name: team-shinchan:debate description: Specialized agents debate to find optimal solutions. Used for "debate", "pros and cons", "gather opinions" requests. user-invocable: false

EXECUTE IMMEDIATELY

Step 0: Validate Input

If args empty: ask user for topic and STOP. If args > 2000 chars: truncate + warn.

All debates invoke Midori via Task tool.

Auto-Trigger Conditions

On detection: announce "Design decision needed: [situation]", proceed Steps 1-3. Record decision in REQUESTS.md (Stage 1) or PROGRESS.md (Stage 2+).

Step 1: Invoke Midori

Task(subagent_type="team-shinchan:midori", model="sonnet",
  prompt="Debate topic: {topic}\nPanel: {panel list}\nProcedure: Announce, collect opinions (parallel), Hiroshi derives consensus, report decision.")

Step 2: Deliver Results

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
💭 Debate Results
📋 Topic: {topic}
🎤 Opinions: [Agent]: {summary} ...
✅ Decision: {conclusion} | 📝 Rationale: {reasoning}
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

Step 3: Confirm + Finalize

Ask user if they agree. If yes: document and proceed. If no: revise reflecting concerns. Never proceed without confirmation.

Panel Selection

See agents/midori.md for full criteria. Quick reference:

Auto-Detection Signals

Keywords: "A or B", "vs", "method1/method2", "schema change", "layer", "structure", "differs from existing pattern", "trade-off", "at the cost of", "auth", "encryption", "permission"