{"id":20622,"plugin_id":"plugins_6aa5c14ce164819185474f73c890b680","kind":"skill","collection_source":null,"comparison_source":null,"observed_at":"2026-09-30T23:16:23.861Z","digest":"ec057c5f30b4cbf56b3eee44c02b6d5a7fd641ff5efee4fccb7cabe914422a87","against":null,"payload":{"description":"Run a multi-round deliberation between reviewers. Unlike /cw-panel (which synthesizes), /cw-debate has reviewers respond to each other's arguments across rounds until tensions resolve or reach acknowledged stalemate.","included_files":[],"name":"cw-debate","skill_md_contents":"---\nname: cw-debate\ndescription: Run a multi-round deliberation between reviewers. Unlike /cw-panel (which synthesizes), /cw-debate has reviewers respond to each other's arguments across rounds until tensions resolve or reach acknowledged stalemate.\n---\n\n# Debate Review\n\nOrchestrate a structured deliberation between reviewer agents. Reviewers don't just give feedback in parallel—they engage with each other's perspectives, challenge each other's recommendations, and work toward resolution (or acknowledged stalemate).\n\n## When to Use /cw-debate vs /cw-panel\n\n| Use | When |\n|-----|------|\n| `/cw-panel` | You want multiple perspectives synthesized. Fast. Tensions surfaced for you to decide. |\n| `/cw-debate` | You want reviewers to actually argue it out. More thorough. Tensions may resolve through deliberation. |\n\n**Choose /cw-debate when:**\n- The piece is high-stakes and worth the extra rounds\n- You want to see how perspectives hold up under challenge\n- You suspect some tensions might resolve if reviewers engaged each other\n- You want proposals and compromises, not just \"you decide\"\n\n**Choose /cw-panel when:**\n- You want comprehensive feedback quickly\n- You're comfortable resolving tensions yourself\n- Time/tokens are a concern\n\n## The Deliberation Flow\n\n### Round 1: Initial Positions\n\nSame as /cw-panel—all reviewers analyze the draft independently.\n\n### Round 2: Challenges\n\nThe moderator identifies tensions and sends challenges to involved reviewers:\n\n```\n## Challenge: The fishing gear paragraph\n\n**The passage:**\n> \"His tackle box was a museum of failures—rusted lures from the '70s...\"\n\n**hemingway said:**\n> \"47 words that could be 15. Cut it.\"\n\n**sedaris said:**\n> \"This is the most alive moment in the piece. Keep every word.\"\n\n---\n\n**hemingway:** sedaris argues the specificity earns its place. How do you respond?\n\n**sedaris:** hemingway argues this slows momentum. How do you respond?\n```\n\n### Round 3: Responses\n\nReviewers respond in character. They may:\n\n- **Concede** — \"On reflection, the detail does earn its place. I withdraw.\"\n- **Hold** — \"The momentum problem remains. Even good details hurt here.\"\n- **Propose** — \"Keep the first image, cut the extended list. Satisfies both.\"\n\n### Round 4: Resolution\n\nFor each tension:\n- **Resolved** — Reviewers agree (via concession or proposal)\n- **Stalemate** — Reviewers hold, fundamental value difference\n- **Proposal on table** — Compromise offered, writer decides\n\nThe moderator may run additional rounds if proposals generate new discussion, but caps at 4 rounds.\n\n## The Flow in Practice\n\n### 1. Load Context\n\nRead `../../references/context-contract.md`, the draft, and any available brief, outline, project instructions, feedback, or status metadata. Establish:\n- piece_type\n- audience\n- stage\n- goals\n\nInfer missing context from the draft and request. Do not require `.status.yaml`.\n\n### 2. Propose Panel\n\nSame as /cw-panel—propose 4-6 reviewers based on context.\n\n```\n## Proposed Panel for Debate\n\n**Draft:** [title]\n**Context:** [piece_type] for [audience], [stage]\n\n| Reviewer | Why |\n|----------|-----|\n| **[name]** | [Rationale] |\n| **[name]** | [Rationale] |\n\n**Note:** Debate works best with reviewers who have natural tensions.\nConsider including pairs like:\n- hemingway + sedaris (economy vs. specificity)\n- reader + hitchcock (orientation vs. mystery)\n- sorkin + vonnegut (speed vs. depth)\n\nProceed with this panel, or adjust?\n```\n\nWait for user confirmation.\n\n### 3. Round 1: Run Reviewers\n\nRun all reviewers in parallel when the active platform supports isolated workers and the user's explicit debate request authorizes them. On Claude, use the matching reviewer agents when available; on Codex or another runtime, use the reviewer skills or isolated workers with those lenses. Collect independent outputs.\n\n### 4. Round 2: Identify Tensions and Generate Challenges\n\nAnalyze outputs for conflicts. For each significant tension:\n- Quote the passage\n- Present both positions\n- Generate challenge prompts\n\nSpawn challenged reviewers with the challenge context.\n\n### 5. Round 3: Collect Responses\n\nEach reviewer responds to their challenge:\n- Concede, Hold, or Propose\n- With reasoning\n\n### 6. Round 4: Resolution and Additional Rounds\n\nDetermine resolution status. If proposals warrant further discussion, run another round (max 4 total).\n\n### 7. Final Synthesis\n\nReturn the debate-moderator's synthesis:\n- Resolved tensions (with how they resolved)\n- Stalemates (with final positions and stakes)\n- Proposals (compromise options)\n- Consensus findings (non-debated)\n- The hard question\n\n## Panel Selection for Debate\n\nDebate is most valuable when the panel includes natural tensions:\n\n| Pairing | Tension Type | Good For |\n|---------|--------------|----------|\n| hemingway + sedaris | Economy vs. Specificity | Deciding what earns its length |\n| reader + hitchcock | Orientation vs. Mystery | Distinguishing strategic withholding from accidental confusion |\n| reader + hemingway | Context vs. Brevity | Deciding what setup a first-time reader needs |\n| reader + sorkin | Orientation vs. Momentum | Giving readers footing without slowing the piece |\n| mom + hemingway | Accessibility vs. Brevity | Deciding what a non-expert audience needs explained |\n| sorkin + sedaris | Momentum vs. Observation | Pacing of descriptive passages |\n| sorkin + vonnegut | Speed vs. Depth | When to slow for character |\n| nemesis + sedaris | Rigor vs. Voice | Tone of argument pieces |\n\n**Recommendation:** Include at least one natural tension pair in your panel.\n\n## Example Session\n\n```\nUser: /cw-debate\n\nAgent: I'll run a deliberation on your draft. Let me read it and propose a panel.\n\n[Reads draft]\n[Loads the brief and project context: personal essay, general audience, revision stage]\n\n## Proposed Panel for Debate\n\n**Draft:** father-and-the-sea\n**Context:** personal essay for general audience, revision stage\n\n| Reviewer | Why |\n|----------|-----|\n| **hemingway** | Check economy—personal essays can get overwritten |\n| **sedaris** | Find the specific, human, funny moments |\n| **sorkin** | Verify pacing through emotional beats |\n| **reader** | Trace the first-time reading experience |\n\n**Natural tensions:** hemingway/sedaris (economy vs. specificity),\nsorkin/sedaris (momentum vs. observation)\n\nProceed with this panel, or adjust?\n\nUser: yes\n\nAgent: Running Round 1—each reviewer analyzing independently.\n\n[Spawns 4 reviewers in parallel]\n[Collects outputs]\n\nRound 1 complete. Analyzing for tensions.\n\n## Tensions Identified\n\n### Tension 1: The tackle box paragraph\n**hemingway** wants to cut; **sedaris** wants to keep every word.\n\n### Tension 2: Section 3 pacing\n**sorkin** says it stalls; **sedaris** says the observation is the point.\n\n### Tension 3: Ending explanation\n**mom** wants more; **sorkin** says trust the reader.\n\nProceeding to Round 2 with 3 challenges.\n\n[Spawns challenged reviewers with challenge context]\n[Collects responses]\n\n## Round 2 Responses\n\n**Tension 1:**\n- hemingway: HOLD — \"Momentum matters more than any single detail.\"\n- sedaris: PROPOSE — \"Keep 'museum of failures' and the pug. Cut the Reagan line.\"\n\n**Tension 2:**\n- sorkin: CONCEDE — \"If the observation is the point, I was wrong to call it a stall.\"\n- sedaris: (no concession needed)\n\n**Tension 3:**\n- mom: HOLD — \"I still don't know what he learned. That's a real gap.\"\n- sorkin: PROPOSE — \"One specific image of the lesson. Not explanation—scene.\"\n\nProceeding to final synthesis.\n\n## Debate Synthesis\n\n**Panel:** hemingway, sedaris, sorkin, mom\n**Rounds:** 2\n**Context:** personal essay for general audience, revision stage\n\n---\n\n### Resolved Tensions\n\n**Section 3 pacing**\n\n> [The observation passage]\n\n**Resolution:** Sorkin conceded. The slow pace is intentional and earns its place.\n**Recommendation:** Keep as-is. The observation is the point.\n\n---\n\n### Stalemates\n\n*None—all tensions reached resolution or proposal.*\n\n---\n\n### Proposals\n\n**The tackle box paragraph**\n\n> \"His tackle box was a museum of failures—rusted lures from the '70s,\n> a reel that hadn't worked since Reagan, hooks that had never caught\n> anything but his own thumb.\"\n\n**Proposed revision:** \"His tackle box was a museum of failures—hooks\nthat had never caught anything but his own thumb.\"\n\n**Proposed by:** sedaris (compromise with hemingway)\n**Would satisfy:** Keeps the voice and \"museum\" metaphor, cuts the catalog\n**Trade-off:** Loses the temporal sweep (\"'70s,\" \"Reagan\")\n\n---\n\n**The ending**\n\n**Proposed revision:** Add one concrete image showing what you learned,\nrather than explaining it.\n\n**Proposed by:** sorkin (compromise with mom)\n**Would satisfy:** Mom gets clarity; sorkin keeps momentum\n**Trade-off:** Requires writing new material\n\n---\n\n### Consensus Findings\n\n| Issue | Flagged By | Recommendation |\n|-------|------------|----------------|\n| Opening is slow | hemingway, sorkin | Start on the boat |\n| \"He taught me everything\" unearned | all | Show, don't declare |\n\n---\n\n### The Deliberation Summary\n\n**Resolved:** 1 tension resolved through concession\n**Stalemates:** 0\n**Proposals:** 2 compromise options for your consideration\n\n---\n\n### The Hard Question\n\n> The debate kept circling one thing: you're telling us the relationship\n> mattered without showing us what made it irreplaceable. The ending\n> proposal addresses this—but only if you write a scene that earns it.\n```\n\n## Cost And Scope\n\nDebate is materially slower and more context-intensive than panel because reviewers receive follow-up challenges. Use it for consequential tensions, cap the number of rounds, and prefer `cw-panel` for routine coverage.\n\n## Error Handling\n\n**If a reviewer fails to spawn:**\n- Note the gap\n- Proceed with available reviewers\n- Tensions involving that reviewer won't be debated\n\n**If no tensions identified:**\n- Skip to synthesis\n- Report as \"No significant tensions—reviewers largely aligned\"\n\n**If a reviewer doesn't respond to challenge:**\n- Treat as implicit HOLD\n- Note non-response in synthesis\n\n**If debate goes circular:**\n- Moderator calls stalemate after positions repeat\n- Cap at 4 rounds regardless\n\n## Relationship to /cw-panel\n\n`/cw-panel` and `/cw-debate` share infrastructure:\n- Same reviewer agents\n- Same context loading from the active project and artifact\n- Same panel proposal logic\n\nThe difference:\n- `/cw-panel` → synthesizer (one pass, tensions surfaced)\n- `/cw-debate` → moderator (multiple rounds, tensions engaged)\n\nA natural workflow: run `/cw-panel` first for quick feedback, then `/cw-debate` on specific tensions if you want them argued out.\n\n## Lessons\n\n[Skill-specific lessons will be added here as they're captured]\n"},"changes":[],"summary":"First saved snapshot. No earlier version is available for comparison.","summary_kind":"deterministic","summary_metadata":{}}