglobby: 0.2.15
optionalDependencies:
feat: implement MCP registry client with multi-source search Add registry client that queries Official, Glama, and Smithery MCP registries with caching, request deduplication, retry logic, and result ranking/dedup. Includes 53 tests covering all components. Also fix null priority values in cancelled tasks (19-21) that broke Task Master, and add new tasks 25-27 for registry completion and CLI discover/install commands. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 03:46:14 +00:00
'@types/node': 25.3.0
first commit
2026-02-21 03:10:39 +00:00
fsevents: 2.3.3
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
jiti: 2.7.0
feat(web): bold redesign — Tailwind v4 + shadcn-style primitives + Skills/Proposals/Revisions UI Phase 6 of the Skills + Revisions + Proposals work. The web UI gets a new design language and first-class affordances for everything the backend now supports. ## Visual direction - Tailwind v4 with custom @theme block (oklch tokens). Dark-mode-only (internal tool — light mode doubles QA surface). - Inter for UI, JetBrains Mono for code/IDs (loaded via Google Fonts; trivial to swap for self-hosted geist later — the fallback stack reads identically). - Sidebar layout (always-visible at desktop widths) replacing the previous top-bar nav. Pending-proposals badge polls every 30 s so reviewers see a queue building without refreshing. - Lucide icons throughout. - Spacing and radii on Tailwind defaults. Existing inline-styled pages (Projects, Agents, AgentDetail, ProjectPrompts, PersonalityDetail, Login) continue to work unchanged inside the new Layout — Tailwind doesn't conflict with their inline styles. A follow-up can migrate them incrementally. ## What's added ### Build infra (src/web/) - package.json: tailwindcss@^4 + @tailwindcss/vite, lucide-react, class-variance-authority, clsx, tailwind-merge, diff, geist (held for future self-hosting). - vite.config.ts: registers the @tailwindcss/vite plugin. - src/index.css: Tailwind import + @theme tokens + @layer base. - src/main.tsx: imports index.css. - src/lib/utils.ts: shadcn-style cn() helper. ### shadcn-style primitives (src/components/ui/) Hand-written rather than generated via `npx shadcn` so the repo doesn't depend on a CLI tool that needs interactive runtime: - button.tsx — variants: primary / secondary / ghost / danger / link; sizes: sm / md / lg / icon. - card.tsx — Card + Header/Title/Description/Content/Footer subparts. - badge.tsx — variants: default / info / success / warning / danger / outline. - input.tsx — Input + Textarea + Label. - tabs.tsx — no-dep accessible Tabs (no Radix needed for our use). - separator.tsx — h/v separator with role=separator. ### Diff component (src/components/Diff.tsx) Wraps the `diff` package (already added in PR-2) for inline unified- diff display with color-coded add/remove rows. Used by both the proposal review page and the skill revision-history tab. ### New pages (src/pages/) - Dashboard.tsx — at-a-glance home. Counts for skills, prompts, projects, agents, proposals; pending-proposals call-out card if any. - Skills.tsx — list view, separated into Global vs Project/Agent- scoped sections. - SkillDetail.tsx — name + semver + description; tabs for SKILL.md / Files / Metadata / History. History tab shows revisions with click-to-diff against the live body. - Proposals.tsx — queue with Pending/Approved/Rejected tabs. Pending count is highlighted in amber. - ProposalDetail.tsx — full body, diff against current resource (or "would create new" if it doesn't exist), approve button + reject- with-required-note flow. ### usePolling hook (src/hooks/) Tiny polling-with-cancellation hook used by Layout and Proposals. ### Layout rewrite (src/components/Layout.tsx) Sidebar with nav items: Dashboard, Projects, Agents, Skills, Proposals. Lucide icons. Active-route highlighting via NavLink. Pending-proposals warning badge on the Proposals item. ### Routes (src/App.tsx) New routes: /dashboard, /skills, /skills/:name, /proposals, /proposals/:id. Default redirects to /dashboard. ### API types (src/api.ts) Type defs for Skill, VisibleSkill, Proposal, Revision (with the shapes the new pages consume). ## Tests Existing 7 web tests still pass (Login + api). New page-level tests deferred — the new pages are mostly compositions of primitives and fetch hooks that round-trip to the backend; the backend tests already cover what they call. PR-7 polish can add render-and-click tests if coverage drift surfaces. ## Verification - `pnpm --filter @mcpctl/web build` clean, no warnings. - `pnpm test:run` whole monorepo: 162 test files / 2157 tests green. - Visual smoke deferred — needs a running mcpd to populate the fixtures. Manual smoke tested locally is the next step. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-07 17:54:55 +01:00
lightningcss: 1.32.0
first commit
2026-02-21 03:10:39 +00:00
tsx: 4.21.0
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
yaml: 2.9.0
first commit
2026-02-21 03:10:39 +00:00
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
vitest@4.0.18(@opentelemetry/api@1.9.0)(@types/node@25.3.0)(jiti@2.7.0)(jsdom@28.1.0)(lightningcss@1.32.0)(tsx@4.21.0)(yaml@2.9.0):
first commit
2026-02-21 03:10:39 +00:00
dependencies:
'@vitest/expect': 4.0.18
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
'@vitest/mocker': 4.0.18(vite@7.3.1(@types/node@25.3.0)(jiti@2.7.0)(lightningcss@1.32.0)(tsx@4.21.0)(yaml@2.9.0))
first commit
2026-02-21 03:10:39 +00:00
'@vitest/pretty-format': 4.0.18
'@vitest/runner': 4.0.18
'@vitest/snapshot': 4.0.18
'@vitest/spy': 4.0.18
'@vitest/utils': 4.0.18
es-module-lexer: 1.7.0
expect-type: 1.3.0
magic-string: 0.30.21
obug: 2.1.1
pathe: 2.0.3
picomatch: 4.0.3
std-env: 3.10.0
tinybench: 2.9.0
tinyexec: 1.0.2
tinyglobby: 0.2.15
tinyrainbow: 3.0.3
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
vite: 7.3.1(@types/node@25.3.0)(jiti@2.7.0)(lightningcss@1.32.0)(tsx@4.21.0)(yaml@2.9.0)
first commit
2026-02-21 03:10:39 +00:00
why-is-node-running: 2.3.0
feat: implement MCP registry client with multi-source search Add registry client that queries Official, Glama, and Smithery MCP registries with caching, request deduplication, retry logic, and result ranking/dedup. Includes 53 tests covering all components. Also fix null priority values in cancelled tasks (19-21) that broke Task Master, and add new tasks 25-27 for registry completion and CLI discover/install commands. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 03:46:14 +00:00
optionalDependencies:
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
'@opentelemetry/api': 1.9.0
feat: implement MCP registry client with multi-source search Add registry client that queries Official, Glama, and Smithery MCP registries with caching, request deduplication, retry logic, and result ranking/dedup. Includes 53 tests covering all components. Also fix null priority values in cancelled tasks (19-21) that broke Task Master, and add new tasks 25-27 for registry completion and CLI discover/install commands. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 03:46:14 +00:00
'@types/node': 25.3.0
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
jsdom: 28.1.0
first commit
2026-02-21 03:10:39 +00:00
transitivePeerDependencies:
- jiti
- less
- lightningcss
- msw
- sass
- sass-embedded
- stylus
- sugarss
- terser
- tsx
- yaml
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
w3c-xmlserializer@5.0.0:
dependencies:
xml-name-validator: 5.0.0
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
walk-up-path@4.0.0: {}
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
web-streams-polyfill@3.3.3: {}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
web-tree-sitter@0.25.10: {}
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
webidl-conversions@3.0.1: {}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
webidl-conversions@8.0.1: {}
whatwg-mimetype@5.0.0: {}
whatwg-url@16.0.1:
dependencies:
'@exodus/bytes': 1.15.0
tr46: 6.0.0
webidl-conversions: 8.0.1
transitivePeerDependencies:
- '@noble/hashes'
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
whatwg-url@5.0.0:
dependencies:
tr46: 0.0.3
webidl-conversions: 3.0.1
first commit
2026-02-21 03:10:39 +00:00
which@2.0.2:
dependencies:
isexe: 2.0.0
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
which@6.0.1:
dependencies:
isexe: 4.0.0
first commit
2026-02-21 03:10:39 +00:00
why-is-node-running@2.3.0:
dependencies:
siginfo: 2.0.0
stackback: 0.0.2
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
wide-align@1.1.5:
dependencies:
string-width: 4.2.3
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
widest-line@6.0.0:
dependencies:
string-width: 8.2.0
first commit
2026-02-21 03:10:39 +00:00
word-wrap@1.2.5: {}
wrap-ansi@6.2.0:
dependencies:
ansi-styles: 4.3.0
string-width: 4.2.3
strip-ansi: 6.0.1
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
wrap-ansi@7.0.0:
dependencies:
ansi-styles: 4.3.0
string-width: 4.2.3
strip-ansi: 6.0.1
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
wrap-ansi@9.0.2:
dependencies:
ansi-styles: 6.2.3
string-width: 7.2.0
strip-ansi: 7.1.2
first commit
2026-02-21 03:10:39 +00:00
wrappy@1.0.2: {}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
write-file-atomic@7.0.1:
dependencies:
signal-exit: 4.1.0
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
ws@8.19.0: {}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
xml-name-validator@5.0.0: {}
xmlchars@2.2.0: {}
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
y18n@5.0.8: {}
feat(web): browser-based prompt + personality editor (Stage 5) New workspace package @mcpctl/web — a Vite + React 19 SPA that talks to mcpd's existing HTTP API. Bundles to a static dist/ which Stage 6 will bake into the RPM and serve from mcpd at /ui via @fastify/static. Pages: /ui/projects list projects /ui/projects/:name/prompts CRUD project prompts (Monaco editor) /ui/agents list agents /ui/agents/:name tabs: Direct prompts | Personalities /ui/personalities/:id bind/unbind prompts to a personality Auth: paste a session token (mcpctl auth login) or PAT (mcpctl_pat_*) once on a login screen, kept in localStorage; logout clears it. API client: 60-line fetch wrapper, attaches the bearer header from storage, throws an ApiError with status + parsed body on non-2xx. A 200-line useFetch hook provides loading/error/data without a state-management library — we are not building Notion. UX: - Dark terminal-adjacent theme so the page feels like the CLI. - Monaco @monaco-editor/react for prompt content (markdown mode, word-wrap, search, multi-cursor). - Personality detail's "attach prompt" picker filters in-scope candidates: agent-direct + same-project + globals. Dev loop: pnpm --filter @mcpctl/web dev (vite at :5173, proxies /api to https://mcpctl.ad.itaz.eu — override with MCPCTL_API_URL). Build: pnpm --filter @mcpctl/web build → src/web/dist/. Tests: 7 vitest cases covering the bearer header / 4xx body / 204 no-content path on the api wrapper, and the login storage round-trip + help toggle. Production build green: 269 KB JS / 84 KB gzipped. Typecheck clean (TS strict + exactOptionalPropertyTypes carried over). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-04-26 19:41:57 +01:00
yallist@3.1.1: {}
feat: implement v2 3-tier architecture (mcpctl → mcplocal → mcpd) - Rename local-proxy to mcplocal with HTTP server, LLM pipeline, mcpd discovery - Add LLM pre-processing: token estimation, filter cache, metrics, Gemini CLI + DeepSeek providers - Add mcpd auth (login/logout) and MCP proxy endpoints - Update CLI: dual URLs (mcplocalUrl/mcpdUrl), auth commands, --direct flag - Add tiered health monitoring, shell completions, e2e integration tests - 57 test files, 597 tests passing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-22 11:42:06 +00:00
yallist@4.0.0: {}
feat(pulumi): @mcpctl/pulumi — generic Pulumi provider for mcpctl New workspace package: a Pulumi dynamic provider (TypeScript, in-process) that manages mcpctl resources declaratively from any Pulumi program. Motivated by model drift: when the deployed vLLM/LiteLLM model changes, mcpctl's llm target must follow, and a Pulumi resource makes that automatic on `pulumi up`. - Generic core `McpctlResource({ kind, name, spec, mcpd })` round-trips ANY mcpctl resource kind; typed `Llm` wrapper for ergonomics. - Engine talks direct HTTP to mcpd's REST API (ported from cli/api-client.ts), mirroring apply's name-keyed upsert (PUT strips immutable name/type) and get -o json's read field-stripping. No mcpctl binary needed → CI-friendly. - CRUD/diff/check with correct replace semantics (name/type/kind/mcpd.url force delete-before-replace), 429 retry, secret token in state. - Validation deferred to mcpd's Zod schemas, so new resource kinds work with no provider release. - CommonJS build so the serialized dynamic provider's module refs are captured. - 18 Vitest tests (mocked mcpd + Pulumi mocks); build + lint clean; README with auth (PAT) setup and consumer example for kubernetes-deployment. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 22:02:22 +01:00
yallist@5.0.0: {}
feat: eager vLLM warmup and smart page titles in paginate stage - Add warmup() to LlmProvider interface for eager subprocess startup - ManagedVllmProvider.warmup() starts vLLM in background on project load - ProviderRegistry.warmupAll() triggers all managed providers - NamedProvider proxies warmup() to inner provider - paginate stage generates LLM-powered descriptive page titles when available, cached by content hash, falls back to generic "Page N" - project-mcp-endpoint calls warmupAll() on router creation so vLLM is loading while the session initializes Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-03-03 19:07:39 +00:00
yaml@2.8.2: {}
fix(pi): repair the /mcpctl menu, skills target, and typecheck the extension The pi extension shipped in `src/pi-ext/` was covered by no tsconfig and no eslint config, so nothing ever checked it against pi's API. Pointing tsc at the published @earendil-works/pi-coding-agent types found the command surface to be inert. Fixes: - `/mcpctl` did nothing. `ctx.ui.select` takes `string[]` and returns the chosen string; it was called with `{value,label}` objects, so the menu rendered five `[object Object]` rows and `choice === "status"` never matched any branch. Labels are now plain strings mapped back to actions. - The headless branch returned a status string from a handler typed `Promise<void>`; pi drops it. Reports via notify instead. - "Sync skills" omitted `--agent pi`, writing into ~/.claude/skills — in an integration whose stated purpose is to not depend on ~/.claude — and said so in its own success message. It also ran execSync with `stdio: "inherit"`, painting raw output over pi's TUI, and interpolated the project name into a shell string. Now execFile with `--agent pi` and captured output. - Tool results typed `content[].type` as `string`; pi's AgentToolResult wants the `"text"` literal. - `callTool` asserted `Promise<unknown>` to `ToolCallResult`. - Sanitising MCP tool names to `[a-z0-9_]` can collide (`docs.search` vs `docs-search`). The colliding tool was silently never registered but still reported active, so its calls were forwarded to the first tool. Names are now disambiguated and tracked with the MCP tool they forward to. - `registerWithPi` rewrote settings.json even when nothing changed. Since parsing strips `//` comments, a no-op run destroyed them. Guards, so this class of bug can't return: - `src/pi-ext/tsconfig.json` checks the extension against the real published pi types (dev dependency, not a shim — a shim drifting from the published API is the exact failure being guarded). Wired into `pnpm typecheck`. - eslint now covers `src/pi-ext/*.ts` like every other source file. - A test fails if the embedded copy in `config/pi-extension.ts` is stale; editing the sources without regenerating silently shipped old code. Also: the branch added `config pi` without regenerating shell completions (the committed-completions test was failing), and the doc advertised `mcpctl pi sync-skills`, which does not exist. Both corrected, plus a note on the session-token vs `mcpctl_pat_` bearer difference that would bite against an authenticated `mcplocal serve`. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017BMXdb2qZbPSh8Q7XpTyjB
2026-08-08 17:30:42 +01:00
yaml@2.9.0: {}
feat: add Docker container management for MCP servers McpOrchestrator interface with DockerContainerManager implementation, instance service for lifecycle management, instance API routes, and docker-compose with mcpd service. 127 tests passing. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-21 04:52:12 +00:00
yargs-parser@21.1.1: {}
yargs@17.7.2:
dependencies:
cliui: 8.0.1
escalade: 3.2.0
get-caller-file: 2.0.5
require-directory: 2.1.1
string-width: 4.2.3
y18n: 5.0.8
yargs-parser: 21.1.1
first commit
2026-02-21 03:10:39 +00:00
yocto-queue@0.1.0: {}
yoctocolors-cjs@2.1.3: {}
feat: interactive MCP console (`mcpctl console <project>`) Ink-based TUI that shows exactly what an LLM sees through MCP. Browse tools/resources/prompts, execute them, and see raw JSON-RPC traffic in a protocol log. Supports gated session flow with begin_session, raw JSON-RPC input, and session reconnect. - McpSession class wrapping HTTP transport with typed methods - 12 React/Ink components (header, protocol-log, menu, tool/resource/prompt views, etc.) - 21 unit tests for McpSession against a mock MCP server - Fish + Bash completions with project name argument - bun compile with --external react-devtools-core Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-02-25 23:56:23 +00:00
yoga-layout@3.2.1: {}
first commit
2026-02-21 03:10:39 +00:00
zod-to-json-schema@3.25.1(zod@3.25.76):
dependencies:
zod: 3.25.76
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
zod-to-json-schema@3.25.1(zod@4.1.8):
dependencies:
zod: 4.1.8
first commit
2026-02-21 03:10:39 +00:00
zod@3.25.76: {}
feat(opencode): native opencode integration — /mcpctl switcher, live project switching, footer indicator Adds `mcpctl config opencode`, two opencode plugins and an `opencode` skills sync target, so an mcpctl project can be switched from inside opencode's TUI and the active one is visible at a glance. Unlike `config claude` / `config prime-agent`, this writes NO MCP entry into the host's config. opencode exposes an HTTP API for its own MCP registry (`POST /mcp`), so the project is mounted through the running app: - the token stays in ~/.mcpctl/opencode-state.json (0600) instead of a mode-0644 opencode.json users paste into bug reports; - switching projects takes effect on the next turn, with no restart. Inside opencode: /mcpctl filterable project picker; switches live /mcpctl-status active project, mount state, gateway URL /mcpctl-skills re-sync this project's skills plus a `mcpctl:<project>` indicator in the prompt footer, next to the model name and one line above the token counter. Design notes: - the MCP server is registered under a constant name, so tools keep a stable `mcpctl_*` prefix and opencode's per-request tool resolution shows the new project's tools by itself — no "your old tool names are dead" message to the model, unlike the pi extension; - an unchanged mount is never re-registered: mcp.add rebuilds the connection and mcplocal binds a gated project's unlocked state to that connection's mcp-session-id, so re-adding would re-lock a project begin_session had just opened; - the server plugin does not mount during setup — setup runs before the server accepts connections and mcp.add calls back into it, which hangs opencode on a blank screen before the TUI draws; - the switcher shells out to this CLI (--skip-plugin --skip-marker) so token minting, state and skills stay in one place; - no usable credential aborts non-zero with the state file untouched, so a failed switch leaves the previous project working rather than swapping it for a mount that 401s. `skills sync --agent opencode` installs into ~/.config/opencode/skill (XDG aware) with the same shared-tree semantics as pi and prime-agent. The credential plumbing shared with `config prime-agent` is lifted to one place and parameterised by agent rather than copied. The plugin sources are embedded in the CLI (generated, freshness-tested) so an installed binary with no source tree can provision them, and are typechecked against the real @opencode-ai/plugin types. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BVwuCjuMoA13gmzYEfcrNP
2026-08-08 21:03:53 +01:00
zod@4.1.8: {}
Reference in New Issue Copy Permalink