[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M6FCSM4TF5ZG1FGT2Z970MT2":3,"skill-same:01M6FCSM4TF5ZG1FGT2Z970MT2":58},{"skillNo":4,"name":5,"nameZh":6,"description":6,"descriptionZh":6,"repositoryFullName":7,"ownerLogin":8,"ownerAvatarUrl":9,"ownerAccountNo":10,"ownerDisplayName":11,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":15,"categories":16,"keywords":27,"publishedAt":31,"linkOnly":32,"markdownContent":33,"markdownBody":33,"frontMatter":34,"refName":35,"commitSha":36,"skillDirectory":37,"forksCount":38,"repositoryHtmlUrl":39,"repositoryHomepageUrl":40,"ownerHtmlUrl":41,"ownerBlogUrl":6,"ownerDescription":42,"updatedAt":43,"install":44,"resources":57},"01M6FCSM4TF5ZG1FGT2Z970MT2","debug-council",null,"ComeOnOliver\u002Fskillshub","ComeOnOliver","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F43215937?v=4","01M6E5JNFECJAAVFE0W7SMWWD2","Haoran Shu","01M6E5JNNDQS7AJZDYFQ8TPZCK",67,"MIT","https:\u002F\u002Fgithub.com\u002FComeOnOliver\u002Fskillshub\u002Fblob\u002Fmain\u002Fskills\u002Fmichaelboeding\u002Fskills\u002Fdebug-council\u002FSKILL.md",[17,22],{"categoryNo":18,"categoryName":19,"taxonomyType":20,"taxonomyTypeDesc":21},"01JBSEEDCAT000000000000023","Skill与智能体开发","SCENARIO","使用场景",{"categoryNo":23,"categoryName":24,"taxonomyType":25,"taxonomyTypeDesc":26},"01JBSEEDCAT000000000000030","开发工程师","PROFESSION","适用职业",[28,29,30],"debug","council","debug council","2026-11-10 10:12:30",false,"# Debug Council: Research-Aligned Self-Consistency\n\nPure implementation of self-consistency (Wang et al., 2022). Each agent receives the **raw user prompt** and explores\u002Fdebugs **independently**. No pre-processing, no shared context. Majority voting selects the answer.\n\n**Use this for bugs and problems with ONE correct answer.**\n\n## Step 0: Ask User How Many Agents\n\nBefore doing anything else, **ask the user how many solver agents to use**:\n\n```\nHow many debug agents would you like me to use? (3-10)\n\nRecommendations:\n- 3 agents: Faster, still reliable\n- 5 agents: Good balance\n- 7 agents: High confidence\n- 10 agents: Maximum confidence (critical bugs)\n\nNote: Each agent will independently explore the codebase and find the bug.\nThis takes longer but provides true independence per the research.\n```\n\nWait for the user's response. If they specified a number (e.g., \"debug council of 5\"), use that.\n\n**Minimum: 3 agents** | **Maximum: 10 agents**\n\n---\n\n## CRITICAL: Pure Research Alignment\n\n### What This Means\n\n1. **NO orchestrator exploration** - Do NOT read files or gather context before spawning agents\n2. **Raw user prompt to all agents** - Each agent gets the user's original request, unchanged\n3. **Each agent explores independently** - Agents discover the codebase themselves\n4. **True independence** - No shared context, no cross-contamination\n\n### Why This Matters\n\nThe research shows that **independent samples** converge on correct answers. If we pre-process or share context, we:\n- Introduce orchestrator bias\n- Reduce independence\n- May miss what individual agents would discover\n\n---\n\n## Workflow\n\n### Step 1: Capture the Raw User Prompt\n\nTake the user's request **exactly as stated**. Do NOT:\n- ❌ Read files first\n- ❌ Explore the codebase\n- ❌ Add context\n- ❌ Rephrase or enhance the prompt\n\nJust capture what the user said.\n\n### Step 2: Spawn Agents IN PARALLEL with RAW PROMPT\n\nSpawn ALL agents simultaneously. Each gets the **exact same raw prompt**:\n\n```\nTask(agent: \"debug-solver-1\", prompt: \"[USER'S EXACT WORDS]\")\nTask(agent: \"debug-solver-2\", prompt: \"[USER'S EXACT WORDS]\")\nTask(agent: \"debug-solver-3\", prompt: \"[USER'S EXACT WORDS]\")\n... (all in the SAME batch - parallel execution)\n```\n\n**DO NOT modify the prompt. DO NOT add context. Raw user words only.**\n\n### Step 3: Agents Work Independently\n\nEach agent will:\n1. Read and understand the user's request\n2. Explore the codebase using their tools (Read, Grep, Glob, LS)\n3. Identify the root cause\n4. Reason through solutions (chain-of-thought)\n5. Generate a complete fix\n\n**Each agent works in complete isolation** - they cannot see what other agents are doing or have found.\n\n### Step 4: Track Progress & Collect Solutions\n\nAs agents complete, **show progress to the user**:\n\n```\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n                     AGENT PROGRESS\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n☑ Agent 1 - Complete\n☑ Agent 2 - Complete  \n☑ Agent 3 - Complete\n☐ Agent 4 - Working...\n☐ Agent 5 - Working...\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n```\n\nUpdate this display as each agent finishes. When all complete:\n\n```\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n                     AGENT PROGRESS\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n☑ Agent 1 - Complete ✓\n☑ Agent 2 - Complete ✓\n☑ Agent 3 - Complete ✓\n☑ Agent 4 - Complete ✓\n☑ Agent 5 - Complete ✓\n\nAll agents finished! Analyzing solutions...\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n```\n\nCollect all outputs for voting.\n\n### Step 5: Majority Voting\n\n**Group solutions by their core approach\u002Fanswer:**\n\n1. Identify the **key decision** in each solution\n2. Group solutions that make the same key decision\n3. Count how many agents chose each approach\n\n**Voting rules:**\n- **Clear majority (≥50%)**: Select that solution, HIGH confidence\n- **Plurality (highest \u003C 50%)**: Select that solution, MEDIUM confidence\n- **No clear winner**: Analyze disagreement, LOW confidence\n\n### Step 6: Implement the Winner\n\nImplement the majority solution. Do NOT synthesize or merge - use the winning answer as-is.\n\n### Step 7: Report Results\n\n```\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n                    DEBUG COUNCIL RESULTS\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n\n## ? Voting Summary\n\n| Approach | Description | Agents | Votes |\n|----------|-------------|--------|-------|\n| ✅ A | [description] | 1, 2, 4, 5, 7 | **5\u002F7** |\n| B | [description] | 3, 6 | 2\u002F7 |\n\n**Winner: Approach A** (71% consensus)\n\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n\n## ? What Each Agent Found\n\n### Agent 1\n- Files explored: [list]\n- Root cause identified: [summary]\n- Solution: [brief]\n\n### Agent 2\n- Files explored: [list]\n- Root cause identified: [summary]\n- Solution: [brief]\n\n... (for each agent)\n\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n\n## ? Reasoning Highlights\n\n### Why majority chose Approach A:\n- Agent 1: \"[key insight]\"\n- Agent 2: \"[key insight]\"\n- Agent 4: \"[key insight]\"\n\n### Why minority chose differently:\n- Agent 3: \"[different perspective]\"\n\n### Valuable minority insight:\n[Any good ideas from minority that might be worth noting]\n\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n\n## ? Confidence: HIGH\u002FMEDIUM\u002FLOW\n\n[Explanation based on voting distribution and reasoning quality]\n\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n\n## ✅ Selected Solution\n\n[The complete winning solution]\n\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n\n## ? Implementation\n\n[The actual code change being made]\n\n━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\n```\n\n---\n\n## Configuration\n\n| Mode | Agents | Use Case |\n|------|--------|----------|\n| `debug council of 3` | 3 | Faster, still reliable |\n| `debug council of 5` | 5 | Good balance |\n| `debug council of 7` | 7 | High confidence |\n| `debug council of 10` | 10 | Maximum confidence |\n\nIf user just says `debug council`, ask them to choose.\n\n---\n\n## Research Basis\n\nBased on \"Self-Consistency Improves Chain of Thought Reasoning in Language Models\" (Wang et al., 2022):\n\n| Principle | Our Implementation |\n|-----------|-------------------|\n| Same prompt to all | Raw user prompt, unmodified |\n| Independent samples | Each agent explores independently |\n| No shared context | No orchestrator pre-processing |\n| Chain-of-thought | Agents use ultrathink |\n| Majority voting | Count approaches, select majority |\n\n---\n\n## Why This is Slower (And Why That's OK)\n\nEach agent independently:\n- Explores the codebase\n- Reads relevant files\n- Reasons through the problem\n- Generates a solution\n\nThis takes **3-10x longer** than shared-context approaches, but provides:\n- **True independence** - no orchestrator bias\n- **Diverse exploration** - agents may find different things\n- **Research alignment** - matches the paper exactly\n- **Maximum reliability** - for when accuracy matters most\n\n**Use this for critical problems where getting it right matters more than getting it fast.**\n\n---\n\n## Agents\n\n10 identical debug solver agents in `agents\u002F` directory:\n- `debug-solver-1` through `debug-solver-10`\n\nAll agents:\n- Same instructions\n- Same temperature (0.7)\n- Same tools (Read, Grep, Glob, LS)\n- Use ultrathink (extended thinking)\n- Focus on finding the ONE correct answer\n\nDiversity comes from sampling randomness and independent exploration, not different prompts.\n",{},"main","def8531e65114c0fca8fb8551c1871ee0eed705c","skills\u002Fmichaelboeding\u002Fskills\u002Fdebug-council",24,"https:\u002F\u002Fgithub.com\u002FComeOnOliver\u002Fskillshub","https:\u002F\u002Fskillshub.wtf","https:\u002F\u002Fgithub.com\u002FComeOnOliver","AI & Blockchain Engineer\r\n","2026-06-24 22:07:19",{"source":7,"skillName":5,"commands":45,"prompt":55,"zipAvailable":56},[46,49,52],{"tool":47,"command":48},"npx","npx skills add ComeOnOliver\u002Fskillshub --skill debug-council",{"tool":50,"command":51},"bunx","bunx skills add ComeOnOliver\u002Fskillshub --skill debug-council",{"tool":53,"command":54},"pnpm","pnpm dlx skills add ComeOnOliver\u002Fskillshub --skill debug-council","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FComeOnOliver\u002Fskillshub\u002Fblob\u002Fmain\u002Fskills\u002Fmichaelboeding\u002Fskills\u002Fdebug-council\u002FSKILL.md",true,[],[59,69,79,88,94,101],{"skillNo":60,"name":61,"nameZh":6,"description":6,"descriptionZh":6,"repositoryFullName":7,"ownerLogin":8,"ownerAvatarUrl":9,"ownerAccountNo":10,"ownerDisplayName":11,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":62,"categories":63,"keywords":65,"publishedAt":31},"01M6FKJN13GMG4KY57J3QD8GY5","model-hierarchy-skill","https:\u002F\u002Fgithub.com\u002FComeOnOliver\u002Fskillshub\u002Fblob\u002Fmain\u002Fskills\u002Fzscole\u002Fmodel-hierarchy-skill\u002Fmodel-hierarchy-skill\u002FSKILL.md",[64],{"categoryNo":18,"categoryName":19,"taxonomyType":20,"taxonomyTypeDesc":21},[66,67,68],"model","model hierarchy","model hierarchy skill",{"skillNo":70,"name":71,"nameZh":6,"description":6,"descriptionZh":6,"repositoryFullName":7,"ownerLogin":8,"ownerAvatarUrl":9,"ownerAccountNo":10,"ownerDisplayName":11,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":72,"categories":73,"keywords":75,"publishedAt":31},"01M6FKJJJ6GCAMR4PPPRD12M5X","zoom-mcp","https:\u002F\u002Fgithub.com\u002FComeOnOliver\u002Fskillshub\u002Fblob\u002Fmain\u002Fskills\u002Fzoom\u002Fskills\u002Fzoom-mcp\u002FSKILL.md",[74],{"categoryNo":18,"categoryName":19,"taxonomyType":20,"taxonomyTypeDesc":21},[76,77,78],"MCP","zoom","zoom mcp",{"skillNo":80,"name":81,"nameZh":6,"description":6,"descriptionZh":6,"repositoryFullName":7,"ownerLogin":8,"ownerAvatarUrl":9,"ownerAccountNo":10,"ownerDisplayName":11,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":82,"categories":83,"keywords":84,"publishedAt":31},"01M6FKJE3KYT3KPEZMEEHYKA6K","zoom-apps-sdk","https:\u002F\u002Fgithub.com\u002FComeOnOliver\u002Fskillshub\u002Fblob\u002Fmain\u002Fskills\u002Fzoom\u002Fskills\u002Fzoom-apps-sdk\u002FSKILL.md",[],[85,86,87,77],"sdk","zoom apps sdk","apps",{"skillNo":89,"name":90,"nameZh":6,"description":6,"descriptionZh":6,"repositoryFullName":7,"ownerLogin":8,"ownerAvatarUrl":9,"ownerAccountNo":10,"ownerDisplayName":11,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":91,"categories":92,"keywords":93,"publishedAt":31},"01M6FKJ9EDH9Q5VWDF8HAJRSP0","windows","https:\u002F\u002Fgithub.com\u002FComeOnOliver\u002Fskillshub\u002Fblob\u002Fmain\u002Fskills\u002Fzoom\u002Fskills\u002Fwindows\u002FSKILL.md",[],[90],{"skillNo":95,"name":96,"nameZh":6,"description":6,"descriptionZh":6,"repositoryFullName":7,"ownerLogin":8,"ownerAvatarUrl":9,"ownerAccountNo":10,"ownerDisplayName":11,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":97,"categories":98,"keywords":100,"publishedAt":31},"01M6FKJ666BEEFC3AE9GJ72Q93","whiteboard","https:\u002F\u002Fgithub.com\u002FComeOnOliver\u002Fskillshub\u002Fblob\u002Fmain\u002Fskills\u002Fzoom\u002Fskills\u002Fwhiteboard\u002FSKILL.md",[99],{"categoryNo":18,"categoryName":19,"taxonomyType":20,"taxonomyTypeDesc":21},[96],{"skillNo":102,"name":103,"nameZh":6,"description":6,"descriptionZh":6,"repositoryFullName":7,"ownerLogin":8,"ownerAvatarUrl":9,"ownerAccountNo":10,"ownerDisplayName":11,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":104,"categories":105,"keywords":106,"publishedAt":31},"01M6FKJ21WFCDGTJQ2Q91T9DVW","websockets","https:\u002F\u002Fgithub.com\u002FComeOnOliver\u002Fskillshub\u002Fblob\u002Fmain\u002Fskills\u002Fzoom\u002Fskills\u002Fwebsockets\u002FSKILL.md",[],[103]]