source-command-self-reflect

source-command-self-reflect

Use this skill when the user asks to run the migrated source command self-reflect.

Command Template

BEADS Self-Reflect

You are performing a self-reflection for the BEADS agent swarm. Your job is to analyze PR review comments, conversation history, and session patterns to extract high-quality, reusable learnings.

Philosophy: Be judicious. Quality over quantity. Each learning should make future development measurably better.

Phase A: PR Comment Analysis

Step 1: Fetch PR Comments

GITHUB_TOKEN=$(gh auth token) npx tsx scripts/beads-fetch-pr-comments.ts --days 7

This outputs PR comments to .beads/temp/pr-comments.json.

Step 2: Extract CodeRabbit's Structured Learnings

cat .beads/temp/pr-comments.json | jq -r '.comments[].body' | grep -A5 "^Learnt from:" | grep "^Learning:" | sed 's/^Learning: //' | sort -u

Evaluate Each CodeRabbit Learning

NOT all CodeRabbit learnings are equal. Evaluate each one:

ACCEPT (High Value)

• Applies to patterns: Applies to **/*.test.ts: ... - Actionable file-scoped rules
• NEVER/ALWAYS rules: Clear, enforceable constraints
• Security/Performance: Critical quality gates
• Gotchas with context: Explains WHY something is problematic

REJECT or DEFER (Low Value)

• PR-specific context: "In PR #X, someone did Y" - Too specific unless the pattern generalizes
• Personality observations: "Developer provides detailed updates" - Not actionable

Quality Filter Questions

Phase B: Conversation & Session Mining (Optional)

Phase C: Config Reflection (Optional)

Step 3: Analyze Comment Patterns

Pattern Detection Checklist

Step 4: Present Candidates for Approval

Candidate Learnings