@yemi33/minions 0.1.229 → 0.1.231

package/CHANGELOG.md

@@ -1,5 +1,15 @@
 # Changelog
 
+## 0.1.231 (2026-04-03)
+
+### Fixes
+-  log CC failures to engine log for debugging
+
+## 0.1.230 (2026-04-03)
+
+### Fixes
+-  remove all 'evaluate' work type — eval loop uses 'review' exclusively
+
 ## 0.1.229 (2026-04-03)
 
 ### Fixes

package/dashboard.js

@@ -1672,45 +1672,6 @@ If nothing to do: { "duplicates": [], "reclassify": [], "remove": [] }`;
     return jsonReply(res, 200, plans);
   }
 
-  async function handlePlansArchiveMove(req, res) {
-    try {
-      const body = await readBody(req);
-      if (!body.file) return jsonReply(res, 400, { error: 'file required' });
-      const file = body.file;
-      if (file.includes('..') || file.includes('\0') || file.includes('/') || file.includes('\\')) return jsonReply(res, 400, { error: 'invalid' });
-
-      const isJson = file.endsWith('.json');
-      const sourceDir = isJson ? PRD_DIR : PLANS_DIR;
-      const archiveDir = path.join(sourceDir, 'archive');
-      const sourcePath = path.join(sourceDir, file);
-
-      if (!fs.existsSync(sourcePath)) return jsonReply(res, 404, { error: 'File not found' });
-      if (!fs.existsSync(archiveDir)) fs.mkdirSync(archiveDir, { recursive: true });
-
-      fs.renameSync(sourcePath, path.join(archiveDir, file));
-
-      // If archiving a PRD .json, also archive its source .md plan
-      let archivedSource = null;
-      if (isJson) {
-        try {
-          const prd = safeJson(path.join(archiveDir, file));
-          if (prd?.source_plan) {
-            const mdPath = path.join(PLANS_DIR, prd.source_plan);
-            if (fs.existsSync(mdPath)) {
-              const planArchive = path.join(PLANS_DIR, 'archive');
-              if (!fs.existsSync(planArchive)) fs.mkdirSync(planArchive, { recursive: true });
-              fs.renameSync(mdPath, path.join(planArchive, prd.source_plan));
-              archivedSource = prd.source_plan;
-            }
-          }
-        } catch { /* optional — source plan may not exist */ }
-      }
-
-      invalidateStatusCache();
-      return jsonReply(res, 200, { ok: true, archivedSource });
-    } catch (e) { return jsonReply(res, 400, { error: e.message }); }
-  }
-
   async function handlePlansUnarchive(req, res) {
     try {
       const body = await readBody(req);
@@ -2988,8 +2949,9 @@ What would you like to discuss or change? When you're happy, say "approve" and I
 
         if (result.code !== 0 || !result.text) {
           const debugInfo = result.code !== 0 ? `(exit code ${result.code})` : '(empty response)';
-          const stderrTail = (result.stderr || '').trim().split('\n').filter(Boolean).slice(-3).join(' | ');
+          const stderrTail = (result.stderr || '').trim().split('\n').filter(Boolean).slice(-5).join(' | ');
           console.error(`[CC] LLM failed after retries ${debugInfo}: ${stderrTail}`);
+          try { shared.log('warn', `CC failed ${debugInfo}: ${stderrTail.slice(0, 300)}`); } catch {}
           const hasSession = !!ccSession.sessionId;
           const retryHint = hasSession
             ? 'Your session is still active — just send your message again to retry.'
@@ -3354,8 +3316,6 @@ What would you like to discuss or change? When you're happy, say "approve" and I
     { method: 'POST', path: '/api/plans/unarchive', desc: 'Restore a plan/PRD from archive', params: 'file', handler: handlePlansUnarchive },
     { method: 'POST', path: '/api/plans/revise', desc: 'Request revision with feedback, dispatches agent to revise', params: 'file, feedback, requestedBy?', handler: handlePlansRevise },
     { method: 'POST', path: '/api/plans/discuss', desc: 'Generate a plan discussion session script for Claude CLI', params: 'file', handler: handlePlansDiscuss },
-    { method: 'POST', path: '/api/plans/archive', desc: 'Archive a plan/PRD (move to archive folder)', params: 'file', handler: handlePlansArchiveMove },
-    { method: 'POST', path: '/api/plans/unarchive', desc: 'Unarchive a plan/PRD (restore from archive folder)', params: 'file', handler: handlePlansUnarchive },
     { method: 'GET', path: /^\/api\/plans\/archive\/([^?]+)$/, desc: 'Read an archived plan file', handler: handlePlansArchiveRead },
     { method: 'GET', path: /^\/api\/plans\/([^?]+)$/, desc: 'Read a full plan (JSON from prd/ or markdown from plans/)', handler: handlePlansRead },
 

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@yemi33/minions",
-  "version": "0.1.229",
+  "version": "0.1.231",
   "description": "Multi-agent AI dev team that runs from ~/.minions/ — five autonomous agents share a single engine, dashboard, and knowledge base",
   "bin": {
     "minions": "bin/minions.js"

package/playbooks/evaluate.md

@@ -1,149 +0,0 @@
-# Playbook: Evaluate
-
-You are {{agent_name}}, the {{agent_role}} on the {{project_name}} project.
-TEAM ROOT: {{team_root}}
-
-## Your Task
-
-Evaluate the implementation quality of a completed work item against its acceptance criteria and code quality standards.
-
-## Work Item Under Evaluation
-
-- **ID:** {{item_id}}
-- **Title:** {{item_title}}
-- **Description:** {{item_description}}
-- **Branch:** `{{branch_name}}`
-- **Project:** {{project_name}} (`{{project_path}}`)
-
-{{#acceptance_criteria}}
-## Acceptance Criteria
-
-{{acceptance_criteria}}
-{{/acceptance_criteria}}
-
-{{#references}}
-## References
-
-{{references}}
-{{/references}}
-
-## Evaluation Rubric
-
-Score each category on a 1-5 scale. A category **passes** at 3 or above.
-
-### 1. Correctness (weight: 30%)
-
-Does the implementation do what the task description and acceptance criteria require?
-
-- **5 — Excellent:** All acceptance criteria met, edge cases handled, no functional gaps
-- **4 — Good:** All core criteria met, minor edge cases not handled
-- **3 — Adequate:** Most criteria met, one minor gap that doesn't block usage
-- **2 — Deficient:** One or more acceptance criteria not met
-- **1 — Failing:** Core functionality missing or broken
-
-**Pass threshold:** 3
-
-### 2. Completeness (weight: 25%)
-
-Is the implementation finished end-to-end? No TODO stubs, no half-wired features, no missing integration points.
-
-- **5 — Excellent:** Fully integrated, no loose ends, documentation updated if applicable
-- **4 — Good:** Feature complete, minor polish items remain (comments, naming)
-- **3 — Adequate:** Core feature works, one non-critical integration point incomplete
-- **2 — Deficient:** Significant pieces missing or stubbed out
-- **1 — Failing:** Skeleton or partial implementation only
-
-**Pass threshold:** 3
-
-### 3. Code Quality (weight: 25%)
-
-Does the code follow existing project patterns, naming conventions, and architectural decisions?
-
-- **5 — Excellent:** Clean, idiomatic, follows all project conventions, well-structured
-- **4 — Good:** Follows conventions, minor style inconsistencies
-- **3 — Adequate:** Generally follows patterns, one area deviates without justification
-- **2 — Deficient:** Multiple convention violations, poor structure
-- **1 — Failing:** Ignores project patterns, introduces anti-patterns
-
-**Pass threshold:** 3
-
-### 4. Test Coverage (weight: 20%)
-
-Are there tests for the new functionality? Do existing tests still pass?
-
-- **5 — Excellent:** Comprehensive tests for happy path and edge cases, all passing
-- **4 — Good:** Tests cover core functionality, existing tests pass
-- **3 — Adequate:** At least one test for the main feature, no regressions
-- **2 — Deficient:** No new tests, but existing tests pass
-- **1 — Failing:** No tests, or existing tests broken
-
-**Pass threshold:** 3
-
-## Evaluation Steps
-
-1. **Fetch and review the diff:**
-   ```bash
-   git fetch origin
-   git diff {{main_branch}}...origin/{{branch_name}}
-   ```
-
-2. **Check acceptance criteria** one by one — mark each as MET or NOT MET with evidence
-
-3. **Review code quality** — check for pattern adherence, naming, structure
-
-4. **Verify tests:**
-   ```bash
-   cd {{project_path}}
-   npm test
-   ```
-
-5. **Calculate scores** using the rubric above
-
-6. **Determine verdict:**
-   - **PASS** — all four categories score 3 or above
-   - **FAIL** — any category scores below 3
-
-## Output Format
-
-Structure your evaluation result as follows:
-
-```
-## Evaluation Result
-
-**Item:** {{item_id}} — {{item_title}}
-**Verdict:** PASS | FAIL
-**Weighted Score:** X.X / 5.0
-
-### Scores
-
-| Category | Score | Pass | Notes |
-|----------|-------|------|-------|
-| Correctness | X/5 | YES/NO | ... |
-| Completeness | X/5 | YES/NO | ... |
-| Code Quality | X/5 | YES/NO | ... |
-| Test Coverage | X/5 | YES/NO | ... |
-
-### Acceptance Criteria Checklist
-
-- [x] Criterion 1 — evidence
-- [ ] Criterion 2 — what's missing
-
-### Issues Found
-
-1. **[severity]** Description (file:line)
-
-### Recommendations
-
-- What to fix before merging (if FAIL)
-- Suggestions for improvement (if PASS)
-```
-
-## Rules
-
-- Base your evaluation on **evidence from the diff and test output** — not assumptions
-- If acceptance criteria are missing, evaluate against the task description
-- A FAIL verdict should include actionable feedback — what specifically needs to change
-- Do NOT modify any code — this is a read-only evaluation
-- NEVER checkout branches in the main working tree — use `git diff` and `git show` only
-
-**Note:** Do NOT write to `agents/*/status.json` — the engine manages your status automatically.