[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M6QVAFJ0S3MSJ4VDJ7D5Y7JJ":3,"skill-same:01M6QVAFJ0S3MSJ4VDJ7D5Y7JJ":69},{"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":39,"linkOnly":40,"markdownContent":41,"markdownBody":42,"frontMatter":43,"refName":22,"commitSha":44,"skillDirectory":45,"forksCount":14,"repositoryHtmlUrl":46,"repositoryHomepageUrl":6,"ownerHtmlUrl":47,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":48,"install":49,"resources":62},"01M6QVAFJ0S3MSJ4VDJ7D5Y7JJ","audit-worktree",null,"Audits an implementation worktree branch against main using a provided implementation summary, loads relevant project skills as needed, fixes all identified issues in the current worktree only, verifies the result, and reports the audit outcome. Use after script-driven worktree implementation when \u003Charness> context has been cleared and the user passes the prior implementation summary.","joe-salamy\u002Fharness-loadouts","joe-salamy","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F241636360?v=4","01M6QV9XR82S4QFQCPTJXK2KTZ","Joe Salamy","01M6QV9XWN08VFW40FDQAE3CHP",0,"MIT","https:\u002F\u002Fgithub.com\u002Fjoe-salamy\u002Fharness-loadouts\u002Fblob\u002Fmain\u002F.omp\u002Fskills\u002Faudit-worktree\u002FSKILL.md",[],[19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38],"audit worktree","worktree","audit","main","result","fixes","script","relevant","loads","as","project","summary","issues","implementation","reports","only","use","outcome","branch","harness","2026-11-10 10:12:30",false,"---\nname: audit-worktree\ndescription: Audits an implementation worktree branch against main using a provided implementation summary, loads relevant project skills as needed, fixes all identified issues in the current worktree only, verifies the result, and reports the audit outcome. Use after script-driven worktree implementation when \u003Charness> context has been cleared and the user passes the prior implementation summary.\n---\n\n# Audit Worktree\n\nUse this skill inside an implementation worktree after another \u003Charness> session has completed a plan and returned a full implementation summary. The goal is to audit the current branch against `main`, use the provided summary as intent\u002Fcontext, load relevant project skills as necessary, identify issues, fix them in this worktree, and verify the final state.\n\n## Required Input\n\n- The implementation summary from the prior script-driven worktree run.\n- The current shell location should be the implementation worktree, not the primary `main` checkout.\n\nIf the summary is missing, continue with the git diff but state that intent coverage is limited. If the current checkout is `main`, stop before editing and tell the user to run the skill from the implementation worktree.\n\n## Safety Rules\n\n- Do not edit the primary\u002Fmain worktree.\n- Do not switch to or edit `main`.\n- Do not merge, rebase, reset, or force-push unless the user explicitly asks.\n- Do not read, write, or diff `scratchpad.md`.\n- Preserve user changes you did not make.\n- Make all fixes only in the current implementation worktree.\n\n## Workflow\n\n1. Confirm the current worktree and branch:\n\n```powershell\ngit worktree list\ngit branch --show-current\ngit status --short\n```\n\nStop before editing if the current branch is `main` or if `git worktree list` shows the current path is the primary `main` checkout.\n\n2. Establish the comparison base:\n\n```powershell\ngit fetch --all --prune\ngit merge-base main HEAD\ngit diff --stat main...HEAD -- . \":(exclude)scratchpad.md\" \":(exclude)docs\u002Fscratchpad.md\"\ngit diff --name-only main...HEAD -- . \":(exclude)scratchpad.md\" \":(exclude)docs\u002Fscratchpad.md\"\n```\n\nIf local `main` is unavailable, inspect branches and use the repository default branch. Do not check out that branch.\n\n3. Read the provided implementation summary and map it to the changed files. Treat the summary as intent, not proof. Verify the actual diff.\n\n4. Audit the full diff against `main`:\n   - Check correctness, regressions, missing edge cases, missing tests, and incomplete plan items.\n   - Check this repo's architecture rules from `AGENTS.md`, especially CLI\u002Fweb behavioral alignment, `EssayPipeline.run_step()`, index-to-text replacement, `_metadata.input_messages`, Supabase RLS\u002Fstorage rules, prompt\u002Fschema pairing, and callback\u002Fusage tracking for LLM calls.\n   - Check that new behavior is represented in focused tests or explain why tests are not feasible.\n   - Inspect relevant existing code paths, not only changed lines, when needed to validate behavior.\n\n5. Load project skills as necessary based on the changed files and risk areas. Read the relevant `.agent-harness\u002Fskills\u002F\u003Cskill-name>\u002FSKILL.md` files directly, then follow their audit guidance. Common mappings:\n   - FastAPI routes, middleware, ARQ worker APIs: `api-review`\n   - SSE or HITL streaming: `sse-audit`\n   - Supabase migrations, RLS, storage, auth, SQL: `supabase`, `supabase-schema`, `supabase-postgres-best-practices`\n   - Python async, typing, pipeline code: `python-pro`\n   - Next.js, React, TypeScript, shadcn, frontend hooks: `ts-pro`, `audit-ts`, `vercel-react-best-practices`, `nextjs-shadcn`\n   - Pydantic schemas or generated frontend types: `cross-stack-types`\n   - Prompt templates or structured-output schemas: `prompt-schema-conventions`\n   - Docker or deployment files: `docker-review`\n   - Broad code quality\u002Fsecurity review: `code-reviewer`\n\n6. Create a concise issue list before editing. Prioritize bugs, security issues, data loss, broken contracts, missing migrations\u002FRLS, missing callback\u002Fusage tracking, CLI\u002Fweb drift, and test gaps.\n\n7. Fix all confirmed issues in the current worktree. Keep edits scoped to the audit findings. Do not make unrelated refactors.\n\n8. Re-run focused verification appropriate to the final diff. If the implementation worktree does not have its own `venv`, run Python\u002Ftest commands from the worktree but activate the original checkout's virtualenv path:\n\n```powershell\n& \u003Cprimary-checkout>\\venv\\Scripts\\Activate.ps1\n```\n\nKeep all file operations in the implementation worktree.\n\nFor frontend changes, prefer the existing package scripts from `frontend\u002Fpackage.json`. If generated frontend types are affected, run the documented generation flow only when the API server requirement can be satisfied; otherwise report the skipped generation clearly.\n\n9. Commit all intended audit fixes on the current branch if changes were made:\n\n```powershell\ngit status --short\ngit add \u003Cfixed-files>\ngit commit -m \"Fix audit findings\"\ngit status --short\n```\n\nDo not include `scratchpad.md` or workflow handoff artifacts. If no fixes were needed, do not create an empty commit. The workflow merge only receives committed branch changes; uncommitted or unstaged files left in the worktree will not be included in the squash\u002Fno-ff integration merge.\n\n10. Final response must include:\n\n- Worktree path\n- Branch name\n- Base branch\u002Fref used for diff\n- Summary received from the prior implementation, briefly restated\n- Skills loaded and why\n- Issues found and fixes applied\n- Files changed by the audit\n- Commit SHA for audit fixes, or note that no commit was created\n- Tests\u002Fchecks run and any skipped checks\n- Residual risks or follow-up work\n\n## Output Standard\n\nBe explicit about both findings and fixes. If an issue is suspected but not confirmed, label it as a residual risk instead of silently editing around it. The final report should let a reviewer understand what changed in the audit pass without reopening the entire diff.\n","# Audit Worktree\n\nUse this skill inside an implementation worktree after another \u003Charness> session has completed a plan and returned a full implementation summary. The goal is to audit the current branch against `main`, use the provided summary as intent\u002Fcontext, load relevant project skills as necessary, identify issues, fix them in this worktree, and verify the final state.\n\n## Required Input\n\n- The implementation summary from the prior script-driven worktree run.\n- The current shell location should be the implementation worktree, not the primary `main` checkout.\n\nIf the summary is missing, continue with the git diff but state that intent coverage is limited. If the current checkout is `main`, stop before editing and tell the user to run the skill from the implementation worktree.\n\n## Safety Rules\n\n- Do not edit the primary\u002Fmain worktree.\n- Do not switch to or edit `main`.\n- Do not merge, rebase, reset, or force-push unless the user explicitly asks.\n- Do not read, write, or diff `scratchpad.md`.\n- Preserve user changes you did not make.\n- Make all fixes only in the current implementation worktree.\n\n## Workflow\n\n1. Confirm the current worktree and branch:\n\n```powershell\ngit worktree list\ngit branch --show-current\ngit status --short\n```\n\nStop before editing if the current branch is `main` or if `git worktree list` shows the current path is the primary `main` checkout.\n\n2. Establish the comparison base:\n\n```powershell\ngit fetch --all --prune\ngit merge-base main HEAD\ngit diff --stat main...HEAD -- . \":(exclude)scratchpad.md\" \":(exclude)docs\u002Fscratchpad.md\"\ngit diff --name-only main...HEAD -- . \":(exclude)scratchpad.md\" \":(exclude)docs\u002Fscratchpad.md\"\n```\n\nIf local `main` is unavailable, inspect branches and use the repository default branch. Do not check out that branch.\n\n3. Read the provided implementation summary and map it to the changed files. Treat the summary as intent, not proof. Verify the actual diff.\n\n4. Audit the full diff against `main`:\n   - Check correctness, regressions, missing edge cases, missing tests, and incomplete plan items.\n   - Check this repo's architecture rules from `AGENTS.md`, especially CLI\u002Fweb behavioral alignment, `EssayPipeline.run_step()`, index-to-text replacement, `_metadata.input_messages`, Supabase RLS\u002Fstorage rules, prompt\u002Fschema pairing, and callback\u002Fusage tracking for LLM calls.\n   - Check that new behavior is represented in focused tests or explain why tests are not feasible.\n   - Inspect relevant existing code paths, not only changed lines, when needed to validate behavior.\n\n5. Load project skills as necessary based on the changed files and risk areas. Read the relevant `.agent-harness\u002Fskills\u002F\u003Cskill-name>\u002FSKILL.md` files directly, then follow their audit guidance. Common mappings:\n   - FastAPI routes, middleware, ARQ worker APIs: `api-review`\n   - SSE or HITL streaming: `sse-audit`\n   - Supabase migrations, RLS, storage, auth, SQL: `supabase`, `supabase-schema`, `supabase-postgres-best-practices`\n   - Python async, typing, pipeline code: `python-pro`\n   - Next.js, React, TypeScript, shadcn, frontend hooks: `ts-pro`, `audit-ts`, `vercel-react-best-practices`, `nextjs-shadcn`\n   - Pydantic schemas or generated frontend types: `cross-stack-types`\n   - Prompt templates or structured-output schemas: `prompt-schema-conventions`\n   - Docker or deployment files: `docker-review`\n   - Broad code quality\u002Fsecurity review: `code-reviewer`\n\n6. Create a concise issue list before editing. Prioritize bugs, security issues, data loss, broken contracts, missing migrations\u002FRLS, missing callback\u002Fusage tracking, CLI\u002Fweb drift, and test gaps.\n\n7. Fix all confirmed issues in the current worktree. Keep edits scoped to the audit findings. Do not make unrelated refactors.\n\n8. Re-run focused verification appropriate to the final diff. If the implementation worktree does not have its own `venv`, run Python\u002Ftest commands from the worktree but activate the original checkout's virtualenv path:\n\n```powershell\n& \u003Cprimary-checkout>\\venv\\Scripts\\Activate.ps1\n```\n\nKeep all file operations in the implementation worktree.\n\nFor frontend changes, prefer the existing package scripts from `frontend\u002Fpackage.json`. If generated frontend types are affected, run the documented generation flow only when the API server requirement can be satisfied; otherwise report the skipped generation clearly.\n\n9. Commit all intended audit fixes on the current branch if changes were made:\n\n```powershell\ngit status --short\ngit add \u003Cfixed-files>\ngit commit -m \"Fix audit findings\"\ngit status --short\n```\n\nDo not include `scratchpad.md` or workflow handoff artifacts. If no fixes were needed, do not create an empty commit. The workflow merge only receives committed branch changes; uncommitted or unstaged files left in the worktree will not be included in the squash\u002Fno-ff integration merge.\n\n10. Final response must include:\n\n- Worktree path\n- Branch name\n- Base branch\u002Fref used for diff\n- Summary received from the prior implementation, briefly restated\n- Skills loaded and why\n- Issues found and fixes applied\n- Files changed by the audit\n- Commit SHA for audit fixes, or note that no commit was created\n- Tests\u002Fchecks run and any skipped checks\n- Residual risks or follow-up work\n\n## Output Standard\n\nBe explicit about both findings and fixes. If an issue is suspected but not confirmed, label it as a residual risk instead of silently editing around it. The final report should let a reviewer understand what changed in the audit pass without reopening the entire diff.\n",{"name":5,"description":7},"c5d81adf56b155b93b6015e0e15e0d8f9ed8d62f",".omp\u002Fskills\u002Faudit-worktree","https:\u002F\u002Fgithub.com\u002Fjoe-salamy\u002Fharness-loadouts","https:\u002F\u002Fgithub.com\u002Fjoe-salamy","2026-09-24 09:29:45",{"source":8,"skillName":5,"commands":50,"prompt":60,"zipAvailable":61},[51,54,57],{"tool":52,"command":53},"npx","npx skills add joe-salamy\u002Fharness-loadouts --skill audit-worktree",{"tool":55,"command":56},"bunx","bunx skills add joe-salamy\u002Fharness-loadouts --skill audit-worktree",{"tool":58,"command":59},"pnpm","pnpm dlx skills add joe-salamy\u002Fharness-loadouts --skill audit-worktree","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fjoe-salamy\u002Fharness-loadouts\u002Fblob\u002Fmain\u002F.omp\u002Fskills\u002Faudit-worktree\u002FSKILL.md",true,[63],{"path":64,"resourceType":65,"resourceTypeDesc":66,"sizeBytes":67,"sourceHtmlUrl":68},"agents\u002Fopenai.yaml","REFERENCE","参考文档",473,"https:\u002F\u002Fgithub.com\u002Fjoe-salamy\u002Fharness-loadouts\u002Fblob\u002Fmain\u002F.omp\u002Fskills\u002Faudit-worktree\u002Fagents\u002Fopenai.yaml",[70,86,92,123,137,159],{"skillNo":71,"name":72,"nameZh":6,"description":73,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":74,"categories":75,"keywords":81,"publishedAt":39},"01M6QVC87213Q24845Q5W8FXRJ","example-skill","Replace this with your skill description","https:\u002F\u002Fgithub.com\u002Fjoe-salamy\u002Fharness-loadouts\u002Fblob\u002Fmain\u002Fexamples\u002Ftypescript\u002F.harness\u002Fskills\u002Fexample-skill\u002FSKILL.md",[76],{"categoryNo":77,"categoryName":78,"taxonomyType":79,"taxonomyTypeDesc":80},"01JBSEEDCAT000000000000023","Skill与智能体开发","SCENARIO","使用场景",[82,83,84,85],"example","example skill","your skill","with",{"skillNo":87,"name":72,"nameZh":6,"description":73,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":88,"categories":89,"keywords":91,"publishedAt":39},"01M6QVC2XCSAFTK7X7QVH094XR","https:\u002F\u002Fgithub.com\u002Fjoe-salamy\u002Fharness-loadouts\u002Fblob\u002Fmain\u002Fexamples\u002Fpython\u002F.harness\u002Fskills\u002Fexample-skill\u002FSKILL.md",[90],{"categoryNo":77,"categoryName":78,"taxonomyType":79,"taxonomyTypeDesc":80},[82,83,84,85],{"skillNo":93,"name":94,"nameZh":6,"description":95,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":96,"categories":97,"keywords":104,"publishedAt":39},"01M6QVBP9NC2T2D6JP8V8A8RNY","publish-polish","Final audit and polish workflow for converting a personal Python project into a publication-ready open-source GitHub repository. Checks security and privacy (including git history), code quality, documentation, packaging metadata, and repo hygiene, then produces a structured readiness report with prioritized findings. Invoke before making a private repo public or before a first release.","https:\u002F\u002Fgithub.com\u002Fjoe-salamy\u002Fharness-loadouts\u002Fblob\u002Fmain\u002F.omp\u002Fskills\u002Fpublish-polish\u002FSKILL.md",[98,99],{"categoryNo":77,"categoryName":78,"taxonomyType":79,"taxonomyTypeDesc":80},{"categoryNo":100,"categoryName":101,"taxonomyType":102,"taxonomyTypeDesc":103},"01JBSEEDCAT000000000000034","行政与办公人员","PROFESSION","适用职业",[105,106,107,108,109,110,111,112,113,114,115,116,21,117,118,119,120,121,29,122],"publish","polish","publish polish","go","workflow","documentation","cleanup","code","release","git","invoke","open source","repository","open","reviewer","packaging","ready","python project",{"skillNo":124,"name":125,"nameZh":6,"description":126,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":127,"categories":128,"keywords":129,"publishedAt":39},"01M6QVBJWV7ZA80CSSN54P92E0","optimize-repo-skills","Audits a repository's harness skills against the actual codebase, proposes conservative skill updates, and after explicit approval updates, creates, or removes repo-specific skills.","https:\u002F\u002Fgithub.com\u002Fjoe-salamy\u002Fharness-loadouts\u002Fblob\u002Fmain\u002F.omp\u002Fskills\u002Foptimize-repo-skills\u002FSKILL.md",[],[130,131,132,133,134,38,117,135,136],"optimize","optimize repo skills","repo","codebase","updates","approval","skill updates",{"skillNo":138,"name":139,"nameZh":6,"description":140,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":141,"categories":142,"keywords":143,"publishedAt":39},"01M6QVBD4V3J6Z7KDGJ8C0M5YF","merge-conflict-resolver","Resolves merge conflicts inside a temporary integration worktree created by the workflow script, preserves approved plan intent and latest base behavior, and writes a conflict-resolution summary without committing.","https:\u002F\u002Fgithub.com\u002Fjoe-salamy\u002Fharness-loadouts\u002Fblob\u002Fmain\u002F.omp\u002Fskills\u002Fmerge-conflict-resolver\u002FSKILL.md",[],[144,145,146,147,148,149,109,20,150,151,152,153,154,30,155,156,25,157,158],"resolver","merge conflict resolver","conflict","conflict resolver","merge","plan","integration","behavior","merge conflicts","conflicts","intent","base","resolution","latest","conflict-resolution",{"skillNo":160,"name":161,"nameZh":6,"description":162,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":163,"categories":164,"keywords":165,"publishedAt":39},"01M6QVB7EW7GQEAJGD8NRXFEGF","implement-worktree","Implements an approved Markdown plan inside the current script-created git worktree, verifies the result, commits the implementation, and writes a handoff summary for audit.","https:\u002F\u002Fgithub.com\u002Fjoe-salamy\u002Fharness-loadouts\u002Fblob\u002Fmain\u002F.omp\u002Fskills\u002Fimplement-worktree\u002FSKILL.md",[],[166,167,20,168,114,21,169,170,171,149,32,30,25,172,23],"implement","implement worktree","handoff","commits","markdown","git worktree","for"]