Merges feat/agent-project-membership (task 166/152). An agent can finally join many projects: membership routes, per-membership role and optional brain override, chat resolving context from the membership, and CLI attach-agent / detach-agent / get memberships. The legacy Agent.projectId fallback is kept for this release as the schema comment promises. Conflict resolved in main.ts: that branch added its RBAC case to mapUrlToPermission inline, while this branch had extracted the function to middleware/url-permission.ts so it could be tested. Kept the extraction and ported the case into it — joining a team is `edit:agents:<name>`, not `create:agents`, which is the authority to mint a whole agent. Also fixes something the extraction makes possible. security.test.ts contained a hand-copied reproduction of mapUrlToPermission, because main.ts self-invokes main() and nothing in it was importable. A security test asserting against a copy stops testing the system the moment the real one changes, and this copy had already drifted. It now imports the real function; all 18 assertions still pass, against the code that actually runs. Local: build clean, workspace 2550 passed, lint unchanged at 794 errors. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01N4wNHWf7xSwnZCWpJcyv9p
2.1 KiB
2.1 KiB
Task ID: 34
Title: Connect mcplocal MCP router to mcpd proxy endpoint
Status: done
Dependencies: 28 ✓, 29 ✓
Priority: high
Description: Update mcplocal's MCP router to forward tool calls to mcpd's new /api/v1/mcp/proxy endpoint instead of connecting to MCP servers directly.
Details:
- Update
src/mcplocal/src/router.tsto use mcpd proxy:class Router { private mcpdClient: McpdClient; async handleToolsCall(request: JsonRpcRequest) { const { name, arguments: args } = request.params; const [serverName, toolName] = name.split('/'); // Pre-process with LLM if enabled const processedArgs = this.config.enablePreprocessing ? await this.llmProcessor.preprocessRequest(toolName, args) : args; // Call mcpd proxy endpoint const result = await this.mcpdClient.post('/api/v1/mcp/proxy', { serverId: serverName, method: 'tools/call', params: { name: toolName, arguments: processedArgs } }); // Post-process response with LLM if enabled return this.config.enablePreprocessing ? await this.llmProcessor.filterResponse(toolName, args, result) : result; } } - Update upstream configuration:
- Remove direct upstream connections for managed servers
- Keep option for local/unmanaged upstreams
- Add server discovery from mcpd:
async refreshServerList() { const servers = await this.mcpdClient.get('/api/v1/servers'); this.updateAvailableTools(servers); } - Handle tools/list by aggregating from mcpd servers
- Handle resources/list and prompts/list similarly
Test Strategy:
- Unit test: Tool calls are forwarded to mcpd proxy correctly
- Unit test: Server name is extracted from namespaced tool name
- Integration test: Full flow Claude -> mcplocal -> mcpd -> container
- Test tools/list aggregates from all mcpd servers
- Test error handling when mcpd is unreachable
- Test LLM preprocessing is applied when enabled