Conduct a structured multi-perspective debate on a topic using 4 specialized subagents.
Orchestrates a structured multi-perspective debate on any topic using four specialized AI agents.
/plugin marketplace add plinde/claude-plugins/plugin install plinde-socratic-debate-socratic-debate@plinde/claude-pluginsConduct a structured multi-perspective debate on a topic using 4 specialized subagents.
/socratic-debate <topic or question>
$ARGUMENTS - The topic, question, or context to debate (can reference current conversation context)You will orchestrate a Socratic debate by spawning 4 subagents. First read the skill at skills/socratic-debate/SKILL.md to understand the debate framework and guidelines.
Phase 1: Spawn Debaters (parallel)
Launch 3 Task subagents simultaneously with subagent_type=general-purpose:
FOR Agent prompt:
You are debating: $ARGUMENTS
**Your role: Advocate FOR / in favor of this position**
Make the strongest possible case FOR this position. Consider:
- Technical merits and benefits
- Risk mitigation / safety improvements
- Precedent and best practices
- Long-term implications of accepting
Be persuasive but intellectually honest. Acknowledge weaknesses only if they strengthen your overall argument.
Provide a concise argument (250-400 words) with a memorable closing line.
AGAINST Agent prompt:
You are debating: $ARGUMENTS
**Your role: Devil's Advocate AGAINST this position**
Make the strongest possible case AGAINST this position (or for deprioritizing it). Consider:
- Costs, overhead, or complexity introduced
- Whether the problem is overstated
- Alternative approaches or tradeoffs
- Opportunity cost of addressing this
Be persuasive but intellectually honest. Your goal is to stress-test the idea, not dismiss it unfairly.
Provide a concise argument (250-400 words) with a memorable closing line.
NEUTRAL Agent prompt:
You are debating: $ARGUMENTS
**Your role: Neutral Analyst**
Objectively analyze both sides of this debate:
- What are the strongest points on each side?
- What are the key tradeoffs?
- What context or constraints affect the decision?
- Are there hybrid approaches or middle grounds?
Provide:
1. Balanced analysis (200 words)
2. Key tradeoffs table (if applicable)
3. Your preliminary verdict with confidence level (low/medium/high)
Remain impartial but willing to take a position based on evidence.
Phase 2: Synthesize (after Phase 1 completes)
Launch the Scribe/Moderator agent with the outputs from all 3 debaters:
You are the Scribe and Moderator for a Socratic debate.
**Original topic:** $ARGUMENTS
**FOR argument:**
[Insert FOR agent output]
**AGAINST argument:**
[Insert AGAINST agent output]
**NEUTRAL analysis:**
[Insert NEUTRAL agent output]
Your task:
1. Synthesize all three perspectives into a well-formatted debate summary
2. Extract the most compelling points from each side
3. Identify any points of agreement across perspectives
4. Provide a final **Moderator's Verdict** with:
- Clear recommendation (accept/reject/modify/defer)
- Confidence level (low/medium/high)
- Key factor(s) that drove the decision
5. If applicable, suggest next actions
Format the output as a polished markdown document suitable for posting as a PR comment, Slack message, or documentation.
Use emoji headers for each perspective (e.g., green for FOR, red for AGAINST, balance scale for NEUTRAL).
Display the Scribe/Moderator's synthesized debate summary to the user.
If the user's original request mentioned posting somewhere (PR comment, Slack, etc.), offer to post the formatted output there.
/socratic-debate Should we fix Copilot's suggested description change before merging?
/socratic-debate Is it worth refactoring this module to use dependency injection?
/socratic-debate Should we adopt Rust for this performance-critical service?
/socratic-debate PR #1893 - evaluate the tradeoffs of the proposed Vault policy changes