[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M739F88XRF2T7EXMNA2FAPA8":3,"skill-same:01M739F88XRF2T7EXMNA2FAPA8":47},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":25,"commitSha":26,"skillDirectory":27,"forksCount":14,"repositoryHtmlUrl":28,"repositoryHomepageUrl":6,"ownerHtmlUrl":29,"ownerBlogUrl":30,"ownerDescription":31,"updatedAt":32,"install":33,"resources":46},"01M739F88XRF2T7EXMNA2FAPA8","eval",null,"Evaluate and rank agent results by metric or LLM judge for an AgentHub session.","kalash33\u002Fkstack","kalash33","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F90259149?v=4","01M7365X8Q0B73NYHA730SMZWN","Kalash Poddar","01M7365X9D0JK47C3Q0YCCK9Z2",0,"MIT","https:\u002F\u002Fgithub.com\u002Fkalash33\u002Fkstack\u002Fblob\u002Fmain\u002Fskills\u002Fnginity-agenthub\u002Fskills\u002Feval\u002FSKILL.md",[],[],"2026-11-11 15:11:14",false,"---\nname: \"eval\"\ndescription: \"Evaluate and rank agent results by metric or LLM judge for an AgentHub session.\"\ncommand: \u002Fhub:eval\n---\n\n# \u002Fhub:eval — Evaluate Agent Results\n\nRank all agent results for a session. Supports metric-based evaluation (run a command), LLM judge (compare diffs), or hybrid.\n\n## Usage\n\n```\n\u002Fhub:eval                           # Eval latest session using configured criteria\n\u002Fhub:eval 20260317-143022           # Eval specific session\n\u002Fhub:eval --judge                   # Force LLM judge mode (ignore metric config)\n```\n\n## What It Does\n\n### Metric Mode (eval command configured)\n\nRun the evaluation command in each agent's worktree:\n\n```bash\npython {skill_path}\u002Fscripts\u002Fresult_ranker.py \\\n  --session {session-id} \\\n  --eval-cmd \"{eval_cmd}\" \\\n  --metric {metric} --direction {direction}\n```\n\nOutput:\n```\nRANK  AGENT       METRIC      DELTA      FILES\n1     agent-2     142ms       -38ms      2\n2     agent-1     165ms       -15ms      3\n3     agent-3     190ms       +10ms      1\n\nWinner: agent-2 (142ms)\n```\n\n### LLM Judge Mode (no eval command, or --judge flag)\n\nFor each agent:\n1. Get the diff: `git diff {base_branch}...{agent_branch}`\n2. Read the agent's result post from `.agenthub\u002Fboard\u002Fresults\u002Fagent-{i}-result.md`\n3. Compare all diffs and rank by:\n   - **Correctness** — Does it solve the task?\n   - **Simplicity** — Fewer lines changed is better (when equal correctness)\n   - **Quality** — Clean execution, good structure, no regressions\n\nPresent rankings with justification.\n\nExample LLM judge output for a content task:\n```\nRANK  AGENT    VERDICT                               WORD COUNT\n1     agent-1  Strong narrative, clear CTA            1480\n2     agent-3  Good data points, weak intro           1520\n3     agent-2  Generic tone, no differentiation       1350\n\nWinner: agent-1 (strongest narrative arc and call-to-action)\n```\n\n### Hybrid Mode\n\n1. Run metric evaluation first\n2. If top agents are within 10% of each other, use LLM judge to break ties\n3. Present both metric and qualitative rankings\n\n## After Eval\n\n1. Update session state:\n```bash\npython {skill_path}\u002Fscripts\u002Fsession_manager.py --update {session-id} --state evaluating\n```\n\n2. Tell the user:\n   - Ranked results with winner highlighted\n   - Next step: `\u002Fhub:merge` to merge the winner\n   - Or `\u002Fhub:merge {session-id} --agent {winner}` to be explicit\n","# \u002Fhub:eval — Evaluate Agent Results\n\nRank all agent results for a session. Supports metric-based evaluation (run a command), LLM judge (compare diffs), or hybrid.\n\n## Usage\n\n```\n\u002Fhub:eval                           # Eval latest session using configured criteria\n\u002Fhub:eval 20260317-143022           # Eval specific session\n\u002Fhub:eval --judge                   # Force LLM judge mode (ignore metric config)\n```\n\n## What It Does\n\n### Metric Mode (eval command configured)\n\nRun the evaluation command in each agent's worktree:\n\n```bash\npython {skill_path}\u002Fscripts\u002Fresult_ranker.py \\\n  --session {session-id} \\\n  --eval-cmd \"{eval_cmd}\" \\\n  --metric {metric} --direction {direction}\n```\n\nOutput:\n```\nRANK  AGENT       METRIC      DELTA      FILES\n1     agent-2     142ms       -38ms      2\n2     agent-1     165ms       -15ms      3\n3     agent-3     190ms       +10ms      1\n\nWinner: agent-2 (142ms)\n```\n\n### LLM Judge Mode (no eval command, or --judge flag)\n\nFor each agent:\n1. Get the diff: `git diff {base_branch}...{agent_branch}`\n2. Read the agent's result post from `.agenthub\u002Fboard\u002Fresults\u002Fagent-{i}-result.md`\n3. Compare all diffs and rank by:\n   - **Correctness** — Does it solve the task?\n   - **Simplicity** — Fewer lines changed is better (when equal correctness)\n   - **Quality** — Clean execution, good structure, no regressions\n\nPresent rankings with justification.\n\nExample LLM judge output for a content task:\n```\nRANK  AGENT    VERDICT                               WORD COUNT\n1     agent-1  Strong narrative, clear CTA            1480\n2     agent-3  Good data points, weak intro           1520\n3     agent-2  Generic tone, no differentiation       1350\n\nWinner: agent-1 (strongest narrative arc and call-to-action)\n```\n\n### Hybrid Mode\n\n1. Run metric evaluation first\n2. If top agents are within 10% of each other, use LLM judge to break ties\n3. Present both metric and qualitative rankings\n\n## After Eval\n\n1. Update session state:\n```bash\npython {skill_path}\u002Fscripts\u002Fsession_manager.py --update {session-id} --state evaluating\n```\n\n2. Tell the user:\n   - Ranked results with winner highlighted\n   - Next step: `\u002Fhub:merge` to merge the winner\n   - Or `\u002Fhub:merge {session-id} --agent {winner}` to be explicit\n",{"name":5,"command":24,"description":7},"\u002Fhub:eval","main","b30820b3248f6676d5389f0ccabd4e5250c9c536","skills\u002Fnginity-agenthub\u002Fskills\u002Feval","https:\u002F\u002Fgithub.com\u002Fkalash33\u002Fkstack","https:\u002F\u002Fgithub.com\u002Fkalash33","kalashpoddar.me","Tech-savvy Full Stack Developer 🚀 | Cybersecurity enthusiast 🔒 | Passionate about innovation and securing the digital world 🌐 | Let's code securely! 💻🔐","2026-05-20 04:11:22",{"source":8,"skillName":5,"commands":34,"prompt":44,"zipAvailable":45},[35,38,41],{"tool":36,"command":37},"npx","npx skills add kalash33\u002Fkstack --skill eval",{"tool":39,"command":40},"bunx","bunx skills add kalash33\u002Fkstack --skill eval",{"tool":42,"command":43},"pnpm","pnpm dlx skills add kalash33\u002Fkstack --skill eval","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fkalash33\u002Fkstack\u002Fblob\u002Fmain\u002Fskills\u002Fnginity-agenthub\u002Fskills\u002Feval\u002FSKILL.md",true,[],[48,55,62,69,76,83],{"skillNo":49,"name":50,"nameZh":6,"description":51,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":52,"categories":53,"keywords":54,"publishedAt":19},"01M73CFERCR43JCAP606E41EFB","zustand-store-ts","Create Zustand stores with TypeScript, subscribeWithSelector middleware, and proper state\u002Faction separation. Use when building React state management, creating global stores, or implementing reacti...","https:\u002F\u002Fgithub.com\u002Fkalash33\u002Fkstack\u002Fblob\u002Fmain\u002Fskills\u002Fzustand-store-ts\u002FSKILL.md",[],[],{"skillNo":56,"name":57,"nameZh":6,"description":58,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":59,"categories":60,"keywords":61,"publishedAt":19},"01M73CFCMFVTAQEMTZ5HR10X8V","zoom-automation","Automate Zoom meeting creation, management, recordings, webinars, and participant tracking via Rube MCP (Composio). Always search tools first for current schemas.","https:\u002F\u002Fgithub.com\u002Fkalash33\u002Fkstack\u002Fblob\u002Fmain\u002Fskills\u002Fzoom-automation\u002FSKILL.md",[],[],{"skillNo":63,"name":64,"nameZh":6,"description":65,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":66,"categories":67,"keywords":68,"publishedAt":19},"01M73CFAA7EBWXJVXDQ1X555MB","zoho-crm-automation","Automate Zoho CRM tasks via Rube MCP (Composio): create\u002Fupdate records, search contacts, manage leads, and convert leads. Always search tools first for current schemas.","https:\u002F\u002Fgithub.com\u002Fkalash33\u002Fkstack\u002Fblob\u002Fmain\u002Fskills\u002Fzoho-crm-automation\u002FSKILL.md",[],[],{"skillNo":70,"name":71,"nameZh":6,"description":72,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":73,"categories":74,"keywords":75,"publishedAt":19},"01M73CF80XCM9K81567MJ8QZFW","zod-validation-expert","Expert in Zod — TypeScript-first schema validation. Covers parsing, custom errors, refinements, type inference, and integration with React Hook Form, Next.js, and tRPC.","https:\u002F\u002Fgithub.com\u002Fkalash33\u002Fkstack\u002Fblob\u002Fmain\u002Fskills\u002Fzod-validation-expert\u002FSKILL.md",[],[],{"skillNo":77,"name":78,"nameZh":6,"description":79,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":80,"categories":81,"keywords":82,"publishedAt":19},"01M73CF4YPMF3RA8ZJDXZSMD93","zeroize-audit","Detects missing zeroization of sensitive data in source code and identifies zeroization removed by compiler optimizations, with assembly-level analysis, and control-flow verification. Use for auditing C\u002FC++\u002FRust code handling secrets, keys, passwords, or other sensitive data.","https:\u002F\u002Fgithub.com\u002Fkalash33\u002Fkstack\u002Fblob\u002Fmain\u002Fskills\u002Fzeroize-audit\u002FSKILL.md",[],[],{"skillNo":84,"name":85,"nameZh":6,"description":86,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":87,"categories":88,"keywords":89,"publishedAt":19},"01M73CF2S25EN3XNAMKY47XHPS","zendesk-automation","Automate Zendesk tasks via Rube MCP (Composio): tickets, users, organizations, replies. Always search tools first for current schemas.","https:\u002F\u002Fgithub.com\u002Fkalash33\u002Fkstack\u002Fblob\u002Fmain\u002Fskills\u002Fzendesk-automation\u002FSKILL.md",[],[]]