michal/mcpctl
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

feat(gate): drop reasoning field from selection (~1.5s) #90

Merged
michal merged 1 commits from feat/gate-drop-reasoning into main 2026-07-25 02:07:31 +00:00
Conversation 0 Commits 1 Files Changed 1 +6 -6
1 changed files with 6 additions and 6 deletions
Download Patch File Download Diff File Expand all files Collapse all files

12
src/mcplocal/src/gate/llm-selector.ts
View File

@@ -95,11 +95,11 @@ export class LlmPromptSelector {
): Promise<LlmSelectionResult> { ): Promise<LlmSelectionResult> {
const { getSystemPromptFn, signal, serverInfer } = opts; const { getSystemPromptFn, signal, serverInfer } = opts;
// Keep the answer SHORT: latency is proportional to output length (a // Keep the answer as SHORT as possible: latency is proportional to output
// no-think model runs ~10-13 tok/s), and a verbose reasoning field pushes a // length (a no-think model runs ~10-18 tok/s). Even an "<=8 words" reasoning
// selection past the gate's responsiveness budget. Ask for compact JSON with // got ignored (models write a full sentence → ~7s). Drop the reasoning field
// a <=8-word reasoning — ~2s vs ~9s for a prose reasoning. // entirely — the gate only needs the names → just {selectedNames} → ~1-2s.
const DEFAULT_SYSTEM_PROMPT = `You are a context selection assistant. Given a developer's task keywords and available project prompts, select the relevant ones. Return ONLY compact JSON: {"selectedNames":[...],"reasoning":"<=8 words"}. No prose outside the JSON. Priority 10 prompts must always be included.`; const DEFAULT_SYSTEM_PROMPT = `You are a context selection assistant. Given a developer's task keywords and available project prompts, select the relevant ones. Return ONLY this JSON and nothing else: {"selectedNames":[...]}. No reasoning, no prose, no explanation. Priority 10 prompts must always be included.`;
const systemPrompt = getSystemPromptFn const systemPrompt = getSystemPromptFn
? await getSystemPromptFn('llm-gate-context-selector', DEFAULT_SYSTEM_PROMPT) ? await getSystemPromptFn('llm-gate-context-selector', DEFAULT_SYSTEM_PROMPT)
: DEFAULT_SYSTEM_PROMPT; : DEFAULT_SYSTEM_PROMPT;
@@ -109,7 +109,7 @@ export class LlmPromptSelector {
Available prompts: Available prompts:
${promptIndex.map((p) => `- ${p.name} (priority: ${p.priority}): ${p.summary ?? 'No summary'}${p.chapters?.length ? `\n Chapters: ${p.chapters.join(', ')}` : ''}`).join('\n')} ${promptIndex.map((p) => `- ${p.name} (priority: ${p.priority}): ${p.summary ?? 'No summary'}${p.chapters?.length ? `\n Chapters: ${p.chapters.join(', ')}` : ''}`).join('\n')}
Select the relevant prompts. Return JSON: { "selectedNames": [...], "reasoning": "..." }`; Select the relevant prompts. Return ONLY: {"selectedNames":[...]}`;
const messages: ChatMessage[] = [ const messages: ChatMessage[] = [
{ role: 'system', content: systemPrompt }, { role: 'system', content: systemPrompt },