{ "schemaVersion": "1.0", "item": { "slug": "mixtureofagents-debate", "name": "MOA-Debate", "source": "tencent", "type": "skill", "category": "AI 智能", "sourceUrl": "https://clawhub.ai/markoxmobs/mixtureofagents-debate", "canonicalUrl": "https://clawhub.ai/markoxmobs/mixtureofagents-debate", "targetPlatform": "OpenClaw" }, "install": { "downloadMode": "redirect", "downloadUrl": "/downloads/mixtureofagents-debate", "sourceDownloadUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=mixtureofagents-debate", "sourcePlatform": "tencent", "targetPlatform": "OpenClaw", "installMethod": "Manual import", "extraction": "Extract archive", "prerequisites": [ "OpenClaw" ], "packageFormat": "ZIP package", "includedAssets": [ "SKILL.md", "references/DEBATE_FORMAT.md" ], "primaryDoc": "SKILL.md", "quickSetup": [ "Download the package from Yavira.", "Extract the archive and review SKILL.md first.", "Import or place the package into your OpenClaw setup." ], "agentAssist": { "summary": "Hand the extracted package to your coding agent with a concrete install brief instead of figuring it out manually.", "steps": [ "Download the package from Yavira.", "Extract it into a folder your agent can access.", "Paste one of the prompts below and point your agent at the extracted folder." ], "prompts": [ { "label": "New install", "body": "I downloaded a skill package from Yavira. Read SKILL.md from the extracted folder and install it by following the included instructions. Tell me what you changed and call out any manual steps you could not complete." }, { "label": "Upgrade existing", "body": "I downloaded an updated skill package from Yavira. Read SKILL.md from the extracted folder, compare it with my current installation, and upgrade it while preserving any custom configuration unless the package docs explicitly say otherwise. Summarize what changed and any follow-up checks I should run." } ] }, "sourceHealth": { "source": "tencent", "status": "healthy", "reason": "direct_download_ok", "recommendedAction": "download", "checkedAt": "2026-04-30T16:55:25.780Z", "expiresAt": "2026-05-07T16:55:25.780Z", "httpStatus": 200, "finalUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=network", "contentType": "application/zip", "probeMethod": "head", "details": { "probeUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=network", "contentDisposition": "attachment; filename=\"network-1.0.0.zip\"", "redirectLocation": null, "bodySnippet": null }, "scope": "source", "summary": "Source download looks usable.", "detail": "Yavira can redirect you to the upstream package for this source.", "primaryActionLabel": "Download for OpenClaw", "primaryActionHref": "/downloads/mixtureofagents-debate" }, "validation": { "installChecklist": [ "Use the Yavira download entry.", "Review SKILL.md after the package is downloaded.", "Confirm the extracted package contains the expected setup assets." ], "postInstallChecks": [ "Confirm the extracted package includes the expected docs or setup files.", "Validate the skill or prompts are available in your target agent workspace.", "Capture any manual follow-up steps the agent could not complete." ] }, "downloadPageUrl": "https://openagent3.xyz/downloads/mixtureofagents-debate", "agentPageUrl": "https://openagent3.xyz/skills/mixtureofagents-debate/agent", "manifestUrl": "https://openagent3.xyz/skills/mixtureofagents-debate/agent.json", "briefUrl": "https://openagent3.xyz/skills/mixtureofagents-debate/agent.md" }, "agentAssist": { "summary": "Hand the extracted package to your coding agent with a concrete install brief instead of figuring it out manually.", "steps": [ "Download the package from Yavira.", "Extract it into a folder your agent can access.", "Paste one of the prompts below and point your agent at the extracted folder." ], "prompts": [ { "label": "New install", "body": "I downloaded a skill package from Yavira. Read SKILL.md from the extracted folder and install it by following the included instructions. Tell me what you changed and call out any manual steps you could not complete." }, { "label": "Upgrade existing", "body": "I downloaded an updated skill package from Yavira. Read SKILL.md from the extracted folder, compare it with my current installation, and upgrade it while preserving any custom configuration unless the package docs explicitly say otherwise. Summarize what changed and any follow-up checks I should run." } ] }, "documentation": { "source": "clawhub", "primaryDoc": "SKILL.md", "sections": [ { "title": "Oxford Union Multi-Agent Debate", "body": "When the user wants to debate a motion, stress-test an argument, prepare for a formal debate, or asks you to run an Oxford Union–style debate, follow this procedure." }, { "title": "Overview", "body": "You will simulate a full Oxford Union formal debate by making sequential LLM calls playing different roles. Three agents debate iteratively:\n\nAgentRoleTemperatureBehaviourPropositionSteelmans the case FOR0.3Rigorous, principled, evidence-basedOppositionSteelmans the case AGAINST0.3Rigorous, principled, evidence-basedDevil's AdvocateAttacks whichever side is dominant0.7Lateral, unexpected, adversarial\n\nPlus a neutral Chair (temp 0.4) and a Completeness Judge (temp 0.2)." }, { "title": "Step 1 — Accept the Motion", "body": "Ask the user for a motion, or if they've already provided one, confirm it. Motions follow the \"This House...\" format. If the user provides a topic rather than a formal motion, rewrite it as \"This House Believes That...\" or similar.\n\nSuggested example motions (if the user needs inspiration):\n\n\"This House Believes That Artificial Intelligence Will Be Humanity's Last Great Invention\"\n\"This House Will Under No Circumstances Fight for Its King and Its Country\"\n\"This House Has No Confidence in His Majesty's Government\"\n\nAsk for optional parameters:\n\nMin Score to Pass (default 7.5, range 5–10) — completeness threshold for early stop\nHard Round Cap (default 5, range 3–8) — maximum debate rounds" }, { "title": "Step 2 — Run the Debate Rounds", "body": "For each round (starting at round 1), execute the following sequence. Present each step to the user as it completes." }, { "title": "2a. Proposition Speech", "body": "Use this system prompt:\n\nYou are a skilled Oxford Union debater arguing FOR the proposition. Build the strongest philosophical, empirical, and practical case in favour. Be precise, structured, and rhetorically persuasive. Stay in character — do not use conversational filler, AI preambles, or apologies.\n\nRound 1 user message:\n\nMotion: \"[MOTION]\"\nIdentify the single strongest principled argument FOR this motion. Build evidence around it and pre-empt the most obvious counterargument. Open with a memorable line.\nDeliver your opening speech. Max 150 words.\n\nRound 2+ user message:\n\nMotion: \"[MOTION]\"\nRound summary so far:\n[PREVIOUS_SUMMARY]\nIdentify the single strongest principled argument FOR this motion. Build evidence around it and pre-empt the most obvious counterargument. Open with a memorable line.\nDevelop your case further. Max 150 words.\n\nPresent the speech to the user labelled \"🔵 Proposition\"." }, { "title": "2b. Point of Information (Opposition → Proposition)", "body": "Generate a POI from the Opposition:\n\nMotion: \"[MOTION]\"\nProposition just said:\n\"[PRO_SPEECH]\"\nYou are the Opposition. Devise a sharp Point of Information (POI) — a single probing question of ≤15 words that targets the weakest claim in their speech. Return ONLY the question.\n\nAccept/decline: Randomly decide (60% accept, 40% decline). If accepted, generate a response:\n\nYou are speaking for the Proposition. The Opposition has raised this Point of Information:\n\"[POI_QUESTION]\"\nYou chose to accept it. Respond confidently in ≤25 words.\n\nPresent the POI and outcome (accepted/declined) to the user." }, { "title": "2c. Opposition Speech", "body": "Use this system prompt:\n\nYou are a skilled Oxford Union debater arguing AGAINST the proposition. Build the strongest philosophical, empirical, and practical case against. Be precise, structured, and rhetorically persuasive. Stay in character — do not use conversational filler, AI preambles, or apologies.\n\nRound 1 user message:\n\nMotion: \"[MOTION]\"\nProposition has argued:\n\"[PRO_SPEECH]\"\nIdentify the single strongest principled argument AGAINST this motion. Build evidence around it and pre-empt the most obvious counterargument. Open with a memorable line.\nDeliver your opening speech. Max 150 words.\n\nRound 2+ — include the previous summary and ask to develop the case further.\n\nPresent the speech to the user labelled \"🔴 Opposition\"." }, { "title": "2d. Point of Information (Proposition → Opposition)", "body": "Same as 2b but reversed: Proposition poses the POI, Opposition accepts/declines (60/40)." }, { "title": "2e. Devil's Advocate", "body": "Use this system prompt:\n\nYou are a devil's advocate in an Oxford Union debate. Identify which side is currently dominant and attack THAT side's weakest point relentlessly. You take no permanent position. Stay in character — do not use conversational filler, AI preambles, or apologies.\n\nUser message:\n\nMotion: \"[MOTION]\"\n[Previous summary if round 2+]\nProposition: \"[PRO_SPEECH]\"\nOpposition: \"[CON_SPEECH]\"\nSummarize in one sentence why a side feels dominant, then mount a devastating contrarian attack on their most vulnerable assumption.\nMax 120 words.\n\nPresent labelled \"🟡 Devil's Advocate\". Collect all devil attacks across rounds." }, { "title": "2f. Round Summary (Neutral Chair)", "body": "No system prompt. User message:\n\nMotion: \"[MOTION]\"\nThis round:\nProposition: [PRO_SPEECH]\nOpposition: [CON_SPEECH]\nDevil's Advocate: [DEVIL_SPEECH]\nAs a neutral Oxford Union President, briefly assess: which side has the stronger case and why, what the key unanswered questions are, and what both sides must address next. Max 120 words.\n\nPresent labelled \"⚖️ Round Summary\"." }, { "title": "2g. Completeness Judge (from round 2 onwards)", "body": "No system prompt. User message:\n\nMotion: \"[MOTION]\"\nDebate this round:\nProposition: [PRO_SPEECH]\nOpposition: [CON_SPEECH]\nScore argument completeness 0–10: have the strongest arguments on BOTH sides been raised and have key objections been addressed? Respond ONLY with JSON like this: {\"score\": 7.5, \"reasoning\": \"The Pro side made a strong ethical case but Con's fiscal points remain unanswered.\"}\n\nParse the JSON response. If score >= threshold, stop the round loop. Otherwise, continue to the next round.\n\nPresent the score to the user as a progress indicator." }, { "title": "Step 3 — Closing Rebuttals", "body": "After the round loop ends (by convergence or hard cap):\n\nProposition Rebuttal:\n\nMotion: \"[MOTION]\"\nYou are the first Proposition speaker delivering your closing rebuttal. The debate is over — no new arguments. Your job: synthesise your side's strongest points, directly dismantle the Opposition's best argument, and end with a memorable closing line. Max 120 words.\n\nOpposition Rebuttal (sees Proposition's rebuttal to counter):\n\nMotion: \"[MOTION]\"\nYou are the first Opposition speaker delivering your closing rebuttal. The debate is over — no new arguments. Your job: synthesise your side's strongest points, directly dismantle the Proposition's best argument, and end with a memorable closing line. Max 120 words.\nProposition's rebuttal to counter:\n\"[PRO_REBUTTAL]\"\n\nChair's Verdict (no system prompt):\n\nMotion: \"[MOTION]\"\nClosing rebuttals:\nProposition: \"[PRO_REBUTTAL]\"\nOpposition: \"[CON_REBUTTAL]\"\nAs a neutral Oxford Union Chair, deliver a brief verdict: who made the stronger closing case and why, noting the key rhetorical and logical moments that swayed the debate. Do NOT declare an outright winner — the House votes. Max 100 words.\n\nPresent all three labelled \"💜 Closing Rebuttals\"." }, { "title": "Step 4 — Generate the Debate Brief", "body": "Compile a final structured brief:\n\nMotion: \"[MOTION]\"\nDebate summary:\n[FINAL_SUMMARY]\nClosing rebuttals:\nProposition: [PRO_REBUTTAL]\nOpposition: [CON_REBUTTAL]\nDevil's advocate attacks:\n[ALL_DEVIL_ATTACKS joined by ---]\nGenerate an Oxford Union debate brief. Respond ONLY with JSON conforming to this example:\n{\"pro\": \"1. Argument A... \\n2. Argument B...\", \"con\": \"1. Argument C... \\n2. Argument D...\", \"rebuttals\": \"Prop: Rebuttal X... \\nOpp: Rebuttal Y...\", \"attacks\": \"Attack 1... -> Rebuttal 1...\", \"balance\": \"One paragraph assessment...\"}" }, { "title": "Step 5 — Present the Brief", "body": "Format the brief as a structured document with these sections:\n\n📋 Debate Brief — header with the motion\n🔵 Proposition Case — strongest pro arguments\n🔴 Opposition Case — strongest con arguments\n💜 Key Rebuttal Lines — sharpest closing lines from each side\n🟡 Likely Floor Attacks — devil's advocate exposures with suggested rebuttals\n⚖️ Balanced Assessment — honest assessment of which side has the stronger case\n\nInclude a stop reason:\n\n✅ Converged — if completeness score reached threshold\n⚠️ Hard cap reached — if max rounds hit without convergence" }, { "title": "Notes", "body": "For reference on the full Oxford Union format, agent prompts, and scoring rubric, see {baseDir}/references/DEBATE_FORMAT.md\nThis skill works with any LLM. Use the AI's own model for all agents. Differentiate agents through system prompts and varied temperature/creativity settings\nThe 60/40 POI accept/decline is a random draw, not reasoned — this mirrors the unpredictability of real debate\nTemperature differences are intentional: low temp (0.3) for main debaters ensures rigorous consistency; high temp (0.7) for Devil's Advocate produces lateral, unexpected challenges" } ], "body": "Oxford Union Multi-Agent Debate\n\nWhen the user wants to debate a motion, stress-test an argument, prepare for a formal debate, or asks you to run an Oxford Union–style debate, follow this procedure.\n\nOverview\n\nYou will simulate a full Oxford Union formal debate by making sequential LLM calls playing different roles. Three agents debate iteratively:\n\nAgent\tRole\tTemperature\tBehaviour\nProposition\tSteelmans the case FOR\t0.3\tRigorous, principled, evidence-based\nOpposition\tSteelmans the case AGAINST\t0.3\tRigorous, principled, evidence-based\nDevil's Advocate\tAttacks whichever side is dominant\t0.7\tLateral, unexpected, adversarial\n\nPlus a neutral Chair (temp 0.4) and a Completeness Judge (temp 0.2).\n\nStep 1 — Accept the Motion\n\nAsk the user for a motion, or if they've already provided one, confirm it. Motions follow the \"This House...\" format. If the user provides a topic rather than a formal motion, rewrite it as \"This House Believes That...\" or similar.\n\nSuggested example motions (if the user needs inspiration):\n\n\"This House Believes That Artificial Intelligence Will Be Humanity's Last Great Invention\"\n\"This House Will Under No Circumstances Fight for Its King and Its Country\"\n\"This House Has No Confidence in His Majesty's Government\"\n\nAsk for optional parameters:\n\nMin Score to Pass (default 7.5, range 5–10) — completeness threshold for early stop\nHard Round Cap (default 5, range 3–8) — maximum debate rounds\nStep 2 — Run the Debate Rounds\n\nFor each round (starting at round 1), execute the following sequence. Present each step to the user as it completes.\n\n2a. Proposition Speech\n\nUse this system prompt:\n\nYou are a skilled Oxford Union debater arguing FOR the proposition. Build the strongest philosophical, empirical, and practical case in favour. Be precise, structured, and rhetorically persuasive. Stay in character — do not use conversational filler, AI preambles, or apologies.\n\nRound 1 user message:\n\nMotion: \"[MOTION]\"\n\nIdentify the single strongest principled argument FOR this motion. Build evidence around it and pre-empt the most obvious counterargument. Open with a memorable line.\n\nDeliver your opening speech. Max 150 words.\n\nRound 2+ user message:\n\nMotion: \"[MOTION]\"\n\nRound summary so far: [PREVIOUS_SUMMARY]\n\nIdentify the single strongest principled argument FOR this motion. Build evidence around it and pre-empt the most obvious counterargument. Open with a memorable line.\n\nDevelop your case further. Max 150 words.\n\nPresent the speech to the user labelled \"🔵 Proposition\".\n\n2b. Point of Information (Opposition → Proposition)\n\nGenerate a POI from the Opposition:\n\nMotion: \"[MOTION]\"\n\nProposition just said: \"[PRO_SPEECH]\"\n\nYou are the Opposition. Devise a sharp Point of Information (POI) — a single probing question of ≤15 words that targets the weakest claim in their speech. Return ONLY the question.\n\nAccept/decline: Randomly decide (60% accept, 40% decline). If accepted, generate a response:\n\nYou are speaking for the Proposition. The Opposition has raised this Point of Information: \"[POI_QUESTION]\"\n\nYou chose to accept it. Respond confidently in ≤25 words.\n\nPresent the POI and outcome (accepted/declined) to the user.\n\n2c. Opposition Speech\n\nUse this system prompt:\n\nYou are a skilled Oxford Union debater arguing AGAINST the proposition. Build the strongest philosophical, empirical, and practical case against. Be precise, structured, and rhetorically persuasive. Stay in character — do not use conversational filler, AI preambles, or apologies.\n\nRound 1 user message:\n\nMotion: \"[MOTION]\"\n\nProposition has argued: \"[PRO_SPEECH]\"\n\nIdentify the single strongest principled argument AGAINST this motion. Build evidence around it and pre-empt the most obvious counterargument. Open with a memorable line.\n\nDeliver your opening speech. Max 150 words.\n\nRound 2+ — include the previous summary and ask to develop the case further.\n\nPresent the speech to the user labelled \"🔴 Opposition\".\n\n2d. Point of Information (Proposition → Opposition)\n\nSame as 2b but reversed: Proposition poses the POI, Opposition accepts/declines (60/40).\n\n2e. Devil's Advocate\n\nUse this system prompt:\n\nYou are a devil's advocate in an Oxford Union debate. Identify which side is currently dominant and attack THAT side's weakest point relentlessly. You take no permanent position. Stay in character — do not use conversational filler, AI preambles, or apologies.\n\nUser message:\n\nMotion: \"[MOTION]\"\n\n[Previous summary if round 2+]\n\nProposition: \"[PRO_SPEECH]\" Opposition: \"[CON_SPEECH]\"\n\nSummarize in one sentence why a side feels dominant, then mount a devastating contrarian attack on their most vulnerable assumption.\n\nMax 120 words.\n\nPresent labelled \"🟡 Devil's Advocate\". Collect all devil attacks across rounds.\n\n2f. Round Summary (Neutral Chair)\n\nNo system prompt. User message:\n\nMotion: \"[MOTION]\"\n\nThis round: Proposition: [PRO_SPEECH] Opposition: [CON_SPEECH] Devil's Advocate: [DEVIL_SPEECH]\n\nAs a neutral Oxford Union President, briefly assess: which side has the stronger case and why, what the key unanswered questions are, and what both sides must address next. Max 120 words.\n\nPresent labelled \"⚖️ Round Summary\".\n\n2g. Completeness Judge (from round 2 onwards)\n\nNo system prompt. User message:\n\nMotion: \"[MOTION]\"\n\nDebate this round: Proposition: [PRO_SPEECH] Opposition: [CON_SPEECH]\n\nScore argument completeness 0–10: have the strongest arguments on BOTH sides been raised and have key objections been addressed? Respond ONLY with JSON like this: {\"score\": 7.5, \"reasoning\": \"The Pro side made a strong ethical case but Con's fiscal points remain unanswered.\"}\n\nParse the JSON response. If score >= threshold, stop the round loop. Otherwise, continue to the next round.\n\nPresent the score to the user as a progress indicator.\n\nStep 3 — Closing Rebuttals\n\nAfter the round loop ends (by convergence or hard cap):\n\nProposition Rebuttal:\n\nMotion: \"[MOTION]\"\n\nYou are the first Proposition speaker delivering your closing rebuttal. The debate is over — no new arguments. Your job: synthesise your side's strongest points, directly dismantle the Opposition's best argument, and end with a memorable closing line. Max 120 words.\n\nOpposition Rebuttal (sees Proposition's rebuttal to counter):\n\nMotion: \"[MOTION]\"\n\nYou are the first Opposition speaker delivering your closing rebuttal. The debate is over — no new arguments. Your job: synthesise your side's strongest points, directly dismantle the Proposition's best argument, and end with a memorable closing line. Max 120 words.\n\nProposition's rebuttal to counter: \"[PRO_REBUTTAL]\"\n\nChair's Verdict (no system prompt):\n\nMotion: \"[MOTION]\"\n\nClosing rebuttals: Proposition: \"[PRO_REBUTTAL]\" Opposition: \"[CON_REBUTTAL]\"\n\nAs a neutral Oxford Union Chair, deliver a brief verdict: who made the stronger closing case and why, noting the key rhetorical and logical moments that swayed the debate. Do NOT declare an outright winner — the House votes. Max 100 words.\n\nPresent all three labelled \"💜 Closing Rebuttals\".\n\nStep 4 — Generate the Debate Brief\n\nCompile a final structured brief:\n\nMotion: \"[MOTION]\"\n\nDebate summary: [FINAL_SUMMARY]\n\nClosing rebuttals: Proposition: [PRO_REBUTTAL] Opposition: [CON_REBUTTAL]\n\nDevil's advocate attacks: [ALL_DEVIL_ATTACKS joined by ---]\n\nGenerate an Oxford Union debate brief. Respond ONLY with JSON conforming to this example: {\"pro\": \"1. Argument A... \\n2. Argument B...\", \"con\": \"1. Argument C... \\n2. Argument D...\", \"rebuttals\": \"Prop: Rebuttal X... \\nOpp: Rebuttal Y...\", \"attacks\": \"Attack 1... -> Rebuttal 1...\", \"balance\": \"One paragraph assessment...\"}\n\nStep 5 — Present the Brief\n\nFormat the brief as a structured document with these sections:\n\n📋 Debate Brief — header with the motion\n🔵 Proposition Case — strongest pro arguments\n🔴 Opposition Case — strongest con arguments\n💜 Key Rebuttal Lines — sharpest closing lines from each side\n🟡 Likely Floor Attacks — devil's advocate exposures with suggested rebuttals\n⚖️ Balanced Assessment — honest assessment of which side has the stronger case\n\nInclude a stop reason:\n\n✅ Converged — if completeness score reached threshold\n⚠️ Hard cap reached — if max rounds hit without convergence\nNotes\nFor reference on the full Oxford Union format, agent prompts, and scoring rubric, see {baseDir}/references/DEBATE_FORMAT.md\nThis skill works with any LLM. Use the AI's own model for all agents. Differentiate agents through system prompts and varied temperature/creativity settings\nThe 60/40 POI accept/decline is a random draw, not reasoned — this mirrors the unpredictability of real debate\nTemperature differences are intentional: low temp (0.3) for main debaters ensures rigorous consistency; high temp (0.7) for Devil's Advocate produces lateral, unexpected challenges" }, "trust": { "sourceLabel": "tencent", "provenanceUrl": "https://clawhub.ai/markoxmobs/mixtureofagents-debate", "publisherUrl": "https://clawhub.ai/markoxmobs/mixtureofagents-debate", "owner": "markoxmobs", "version": "1.0.0", "license": null, "verificationStatus": "Indexed source record" }, "links": { "detailUrl": "https://openagent3.xyz/skills/mixtureofagents-debate", "downloadUrl": "https://openagent3.xyz/downloads/mixtureofagents-debate", "agentUrl": "https://openagent3.xyz/skills/mixtureofagents-debate/agent", "manifestUrl": "https://openagent3.xyz/skills/mixtureofagents-debate/agent.json", "briefUrl": "https://openagent3.xyz/skills/mixtureofagents-debate/agent.md" } }