[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill-microsoft-council-review":3,"mdc--ocsjl7-key":36,"related-org-microsoft-council-review":577,"related-repo-microsoft-council-review":772},{"slug":4,"name":4,"fn":5,"description":6,"org":7,"tags":11,"stars":25,"repoUrl":26,"updatedAt":27,"license":28,"forks":29,"topics":30,"repo":31,"sourceUrl":34,"mdContent":35},"council-review","perform multi-model council code reviews","Multi-model council review for diffs, pull requests, and risky changes. It is CRITICAL to use this skill after ANY non-trivial development phase — changes are usually not 100% correct. Development CANNOT be called done until the code has been reviewed. Use whenever the user asks for a review, PR feedback, a bug hunt in recent edits, or wants multiple models to inspect the same change independently. Also use when the user asks to 'discuss findings', 'cross-review', or 'debate' review results between models.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},"microsoft","Microsoft","https:\u002F\u002Fpexgzepcugksgbtrxkhf.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Forg-logos\u002Fmicrosoft.png",[12,16,19,22],{"name":13,"slug":14,"type":15},"Code Review","code-review","tag",{"name":17,"slug":18,"type":15},"Engineering","engineering",{"name":20,"slug":21,"type":15},"Multi-Agent","multi-agent",{"name":23,"slug":24,"type":15},"Pull Requests","pull-requests",25,"https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fvscode-team-kit","2026-05-13T06:14:07.202801",null,5,[],{"repoUrl":26,"stars":25,"forks":29,"topics":32,"description":33},[],"Plugins the VS Code team uses","https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fvscode-team-kit\u002Ftree\u002FHEAD\u002Fmodel-council\u002Fskills\u002Fcouncil-review","---\nname: council-review\ndescription: \"Multi-model council review for diffs, pull requests, and risky changes. It is CRITICAL to use this skill after ANY non-trivial development phase — changes are usually not 100% correct. Development CANNOT be called done until the code has been reviewed. Use whenever the user asks for a review, PR feedback, a bug hunt in recent edits, or wants multiple models to inspect the same change independently. Also use when the user asks to 'discuss findings', 'cross-review', or 'debate' review results between models.\"\n---\n\n# Skill: Review Council\n\nRead-only review powered by a council of model-pinned reviewers. Changes are usually not 100% correct — this skill exists to catch what slipped through. The goal is not broad commentary; it is a short list of concrete issues where the council agrees, plus transparent disclosure of where they disagree.\n\n## Council Models\n\nSpawn subagents with the `model` parameter to pin each to a different model. Use all three when available, at least two otherwise:\n\n- `GPT-5.5`\n- `Claude Opus 4.6`\n- `GPT-5.3-Codex`\n\nEach subagent receives the same system preamble:\n\n> You are an independent subagent for read-only code research.\n> MUST stay read-only. Stay within the requested scope.\n> Do not speculate. Do not suggest patches.\n> For every finding, cite exact files and lines.\n> Form your own view from first principles. Do not anchor to the provided context.\n\n## Workflow\n\n### Phase 1 — Scope\n\n- Prefer the active pull request, current git diff, completed edits during the session, or an explicit file list from the user.\n- If there is nothing to review, ask the user to point at a branch, PR, diff, or file set.\n\n### Phase 2 — Build a Change Summary\n\nBefore fanning out, build a **preliminary orientation** that each reviewer will receive. This is a starting point, not ground truth — reviewers are expected to contradict it if their own investigation leads elsewhere. Do **not** paste the raw diff — reviewers have tools to read code themselves.\n\nThe summary should include only:\n1. **Intent** — what the change is trying to accomplish (from current plan, PR description, commit messages, or conversation context).\n2. **Changed files** — list each file path with a one-line description of what was touched (e.g., \"modified\", \"added\", \"deleted\") — not what the reviewer should think about it.\n3. **How to inspect** — tell the reviewers the branch\u002FPR\u002Fcommit info so they can use their tools to read the actual code.\n\nKeep the summary under ~50 lines. Do not flag risk areas or pre-diagnose issues — leave reviewers to form independent assessments. They get better results reading code in context than scanning a pre-interpreted summary.\n\n### Phase 3 — Independent First-Pass Reviews\n\nSpawn all subagents in parallel at once:\n- One subagent per model using the `model` parameter. Prepend the system preamble to the reviewer prompt below.\n- Do **not** show one reviewer's findings to another before this phase completes.\n\n### Phase 4 — Consensus Synthesis\n\nClassify every finding into one of three buckets:\n\n1. **Consensus** — two or more reviewers independently flag the same underlying issue. These are high-confidence findings.\n2. **Single-reviewer, strongly evidenced** — only one reviewer flags it, but the evidence is concrete (compile error, failing test, clear correctness bug, security vulnerability, measurable regression). Keep these with a note that only one model flagged them.\n3. **Contested or speculative** — a finding one reviewer flags that another reviewer's analysis implicitly or explicitly contradicts. Mention these briefly under \"Contested\" so the user can decide, but do not present them as confirmed issues.\n\nDrop style chatter, linter-catchable nits, and low-confidence speculation entirely.\n\n### Phase 5 — Cross-Review (optional)\n\nTrigger this phase when the user asks to \"discuss\", \"debate\", or \"cross-review\" the findings, or when Phase 4 produces contested findings that could benefit from a second look.\n\n1. Share the anonymized synthesis (without attributing which model said what) back to each council agent.\n2. Ask each agent: _\"Review these findings. For each, state whether you agree, disagree, or want to add nuance. Cite evidence.\"_\n3. Spawn all subagents in parallel at once.\n4. Update the synthesis: promote contested findings to consensus if the cross-review resolves the disagreement, or note the persisting split.\n\nThis deliberation step is lightweight — it only re-examines the synthesized findings, not the full codebase.\n\n## Reviewer Prompt\n\nUse this prompt shape for each council agent. Pass the change summary from Phase 2 — never paste raw diffs.\n\n```text\nReview this change set in read-only mode.\n\n## What changed\n\u003Cchange summary from Phase 2: intent, changed files with one-line descriptions>\n\n## How to inspect\n\u003Cbranch, PR number, or commit range the reviewer can use with their tools>\n\nUse your tools to read the changed files, check diagnostics, and run focused tests. Do not rely solely on this summary — dig into the code yourself.\n\n**The summary above is preliminary orientation, not ground truth.** If your own investigation contradicts the stated intent or file descriptions, trust your findings over the summary and say so explicitly.\n\nReturn only high-signal findings that would block approval or clearly require follow-up.\nFocus on correctness, regressions, security, missing tests for risky behavior changes, and concrete performance problems.\nDo not suggest code edits.\nDo not report style issues or speculative concerns.\nCite exact files and lines for every finding.\nIf you do not find a blocking issue, say \"No blocking issues found.\"\n```\n\n## Cross-Review Prompt\n\nUse this prompt shape when running Phase 5:\n\n```text\nThe following findings were produced by a council of independent reviewers for this change set:\n\n\u003Csynthesized findings, without model attribution>\n\nFor each finding, state whether you:\n- AGREE (the issue is real and correctly described)\n- DISAGREE (explain why with evidence)\n- NUANCE (the issue is partially correct but needs clarification)\n\nDo not introduce new findings. Stay read-only. Cite files and lines.\n```\n\n## Saving Findings\n\nAfter synthesis, always save the findings to the session folder as `review.md`. This makes them available for follow-up turns, fix planning, and cross-referencing with future reviews.\n\n## Default Behavior: Review → Fix\n\nBy default, this skill is biased towards **addressing** the issues it finds, not just reporting them. The review is the means, not the end.\n\nAfter synthesis, determine the mode:\n\n- **Review-only mode** — the user said \"only review\", \"just review\", \"don't change anything\", \"read-only\", or similar. Stop after the output shape below. Do not edit files.\n- **Review-and-fix mode** (default) — plan and apply fixes for consensus and single-reviewer findings after presenting them.\n\nWhen in review-and-fix mode:\n1. Present the findings first (same output shape as below) so the user sees what was found.\n2. For each fixable finding, launch one or more parallel Explore subagents to investigate the fix.\n3. Apply the fixes in severity order.\n4. Skip contested findings — those need the user's judgment.\n5. After applying fixes, run a lightweight re-review (read the changed lines, check diagnostics) to validate.\n6. Report what was fixed and any remaining items that need the user's input.\n\nThis keeps the feedback loop tight: review surfaces the problems, Explore agents plan the fixes, then the same turn resolves them.\n\n## Output Shape\n\n### After Phase 4 (standard review)\n\n**Consensus findings** — issues independently identified by multiple reviewers, ordered by severity. Each with a concise explanation and file references.\n\n**Single-reviewer findings** — high-evidence issues flagged by only one model. Note which model flagged it.\n\n**Contested** (if any) — briefly list findings where reviewers disagreed, with a one-line summary of the disagreement. Suggest running a cross-review if the user wants resolution.\n\n**Summary** — one short paragraph. If no blocking issues survive synthesis, say so explicitly and mention any meaningful testing gaps.\n\n### After Phase 5 (cross-review)\n\nSame structure as above, but update the buckets based on the deliberation:\n- Contested findings that reached agreement move to Consensus or get dropped.\n- Note any findings where the split persists after deliberation.",{"data":37,"body":38},{"name":4,"description":6},{"type":39,"children":40},"root",[41,50,56,63,77,109,114,123,129,136,149,155,175,180,214,219,225,230,256,262,267,300,305,311,316,345,350,356,361,373,379,384,393,399,412,418,430,435,458,463,496,501,507,513,523,533,543,553,559,564],{"type":42,"tag":43,"props":44,"children":46},"element","h1",{"id":45},"skill-review-council",[47],{"type":48,"value":49},"text","Skill: Review Council",{"type":42,"tag":51,"props":52,"children":53},"p",{},[54],{"type":48,"value":55},"Read-only review powered by a council of model-pinned reviewers. Changes are usually not 100% correct — this skill exists to catch what slipped through. The goal is not broad commentary; it is a short list of concrete issues where the council agrees, plus transparent disclosure of where they disagree.",{"type":42,"tag":57,"props":58,"children":60},"h2",{"id":59},"council-models",[61],{"type":48,"value":62},"Council Models",{"type":42,"tag":51,"props":64,"children":65},{},[66,68,75],{"type":48,"value":67},"Spawn subagents with the ",{"type":42,"tag":69,"props":70,"children":72},"code",{"className":71},[],[73],{"type":48,"value":74},"model",{"type":48,"value":76}," parameter to pin each to a different model. Use all three when available, at least two otherwise:",{"type":42,"tag":78,"props":79,"children":80},"ul",{},[81,91,100],{"type":42,"tag":82,"props":83,"children":84},"li",{},[85],{"type":42,"tag":69,"props":86,"children":88},{"className":87},[],[89],{"type":48,"value":90},"GPT-5.5",{"type":42,"tag":82,"props":92,"children":93},{},[94],{"type":42,"tag":69,"props":95,"children":97},{"className":96},[],[98],{"type":48,"value":99},"Claude Opus 4.6",{"type":42,"tag":82,"props":101,"children":102},{},[103],{"type":42,"tag":69,"props":104,"children":106},{"className":105},[],[107],{"type":48,"value":108},"GPT-5.3-Codex",{"type":42,"tag":51,"props":110,"children":111},{},[112],{"type":48,"value":113},"Each subagent receives the same system preamble:",{"type":42,"tag":115,"props":116,"children":117},"blockquote",{},[118],{"type":42,"tag":51,"props":119,"children":120},{},[121],{"type":48,"value":122},"You are an independent subagent for read-only code research.\nMUST stay read-only. Stay within the requested scope.\nDo not speculate. Do not suggest patches.\nFor every finding, cite exact files and lines.\nForm your own view from first principles. Do not anchor to the provided context.",{"type":42,"tag":57,"props":124,"children":126},{"id":125},"workflow",[127],{"type":48,"value":128},"Workflow",{"type":42,"tag":130,"props":131,"children":133},"h3",{"id":132},"phase-1-scope",[134],{"type":48,"value":135},"Phase 1 — Scope",{"type":42,"tag":78,"props":137,"children":138},{},[139,144],{"type":42,"tag":82,"props":140,"children":141},{},[142],{"type":48,"value":143},"Prefer the active pull request, current git diff, completed edits during the session, or an explicit file list from the user.",{"type":42,"tag":82,"props":145,"children":146},{},[147],{"type":48,"value":148},"If there is nothing to review, ask the user to point at a branch, PR, diff, or file set.",{"type":42,"tag":130,"props":150,"children":152},{"id":151},"phase-2-build-a-change-summary",[153],{"type":48,"value":154},"Phase 2 — Build a Change Summary",{"type":42,"tag":51,"props":156,"children":157},{},[158,160,166,168,173],{"type":48,"value":159},"Before fanning out, build a ",{"type":42,"tag":161,"props":162,"children":163},"strong",{},[164],{"type":48,"value":165},"preliminary orientation",{"type":48,"value":167}," that each reviewer will receive. This is a starting point, not ground truth — reviewers are expected to contradict it if their own investigation leads elsewhere. Do ",{"type":42,"tag":161,"props":169,"children":170},{},[171],{"type":48,"value":172},"not",{"type":48,"value":174}," paste the raw diff — reviewers have tools to read code themselves.",{"type":42,"tag":51,"props":176,"children":177},{},[178],{"type":48,"value":179},"The summary should include only:",{"type":42,"tag":181,"props":182,"children":183},"ol",{},[184,194,204],{"type":42,"tag":82,"props":185,"children":186},{},[187,192],{"type":42,"tag":161,"props":188,"children":189},{},[190],{"type":48,"value":191},"Intent",{"type":48,"value":193}," — what the change is trying to accomplish (from current plan, PR description, commit messages, or conversation context).",{"type":42,"tag":82,"props":195,"children":196},{},[197,202],{"type":42,"tag":161,"props":198,"children":199},{},[200],{"type":48,"value":201},"Changed files",{"type":48,"value":203}," — list each file path with a one-line description of what was touched (e.g., \"modified\", \"added\", \"deleted\") — not what the reviewer should think about it.",{"type":42,"tag":82,"props":205,"children":206},{},[207,212],{"type":42,"tag":161,"props":208,"children":209},{},[210],{"type":48,"value":211},"How to inspect",{"type":48,"value":213}," — tell the reviewers the branch\u002FPR\u002Fcommit info so they can use their tools to read the actual code.",{"type":42,"tag":51,"props":215,"children":216},{},[217],{"type":48,"value":218},"Keep the summary under ~50 lines. Do not flag risk areas or pre-diagnose issues — leave reviewers to form independent assessments. They get better results reading code in context than scanning a pre-interpreted summary.",{"type":42,"tag":130,"props":220,"children":222},{"id":221},"phase-3-independent-first-pass-reviews",[223],{"type":48,"value":224},"Phase 3 — Independent First-Pass Reviews",{"type":42,"tag":51,"props":226,"children":227},{},[228],{"type":48,"value":229},"Spawn all subagents in parallel at once:",{"type":42,"tag":78,"props":231,"children":232},{},[233,245],{"type":42,"tag":82,"props":234,"children":235},{},[236,238,243],{"type":48,"value":237},"One subagent per model using the ",{"type":42,"tag":69,"props":239,"children":241},{"className":240},[],[242],{"type":48,"value":74},{"type":48,"value":244}," parameter. Prepend the system preamble to the reviewer prompt below.",{"type":42,"tag":82,"props":246,"children":247},{},[248,250,254],{"type":48,"value":249},"Do ",{"type":42,"tag":161,"props":251,"children":252},{},[253],{"type":48,"value":172},{"type":48,"value":255}," show one reviewer's findings to another before this phase completes.",{"type":42,"tag":130,"props":257,"children":259},{"id":258},"phase-4-consensus-synthesis",[260],{"type":48,"value":261},"Phase 4 — Consensus Synthesis",{"type":42,"tag":51,"props":263,"children":264},{},[265],{"type":48,"value":266},"Classify every finding into one of three buckets:",{"type":42,"tag":181,"props":268,"children":269},{},[270,280,290],{"type":42,"tag":82,"props":271,"children":272},{},[273,278],{"type":42,"tag":161,"props":274,"children":275},{},[276],{"type":48,"value":277},"Consensus",{"type":48,"value":279}," — two or more reviewers independently flag the same underlying issue. These are high-confidence findings.",{"type":42,"tag":82,"props":281,"children":282},{},[283,288],{"type":42,"tag":161,"props":284,"children":285},{},[286],{"type":48,"value":287},"Single-reviewer, strongly evidenced",{"type":48,"value":289}," — only one reviewer flags it, but the evidence is concrete (compile error, failing test, clear correctness bug, security vulnerability, measurable regression). Keep these with a note that only one model flagged them.",{"type":42,"tag":82,"props":291,"children":292},{},[293,298],{"type":42,"tag":161,"props":294,"children":295},{},[296],{"type":48,"value":297},"Contested or speculative",{"type":48,"value":299}," — a finding one reviewer flags that another reviewer's analysis implicitly or explicitly contradicts. Mention these briefly under \"Contested\" so the user can decide, but do not present them as confirmed issues.",{"type":42,"tag":51,"props":301,"children":302},{},[303],{"type":48,"value":304},"Drop style chatter, linter-catchable nits, and low-confidence speculation entirely.",{"type":42,"tag":130,"props":306,"children":308},{"id":307},"phase-5-cross-review-optional",[309],{"type":48,"value":310},"Phase 5 — Cross-Review (optional)",{"type":42,"tag":51,"props":312,"children":313},{},[314],{"type":48,"value":315},"Trigger this phase when the user asks to \"discuss\", \"debate\", or \"cross-review\" the findings, or when Phase 4 produces contested findings that could benefit from a second look.",{"type":42,"tag":181,"props":317,"children":318},{},[319,324,335,340],{"type":42,"tag":82,"props":320,"children":321},{},[322],{"type":48,"value":323},"Share the anonymized synthesis (without attributing which model said what) back to each council agent.",{"type":42,"tag":82,"props":325,"children":326},{},[327,329],{"type":48,"value":328},"Ask each agent: ",{"type":42,"tag":330,"props":331,"children":332},"em",{},[333],{"type":48,"value":334},"\"Review these findings. For each, state whether you agree, disagree, or want to add nuance. Cite evidence.\"",{"type":42,"tag":82,"props":336,"children":337},{},[338],{"type":48,"value":339},"Spawn all subagents in parallel at once.",{"type":42,"tag":82,"props":341,"children":342},{},[343],{"type":48,"value":344},"Update the synthesis: promote contested findings to consensus if the cross-review resolves the disagreement, or note the persisting split.",{"type":42,"tag":51,"props":346,"children":347},{},[348],{"type":48,"value":349},"This deliberation step is lightweight — it only re-examines the synthesized findings, not the full codebase.",{"type":42,"tag":57,"props":351,"children":353},{"id":352},"reviewer-prompt",[354],{"type":48,"value":355},"Reviewer Prompt",{"type":42,"tag":51,"props":357,"children":358},{},[359],{"type":48,"value":360},"Use this prompt shape for each council agent. Pass the change summary from Phase 2 — never paste raw diffs.",{"type":42,"tag":362,"props":363,"children":368},"pre",{"className":364,"code":366,"language":48,"meta":367},[365],"language-text","Review this change set in read-only mode.\n\n## What changed\n\u003Cchange summary from Phase 2: intent, changed files with one-line descriptions>\n\n## How to inspect\n\u003Cbranch, PR number, or commit range the reviewer can use with their tools>\n\nUse your tools to read the changed files, check diagnostics, and run focused tests. Do not rely solely on this summary — dig into the code yourself.\n\n**The summary above is preliminary orientation, not ground truth.** If your own investigation contradicts the stated intent or file descriptions, trust your findings over the summary and say so explicitly.\n\nReturn only high-signal findings that would block approval or clearly require follow-up.\nFocus on correctness, regressions, security, missing tests for risky behavior changes, and concrete performance problems.\nDo not suggest code edits.\nDo not report style issues or speculative concerns.\nCite exact files and lines for every finding.\nIf you do not find a blocking issue, say \"No blocking issues found.\"\n","",[369],{"type":42,"tag":69,"props":370,"children":371},{"__ignoreMap":367},[372],{"type":48,"value":366},{"type":42,"tag":57,"props":374,"children":376},{"id":375},"cross-review-prompt",[377],{"type":48,"value":378},"Cross-Review Prompt",{"type":42,"tag":51,"props":380,"children":381},{},[382],{"type":48,"value":383},"Use this prompt shape when running Phase 5:",{"type":42,"tag":362,"props":385,"children":388},{"className":386,"code":387,"language":48,"meta":367},[365],"The following findings were produced by a council of independent reviewers for this change set:\n\n\u003Csynthesized findings, without model attribution>\n\nFor each finding, state whether you:\n- AGREE (the issue is real and correctly described)\n- DISAGREE (explain why with evidence)\n- NUANCE (the issue is partially correct but needs clarification)\n\nDo not introduce new findings. Stay read-only. Cite files and lines.\n",[389],{"type":42,"tag":69,"props":390,"children":391},{"__ignoreMap":367},[392],{"type":48,"value":387},{"type":42,"tag":57,"props":394,"children":396},{"id":395},"saving-findings",[397],{"type":48,"value":398},"Saving Findings",{"type":42,"tag":51,"props":400,"children":401},{},[402,404,410],{"type":48,"value":403},"After synthesis, always save the findings to the session folder as ",{"type":42,"tag":69,"props":405,"children":407},{"className":406},[],[408],{"type":48,"value":409},"review.md",{"type":48,"value":411},". This makes them available for follow-up turns, fix planning, and cross-referencing with future reviews.",{"type":42,"tag":57,"props":413,"children":415},{"id":414},"default-behavior-review-fix",[416],{"type":48,"value":417},"Default Behavior: Review → Fix",{"type":42,"tag":51,"props":419,"children":420},{},[421,423,428],{"type":48,"value":422},"By default, this skill is biased towards ",{"type":42,"tag":161,"props":424,"children":425},{},[426],{"type":48,"value":427},"addressing",{"type":48,"value":429}," the issues it finds, not just reporting them. The review is the means, not the end.",{"type":42,"tag":51,"props":431,"children":432},{},[433],{"type":48,"value":434},"After synthesis, determine the mode:",{"type":42,"tag":78,"props":436,"children":437},{},[438,448],{"type":42,"tag":82,"props":439,"children":440},{},[441,446],{"type":42,"tag":161,"props":442,"children":443},{},[444],{"type":48,"value":445},"Review-only mode",{"type":48,"value":447}," — the user said \"only review\", \"just review\", \"don't change anything\", \"read-only\", or similar. Stop after the output shape below. Do not edit files.",{"type":42,"tag":82,"props":449,"children":450},{},[451,456],{"type":42,"tag":161,"props":452,"children":453},{},[454],{"type":48,"value":455},"Review-and-fix mode",{"type":48,"value":457}," (default) — plan and apply fixes for consensus and single-reviewer findings after presenting them.",{"type":42,"tag":51,"props":459,"children":460},{},[461],{"type":48,"value":462},"When in review-and-fix mode:",{"type":42,"tag":181,"props":464,"children":465},{},[466,471,476,481,486,491],{"type":42,"tag":82,"props":467,"children":468},{},[469],{"type":48,"value":470},"Present the findings first (same output shape as below) so the user sees what was found.",{"type":42,"tag":82,"props":472,"children":473},{},[474],{"type":48,"value":475},"For each fixable finding, launch one or more parallel Explore subagents to investigate the fix.",{"type":42,"tag":82,"props":477,"children":478},{},[479],{"type":48,"value":480},"Apply the fixes in severity order.",{"type":42,"tag":82,"props":482,"children":483},{},[484],{"type":48,"value":485},"Skip contested findings — those need the user's judgment.",{"type":42,"tag":82,"props":487,"children":488},{},[489],{"type":48,"value":490},"After applying fixes, run a lightweight re-review (read the changed lines, check diagnostics) to validate.",{"type":42,"tag":82,"props":492,"children":493},{},[494],{"type":48,"value":495},"Report what was fixed and any remaining items that need the user's input.",{"type":42,"tag":51,"props":497,"children":498},{},[499],{"type":48,"value":500},"This keeps the feedback loop tight: review surfaces the problems, Explore agents plan the fixes, then the same turn resolves them.",{"type":42,"tag":57,"props":502,"children":504},{"id":503},"output-shape",[505],{"type":48,"value":506},"Output Shape",{"type":42,"tag":130,"props":508,"children":510},{"id":509},"after-phase-4-standard-review",[511],{"type":48,"value":512},"After Phase 4 (standard review)",{"type":42,"tag":51,"props":514,"children":515},{},[516,521],{"type":42,"tag":161,"props":517,"children":518},{},[519],{"type":48,"value":520},"Consensus findings",{"type":48,"value":522}," — issues independently identified by multiple reviewers, ordered by severity. Each with a concise explanation and file references.",{"type":42,"tag":51,"props":524,"children":525},{},[526,531],{"type":42,"tag":161,"props":527,"children":528},{},[529],{"type":48,"value":530},"Single-reviewer findings",{"type":48,"value":532}," — high-evidence issues flagged by only one model. Note which model flagged it.",{"type":42,"tag":51,"props":534,"children":535},{},[536,541],{"type":42,"tag":161,"props":537,"children":538},{},[539],{"type":48,"value":540},"Contested",{"type":48,"value":542}," (if any) — briefly list findings where reviewers disagreed, with a one-line summary of the disagreement. Suggest running a cross-review if the user wants resolution.",{"type":42,"tag":51,"props":544,"children":545},{},[546,551],{"type":42,"tag":161,"props":547,"children":548},{},[549],{"type":48,"value":550},"Summary",{"type":48,"value":552}," — one short paragraph. If no blocking issues survive synthesis, say so explicitly and mention any meaningful testing gaps.",{"type":42,"tag":130,"props":554,"children":556},{"id":555},"after-phase-5-cross-review",[557],{"type":48,"value":558},"After Phase 5 (cross-review)",{"type":42,"tag":51,"props":560,"children":561},{},[562],{"type":48,"value":563},"Same structure as above, but update the buckets based on the deliberation:",{"type":42,"tag":78,"props":565,"children":566},{},[567,572],{"type":42,"tag":82,"props":568,"children":569},{},[570],{"type":48,"value":571},"Contested findings that reached agreement move to Consensus or get dropped.",{"type":42,"tag":82,"props":573,"children":574},{},[575],{"type":48,"value":576},"Note any findings where the split persists after deliberation.",{"items":578,"total":771},[579,599,620,641,656,673,684,697,712,727,746,759],{"slug":580,"name":580,"fn":581,"description":582,"org":583,"tags":584,"stars":596,"repoUrl":597,"updatedAt":598},"rushstack-best-practices","manage Rush monorepos with best practices","Provides best practices and guidance for working with Rush monorepos. Use when the user is working in a Rush-based repository, asks about Rush commands (install, update, build, rebuild), needs help with project selection, dependency management, build caching, subspace configuration, or troubleshooting Rush-specific issues.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[585,586,589,590,593],{"name":17,"slug":18,"type":15},{"name":587,"slug":588,"type":15},"Local Development","local-development",{"name":9,"slug":8,"type":15},{"name":591,"slug":592,"type":15},"Project Management","project-management",{"name":594,"slug":595,"type":15},"Rush","rush",6484,"https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Frushstack","2026-04-06T18:34:44.965032",{"slug":600,"name":600,"fn":601,"description":602,"org":603,"tags":604,"stars":617,"repoUrl":618,"updatedAt":619},"azure-ai-agents-persistent-dotnet","build AI agents with Azure .NET SDK","Azure AI Agents Persistent SDK for .NET. Low-level SDK for creating and managing AI agents with threads, messages, runs, and tools. Use for agent CRUD, conversation threads, streaming responses, function calling, file search, and code interpreter. Triggers: \"PersistentAgentsClient\", \"persistent agents\", \"agent threads\", \"agent runs\", \"streaming agents\", \"function calling agents .NET\".\n",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[605,608,611,614],{"name":606,"slug":607,"type":15},".NET","net",{"name":609,"slug":610,"type":15},"Agents","agents",{"name":612,"slug":613,"type":15},"Azure","azure",{"name":615,"slug":616,"type":15},"LLM","llm",2804,"https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fskills","2026-07-03T16:32:10.297433",{"slug":621,"name":621,"fn":622,"description":623,"org":624,"tags":625,"stars":617,"repoUrl":618,"updatedAt":640},"azure-ai-anomalydetector-java","build anomaly detection applications with Java","Build anomaly detection applications with Azure AI Anomaly Detector SDK for Java. Use when implementing univariate\u002Fmultivariate anomaly detection, time-series analysis, or AI-powered monitoring.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[626,629,630,633,636,637],{"name":627,"slug":628,"type":15},"Analytics","analytics",{"name":612,"slug":613,"type":15},{"name":631,"slug":632,"type":15},"Data Analysis","data-analysis",{"name":634,"slug":635,"type":15},"Java","java",{"name":9,"slug":8,"type":15},{"name":638,"slug":639,"type":15},"Monitoring","monitoring","2026-05-13T06:14:16.261754",{"slug":642,"name":642,"fn":643,"description":644,"org":645,"tags":646,"stars":617,"repoUrl":618,"updatedAt":655},"azure-ai-contentsafety-java","build content moderation applications with Azure AI","Build content moderation applications with Azure AI Content Safety SDK for Java. Use when implementing text\u002Fimage analysis, blocklist management, or harm detection for hate, violence, sexual content, and self-harm.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[647,650,651,652],{"name":648,"slug":649,"type":15},"AI Infrastructure","ai-infrastructure",{"name":612,"slug":613,"type":15},{"name":634,"slug":635,"type":15},{"name":653,"slug":654,"type":15},"Security","security","2026-07-07T06:53:31.293235",{"slug":657,"name":657,"fn":658,"description":659,"org":660,"tags":661,"stars":617,"repoUrl":618,"updatedAt":672},"azure-ai-contentsafety-py","detect harmful content with Azure AI Content Safety","Azure AI Content Safety SDK for Python. Use for detecting harmful content in text and images with multi-severity classification.\nTriggers: \"azure-ai-contentsafety\", \"ContentSafetyClient\", \"content moderation\", \"harmful content\", \"text analysis\", \"image analysis\".\n",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[662,663,666,667,668,671],{"name":612,"slug":613,"type":15},{"name":664,"slug":665,"type":15},"Compliance","compliance",{"name":615,"slug":616,"type":15},{"name":9,"slug":8,"type":15},{"name":669,"slug":670,"type":15},"Python","python",{"name":653,"slug":654,"type":15},"2026-07-18T05:14:23.017504",{"slug":674,"name":674,"fn":675,"description":676,"org":677,"tags":678,"stars":617,"repoUrl":618,"updatedAt":683},"azure-ai-language-conversations-py","implement conversational language understanding with Python","Implement Conversational Language Understanding (CLU) using the azure-ai-language-conversations Python SDK. Use when working with ConversationAnalysisClient to analyze conversation intent and entities, building NLP features, or integrating language understanding into applications.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[679,680,681,682],{"name":627,"slug":628,"type":15},{"name":612,"slug":613,"type":15},{"name":615,"slug":616,"type":15},{"name":669,"slug":670,"type":15},"2026-07-31T05:54:29.068751",{"slug":685,"name":685,"fn":686,"description":687,"org":688,"tags":689,"stars":617,"repoUrl":618,"updatedAt":696},"azure-ai-translation-text-py","translate text using Azure AI services","Azure AI Text Translation SDK for real-time text translation, transliteration, language detection, and dictionary lookup. Use for translating text content in applications.\nTriggers: \"text translation\", \"translator\", \"translate text\", \"transliterate\", \"TextTranslationClient\".\n",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[690,693,694,695],{"name":691,"slug":692,"type":15},"API Development","api-development",{"name":612,"slug":613,"type":15},{"name":9,"slug":8,"type":15},{"name":669,"slug":670,"type":15},"2026-07-18T05:14:16.988376",{"slug":698,"name":698,"fn":699,"description":700,"org":701,"tags":702,"stars":617,"repoUrl":618,"updatedAt":711},"azure-ai-vision-imageanalysis-py","analyze images with Azure AI Vision","Azure AI Vision Image Analysis SDK for captions, tags, objects, OCR, people detection, and smart cropping. Use for computer vision and image understanding tasks.\nTriggers: \"image analysis\", \"computer vision\", \"OCR\", \"object detection\", \"ImageAnalysisClient\", \"image caption\".\n",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[703,704,707,710],{"name":612,"slug":613,"type":15},{"name":705,"slug":706,"type":15},"Computer Vision","computer-vision",{"name":708,"slug":709,"type":15},"Images","images",{"name":669,"slug":670,"type":15},"2026-07-18T05:14:18.007737",{"slug":713,"name":713,"fn":714,"description":715,"org":716,"tags":717,"stars":617,"repoUrl":618,"updatedAt":726},"azure-appconfiguration-java","manage configuration with Azure App Configuration","Azure App Configuration SDK for Java. Centralized application configuration management with key-value settings, feature flags, and snapshots.\nTriggers: \"ConfigurationClient java\", \"app configuration java\", \"feature flag java\", \"configuration setting java\", \"azure config java\".\n",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[718,719,722,725],{"name":612,"slug":613,"type":15},{"name":720,"slug":721,"type":15},"Configuration","configuration",{"name":723,"slug":724,"type":15},"Feature Flags","feature-flags",{"name":634,"slug":635,"type":15},"2026-07-03T16:32:01.278468",{"slug":728,"name":728,"fn":729,"description":730,"org":731,"tags":732,"stars":617,"repoUrl":618,"updatedAt":745},"azure-cosmos-rust","build applications with Azure Cosmos DB","Azure Cosmos DB library for Rust (NoSQL API). Document CRUD, containers, and globally distributed data.\nTriggers: \"cosmos db rust\", \"CosmosClient rust\", \"document crud rust\", \"NoSQL rust\", \"partition key rust\".\n",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[733,736,739,742],{"name":734,"slug":735,"type":15},"Cosmos DB","cosmos-db",{"name":737,"slug":738,"type":15},"Database","database",{"name":740,"slug":741,"type":15},"NoSQL","nosql",{"name":743,"slug":744,"type":15},"Rust","rust","2026-07-31T05:54:27.021432",{"slug":747,"name":747,"fn":729,"description":748,"org":749,"tags":750,"stars":617,"repoUrl":618,"updatedAt":758},"azure-cosmos-ts","Azure Cosmos DB JavaScript\u002FTypeScript SDK (@azure\u002Fcosmos) for data plane operations. Use for CRUD operations on documents, queries, bulk operations, and container management. Triggers: \"Cosmos DB\", \"@azure\u002Fcosmos\", \"CosmosClient\", \"document CRUD\", \"NoSQL queries\", \"bulk operations\", \"partition key\", \"container.items\".\n",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[751,752,753,754,755],{"name":734,"slug":735,"type":15},{"name":737,"slug":738,"type":15},{"name":9,"slug":8,"type":15},{"name":740,"slug":741,"type":15},{"name":756,"slug":757,"type":15},"TypeScript","typescript","2026-07-03T16:31:19.368382",{"slug":760,"name":760,"fn":761,"description":762,"org":763,"tags":764,"stars":617,"repoUrl":618,"updatedAt":770},"azure-data-tables-java","build table storage applications with Java","Build table storage applications with Azure Tables SDK for Java. Use when working with Azure Table Storage or Cosmos DB Table API for NoSQL key-value data, schemaless storage, or structured data at scale.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[765,766,767,768,769],{"name":612,"slug":613,"type":15},{"name":734,"slug":735,"type":15},{"name":737,"slug":738,"type":15},{"name":634,"slug":635,"type":15},{"name":740,"slug":741,"type":15},"2026-05-13T06:14:17.582229",267,{"items":773,"total":851},[774,789,796,808,819,830,839],{"slug":775,"name":775,"fn":776,"description":777,"org":778,"tags":779,"stars":25,"repoUrl":26,"updatedAt":788},"council-plan","perform multi-model council planning","Multi-model council planning for implementation and architecture decisions. Use when the user asks to 'plan with council', 'multi-model plan', 'get different perspectives on approach', 'council plan', or wants multiple models to independently propose plans for the same task. Also use when the user asks to 'debate approaches', 'compare strategies', or 'cross-plan'.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[780,783,784,785],{"name":781,"slug":782,"type":15},"Architecture","architecture",{"name":17,"slug":18,"type":15},{"name":20,"slug":21,"type":15},{"name":786,"slug":787,"type":15},"System Design","system-design","2026-05-13T06:14:08.529779",{"slug":4,"name":4,"fn":5,"description":6,"org":790,"tags":791,"stars":25,"repoUrl":26,"updatedAt":27},{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[792,793,794,795],{"name":13,"slug":14,"type":15},{"name":17,"slug":18,"type":15},{"name":20,"slug":21,"type":15},{"name":23,"slug":24,"type":15},{"slug":797,"name":797,"fn":798,"description":799,"org":800,"tags":801,"stars":25,"repoUrl":26,"updatedAt":807},"follow-goal","manage agent goals and stopping conditions","Give the agent a durable objective with a verifiable stopping condition, then keep iterating across turns until that condition is met. Use when the user says 'set a goal', 'follow a goal', '\u002Fgoal …', 'keep working until …', or asks for a long-running task with a clear end state (migrations, large refactors, retry-until-green loops, experiments).",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[802,803,804],{"name":609,"slug":610,"type":15},{"name":9,"slug":8,"type":15},{"name":805,"slug":806,"type":15},"Productivity","productivity","2026-07-07T06:54:04.263607",{"slug":809,"name":809,"fn":810,"description":811,"org":812,"tags":813,"stars":25,"repoUrl":26,"updatedAt":818},"inbox-add-reaction","add reactions to GitHub issues and PRs","Add a reaction to a GitHub issue or pull request",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[814,817],{"name":815,"slug":816,"type":15},"GitHub","github",{"name":23,"slug":24,"type":15},"2026-07-07T06:53:33.958298",{"slug":820,"name":820,"fn":821,"description":822,"org":823,"tags":824,"stars":25,"repoUrl":26,"updatedAt":829},"inbox-install-gh-cli","install GitHub CLI","Install the GitHub CLI (gh) if not already installed",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[825,828],{"name":826,"slug":827,"type":15},"CLI","cli",{"name":815,"slug":816,"type":15},"2026-05-13T06:14:05.911879",{"slug":831,"name":831,"fn":832,"description":833,"org":834,"tags":835,"stars":25,"repoUrl":26,"updatedAt":838},"inbox-mark-all-read","mark GitHub notifications as read","Mark all GitHub notifications as read",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[836,837],{"name":815,"slug":816,"type":15},{"name":805,"slug":806,"type":15},"2026-05-13T06:14:02.019522",{"slug":840,"name":840,"fn":841,"description":842,"org":843,"tags":844,"stars":25,"repoUrl":26,"updatedAt":850},"inbox-memory","manage agent memory and rules","Read and write agent storage — memory and rules with environment-aware fallback",{"slug":8,"name":9,"logoUrl":10,"githubOrg":8},[845,846,847],{"name":609,"slug":610,"type":15},{"name":720,"slug":721,"type":15},{"name":848,"slug":849,"type":15},"Memory","memory","2026-05-13T06:14:00.740431",14]