[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M69BDCS0CFAC1240TJKR91FY":3,"skill-same:01M69BDCS0CFAC1240TJKR91FY":179},{"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":23,"linkOnly":24,"markdownContent":25,"markdownBody":26,"frontMatter":27,"refName":28,"commitSha":29,"skillDirectory":30,"forksCount":31,"repositoryHtmlUrl":32,"repositoryHomepageUrl":6,"ownerHtmlUrl":33,"ownerBlogUrl":34,"ownerDescription":35,"updatedAt":36,"install":37,"resources":50},"01M69BDCS0CFAC1240TJKR91FY","chaos-study",null,"Guided help for running a Chaos Studio study against an Azure workspace scope: frame a steady-state question with you, discover live which actions Chaos Studio offers for that region, freeze a plan, run a validated scenario configuration only with your explicit consent, collect the before\u002Fduring\u002Fafter signals you chose, and write a self-contained HTML report for your review. Start here for questions like 'what happens to my service when X?' or 'is my service resilient to X?'.","microsoft\u002Fchaos-studio","microsoft","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F6154722?v=4","01M698VYDXZ6JW5WH3CRV650R8","Microsoft","01M69BCX96C2ZS87TPZT4BZGP8",3,"MIT","https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fchaos-studio\u002Fblob\u002Fmain\u002Fcopilot-cli-plugin\u002Fskills\u002Fchaos-study\u002FSKILL.md",[],[19,20,21,22],"chaos study","review","azure","run","2026-11-10 10:12:30",false,"---\nname: chaos-study\ndescription: \"Guided help for running a Chaos Studio study against an Azure workspace scope: frame a steady-state question with you, discover live which actions Chaos Studio offers for that region, freeze a plan, run a validated scenario configuration only with your explicit consent, collect the before\u002Fduring\u002Fafter signals you chose, and write a self-contained HTML report for your review. Start here for questions like 'what happens to my service when X?' or 'is my service resilient to X?'.\"\n---\n\n# chaos-study — guided Chaos studies, end to end\n\nChaos engineering works best as an experiment. This skill helps structure one:\nit starts from a question the run could answer *no* to, runs one action against\nyour resources only with your consent, and reports the signals it collected for\nyou to review.\n\n> Does **`\u003Csteady state>`** hold when I execute **`\u003Caction>`** across\n> **`\u003Cworkspace scope>`**?\n\nYou do not need that sentence to start — the front door builds it with you.\n\"Test resilience\" is not a study: nothing could falsify it.\n\n## Principles\n\n**A study is a falsifiable claim.** Every study needs a steady-state predicate\nmeasurable beforehand that could plausibly break during the run.\n\n**What to test is decided before it is scoped, and never by this skill alone.**\nPicking an action because it is available is how studies end up proving nothing.\nThe front door therefore *always* runs design first: it reads the system, then\nstops and asks you, one question at a time, until the purpose and the risk you\nwill accept are stated in your own words and confirmed. There is no flag that\nskips the interview.\n\n**The platform decides what is possible, not this plugin.** The available\nactions are read live from Chaos Studio for the workspace's region, every time.\nThis suite ships **no** action catalogue, and there is no offline fallback: if\nthe service cannot be asked, scoping stops with exit `16` rather than guessing.\nLikewise the resources a run will touch come from the workspace's discovered\nresources, not from an assertion in the plan.\n\n**Evidence beats vibes.** A verdict is only as good as the signals behind it. If\na signal was not collected, the report says *not measured* — never `0`, never a\nplausible-looking guess.\n\n**Nothing runs without consent.** `-DryRun` defaults to `$true`. Arming a study\ntakes both `-DryRun:$false` and typing back the exact phrase the plan prints.\nThere is no flag that skips this.\n\n**Blast radius is bounded before it is armed.** Workspace, scope, filters,\nexclusions, action, parameters and duration are frozen into the plan and hashed.\nIf the plan changed after you approved it, the run refuses.\n\n## The six skills\n\nThis is a suite, not a monolith. Use the front door — it is the only path that\nenforces the design gate. Calling a phase skill directly is the expert escape\nhatch and skips that enforcement deliberately.\n\n| Skill | Owns |\n| --- | --- |\n| **chaos-study** | chains the phases below; the default entry point |\n| **chaos-study-design** | reading the system, interviewing you, recommending what to test |\n| **chaos-study-scope** | workspace, scope, live action discovery, freezing the plan |\n| **chaos-study-run** | configuration, validation, consent, execution, evidence |\n| **chaos-study-report** | interpretation, findings, the HTML report |\n| **chaos-study-history** | list, compare, and re-run past studies |\n\nEach phase persists its own output, so a chain that fails partway resumes there.\n\n## Usage\n\n**1. Name the system. That is all you start with.** The front door always opens\nthe design phase first: it reads the system, ranks candidate hypotheses against\nreal dependency edges, then holds the study at **exactly one question**:\n\n```powershell\n.\u002Fscripts\u002FInvoke-ChaosStudy.ps1 -System orders-api `\n    -SubscriptionId $sub -ResourceGroup rg-prod -WorkspaceName ws-payments\n```\n\nExit **26** means the study is open and waiting on a person — not that it\nfailed. One question is printed, and repeated in a `CHAOS-QUESTION` JSON block.\n**Put that question to the customer with `ask_user`** (or your host's equivalent\ninteractive mechanism), wait for the reply, then resume the same brief:\n\n```powershell\n.\u002Fscripts\u002FInvoke-ChaosStudy.ps1 -System orders-api -BriefId \u003Cid> `\n    -QuestionId \u003Cid from the block> -Answer '\u003Cwhat they actually said>'\n```\n\nOne question per turn. Do not batch them, do not print a questionnaire, do not\nanswer on the customer's behalf, and do not fall through to scoping. Exit **24**\nmeans the answer was too vague to test and the same question is still open.\nPassing `-Scenario`\u002F`-Action` up front does **not** bypass any of this: nothing\nis scoped until the brief is `CONFIRMED` by the customer.\n\n**2. Ask the platform what it can do here.** The answer depends on the region and\nwhat is in scope, and changes as Chaos Studio ships new actions:\n\n```powershell\n.\u002Fscripts\u002FInvoke-ChaosStudy.ps1 -ListActions `\n    -SubscriptionId $sub -ResourceGroup rg-prod -WorkspaceName ws-payments\n# ...and -ListScenarios for the scenarios the workspace recommends\n```\n\nEach action is printed with its canonical id, type, what it applies to and its\nparameter schema — all as the service reported them.\n\n**3. Plan and preview.** Once the brief is confirmed the same command carries on\ninto scoping and stops there, executing nothing. Read the preview: what is in\nscope after filters and exclusions, the windows, and the configuration that\nwould be created. Overrides such as `-SteadyState` apply on top of the brief.\n**4. Arm it.** Only after the preview matches your intent:\n\n```powershell\n.\u002Fscripts\u002FInvoke-ChaosStudy.ps1 ... -DryRun:$false -Consent '\u003Cphrase from the preview>'\n```\n\n## Reading the verdict\n\n| Verdict | Means |\n| --- | --- |\n| **Steady state held** | The action landed and the objective survived it |\n| **Degraded but recovered** | Breached during the run, recovered afterwards |\n| **Steady state breached** | Breached and still breached after recovery |\n| **Not exercised** | The action never applied enough work to test anything |\n| **Inconclusive** | The action could not be proven to have landed |\n\nSeverity follows recovery, not drama: a breach that never recovers is `critical`;\nthe same breach that self-heals is `medium`.\n\n## Choosing signals\n\nThe suite collects only what you name, because what counts as evidence depends\non the system under study. Two forms:\n\n- `metrics:\u003Cname>` — an Azure Monitor metric on the single scoped resource, or\n  `@\u003CresourceId>` to pin one, `|\u003Caggregation>` to override the aggregation. A\n  scope holding several resources has no implied metric target, so pin it.\n- `logs:\u003CworkspaceId>#\u003Ckql>` — a Log Analytics query\n\nName at least one signal expected to *move* under the action. That is what\nseparates \"resilient\" from \"the action never landed\". Scoping checks readiness\nfirst: a study whose objective cannot be measured, or whose parameters do not\nsatisfy the live schema, teaches nothing, so exit `10` is a real answer.\n\n## Progressive discovery\n\nAction identifiers, parameters and required permissions are **not** documented\nhere, because they are the service's to state and this plugin's to relay. Get\nthem from `-ListActions` — Kubernetes\u002FAKS guidance is deliberately absent until\nthose actions appear there. What lives here is the method:\n\n- **`references\u002Fdesign-method.md`** — how a study is chosen before it is scoped\n- **`references\u002Fstudy-method.md`** — the six-phase method and why it is split\n- **`references\u002Freport-contract.md`** — what the report may and may not claim\n\n## Exit codes\n\n| Code | Meaning |\n| --- | --- |\n| `0` | Success |\n| `1` | Error |\n| `4` | Reserved by the shipped chaos skills (broad fix unconsented) — this suite never emits it |\n| `10` | Readiness gates failed |\n| `11` | Consent declined or phrase mismatch |\n| `12` | Plan changed after it was frozen |\n| `13` | Study or brief already sealed |\n| `14` | Scope unverified — nothing in scope, or the action does not fit it |\n| `15` | Studies not comparable |\n| `16` | Live action discovery unavailable — there is no fallback |\n| `17` | Scenario configuration failed validation |\n| `18` | Paused: an Azure operation must be executed by the host, then resumed |\n| `19` | Paused: role assignments need their own, separate approval |\n| `20` | Scenario would run fewer legs than it declares, and that was not accepted |\n| `21` | The run would not exercise the failure often enough to mean anything |\n| `22` | No adapter can reach Azure — say which one to use |\n| `23` | Study was written by an older contract version |\n| `24` | Design incomplete — an unread area, a vague answer, or a missing field |\n| `25` | A past study cannot be re-run exactly, so no command was printed |\n| `26` | **Paused: waiting on the customer.** One question is open — ask it |\n\n`18`, `19` and `21` are **pauses, not failures**. Each names the one thing it\nneeds; supply it and re-run the same command. Do not fill in the missing part\nyourself — a study that cannot show where its evidence came from cannot be sealed.\n\n## Running where `az` is not usable\n\nEvery Azure call goes through one adapter seam, selected with `-Adapter`:\n`local-az` runs `az` in-process; `external` cannot, so it writes each operation\ninto the study's `operations\u002F` directory and exits `18`. Under `external` you are\nthe executor: read the pending request, run exactly that call with your own\nauthenticated tooling, write the response back as the matching result file, and\nre-run the identical command. Results are bound to the request hash and folded\ninto the study's provenance; resuming is idempotent.\n\n## Notes\n\n- Chaos Studio **V2** only: workspaces are the lifecycle root and resource\n  selection flows through workspace scopes. There is no V1 path or fallback.\n- Core is `az` \u002F Azure Resource Manager: no SRE Agent or MCP dependency.\n- Sealed studies are immutable — re-testing scopes a **new** study.\n- `-Parameters` (scenario), `-ActionParameters` (action), `-AcknowledgeFaultDurationSeconds` and `-AbortCriteria` (the customer's stop rule: `@{ statement; source = 'customer'; signal; condition; resourceCorrelation }`) all cross the front door.\n","# chaos-study — guided Chaos studies, end to end\n\nChaos engineering works best as an experiment. This skill helps structure one:\nit starts from a question the run could answer *no* to, runs one action against\nyour resources only with your consent, and reports the signals it collected for\nyou to review.\n\n> Does **`\u003Csteady state>`** hold when I execute **`\u003Caction>`** across\n> **`\u003Cworkspace scope>`**?\n\nYou do not need that sentence to start — the front door builds it with you.\n\"Test resilience\" is not a study: nothing could falsify it.\n\n## Principles\n\n**A study is a falsifiable claim.** Every study needs a steady-state predicate\nmeasurable beforehand that could plausibly break during the run.\n\n**What to test is decided before it is scoped, and never by this skill alone.**\nPicking an action because it is available is how studies end up proving nothing.\nThe front door therefore *always* runs design first: it reads the system, then\nstops and asks you, one question at a time, until the purpose and the risk you\nwill accept are stated in your own words and confirmed. There is no flag that\nskips the interview.\n\n**The platform decides what is possible, not this plugin.** The available\nactions are read live from Chaos Studio for the workspace's region, every time.\nThis suite ships **no** action catalogue, and there is no offline fallback: if\nthe service cannot be asked, scoping stops with exit `16` rather than guessing.\nLikewise the resources a run will touch come from the workspace's discovered\nresources, not from an assertion in the plan.\n\n**Evidence beats vibes.** A verdict is only as good as the signals behind it. If\na signal was not collected, the report says *not measured* — never `0`, never a\nplausible-looking guess.\n\n**Nothing runs without consent.** `-DryRun` defaults to `$true`. Arming a study\ntakes both `-DryRun:$false` and typing back the exact phrase the plan prints.\nThere is no flag that skips this.\n\n**Blast radius is bounded before it is armed.** Workspace, scope, filters,\nexclusions, action, parameters and duration are frozen into the plan and hashed.\nIf the plan changed after you approved it, the run refuses.\n\n## The six skills\n\nThis is a suite, not a monolith. Use the front door — it is the only path that\nenforces the design gate. Calling a phase skill directly is the expert escape\nhatch and skips that enforcement deliberately.\n\n| Skill | Owns |\n| --- | --- |\n| **chaos-study** | chains the phases below; the default entry point |\n| **chaos-study-design** | reading the system, interviewing you, recommending what to test |\n| **chaos-study-scope** | workspace, scope, live action discovery, freezing the plan |\n| **chaos-study-run** | configuration, validation, consent, execution, evidence |\n| **chaos-study-report** | interpretation, findings, the HTML report |\n| **chaos-study-history** | list, compare, and re-run past studies |\n\nEach phase persists its own output, so a chain that fails partway resumes there.\n\n## Usage\n\n**1. Name the system. That is all you start with.** The front door always opens\nthe design phase first: it reads the system, ranks candidate hypotheses against\nreal dependency edges, then holds the study at **exactly one question**:\n\n```powershell\n.\u002Fscripts\u002FInvoke-ChaosStudy.ps1 -System orders-api `\n    -SubscriptionId $sub -ResourceGroup rg-prod -WorkspaceName ws-payments\n```\n\nExit **26** means the study is open and waiting on a person — not that it\nfailed. One question is printed, and repeated in a `CHAOS-QUESTION` JSON block.\n**Put that question to the customer with `ask_user`** (or your host's equivalent\ninteractive mechanism), wait for the reply, then resume the same brief:\n\n```powershell\n.\u002Fscripts\u002FInvoke-ChaosStudy.ps1 -System orders-api -BriefId \u003Cid> `\n    -QuestionId \u003Cid from the block> -Answer '\u003Cwhat they actually said>'\n```\n\nOne question per turn. Do not batch them, do not print a questionnaire, do not\nanswer on the customer's behalf, and do not fall through to scoping. Exit **24**\nmeans the answer was too vague to test and the same question is still open.\nPassing `-Scenario`\u002F`-Action` up front does **not** bypass any of this: nothing\nis scoped until the brief is `CONFIRMED` by the customer.\n\n**2. Ask the platform what it can do here.** The answer depends on the region and\nwhat is in scope, and changes as Chaos Studio ships new actions:\n\n```powershell\n.\u002Fscripts\u002FInvoke-ChaosStudy.ps1 -ListActions `\n    -SubscriptionId $sub -ResourceGroup rg-prod -WorkspaceName ws-payments\n# ...and -ListScenarios for the scenarios the workspace recommends\n```\n\nEach action is printed with its canonical id, type, what it applies to and its\nparameter schema — all as the service reported them.\n\n**3. Plan and preview.** Once the brief is confirmed the same command carries on\ninto scoping and stops there, executing nothing. Read the preview: what is in\nscope after filters and exclusions, the windows, and the configuration that\nwould be created. Overrides such as `-SteadyState` apply on top of the brief.\n**4. Arm it.** Only after the preview matches your intent:\n\n```powershell\n.\u002Fscripts\u002FInvoke-ChaosStudy.ps1 ... -DryRun:$false -Consent '\u003Cphrase from the preview>'\n```\n\n## Reading the verdict\n\n| Verdict | Means |\n| --- | --- |\n| **Steady state held** | The action landed and the objective survived it |\n| **Degraded but recovered** | Breached during the run, recovered afterwards |\n| **Steady state breached** | Breached and still breached after recovery |\n| **Not exercised** | The action never applied enough work to test anything |\n| **Inconclusive** | The action could not be proven to have landed |\n\nSeverity follows recovery, not drama: a breach that never recovers is `critical`;\nthe same breach that self-heals is `medium`.\n\n## Choosing signals\n\nThe suite collects only what you name, because what counts as evidence depends\non the system under study. Two forms:\n\n- `metrics:\u003Cname>` — an Azure Monitor metric on the single scoped resource, or\n  `@\u003CresourceId>` to pin one, `|\u003Caggregation>` to override the aggregation. A\n  scope holding several resources has no implied metric target, so pin it.\n- `logs:\u003CworkspaceId>#\u003Ckql>` — a Log Analytics query\n\nName at least one signal expected to *move* under the action. That is what\nseparates \"resilient\" from \"the action never landed\". Scoping checks readiness\nfirst: a study whose objective cannot be measured, or whose parameters do not\nsatisfy the live schema, teaches nothing, so exit `10` is a real answer.\n\n## Progressive discovery\n\nAction identifiers, parameters and required permissions are **not** documented\nhere, because they are the service's to state and this plugin's to relay. Get\nthem from `-ListActions` — Kubernetes\u002FAKS guidance is deliberately absent until\nthose actions appear there. What lives here is the method:\n\n- **`references\u002Fdesign-method.md`** — how a study is chosen before it is scoped\n- **`references\u002Fstudy-method.md`** — the six-phase method and why it is split\n- **`references\u002Freport-contract.md`** — what the report may and may not claim\n\n## Exit codes\n\n| Code | Meaning |\n| --- | --- |\n| `0` | Success |\n| `1` | Error |\n| `4` | Reserved by the shipped chaos skills (broad fix unconsented) — this suite never emits it |\n| `10` | Readiness gates failed |\n| `11` | Consent declined or phrase mismatch |\n| `12` | Plan changed after it was frozen |\n| `13` | Study or brief already sealed |\n| `14` | Scope unverified — nothing in scope, or the action does not fit it |\n| `15` | Studies not comparable |\n| `16` | Live action discovery unavailable — there is no fallback |\n| `17` | Scenario configuration failed validation |\n| `18` | Paused: an Azure operation must be executed by the host, then resumed |\n| `19` | Paused: role assignments need their own, separate approval |\n| `20` | Scenario would run fewer legs than it declares, and that was not accepted |\n| `21` | The run would not exercise the failure often enough to mean anything |\n| `22` | No adapter can reach Azure — say which one to use |\n| `23` | Study was written by an older contract version |\n| `24` | Design incomplete — an unread area, a vague answer, or a missing field |\n| `25` | A past study cannot be re-run exactly, so no command was printed |\n| `26` | **Paused: waiting on the customer.** One question is open — ask it |\n\n`18`, `19` and `21` are **pauses, not failures**. Each names the one thing it\nneeds; supply it and re-run the same command. Do not fill in the missing part\nyourself — a study that cannot show where its evidence came from cannot be sealed.\n\n## Running where `az` is not usable\n\nEvery Azure call goes through one adapter seam, selected with `-Adapter`:\n`local-az` runs `az` in-process; `external` cannot, so it writes each operation\ninto the study's `operations\u002F` directory and exits `18`. Under `external` you are\nthe executor: read the pending request, run exactly that call with your own\nauthenticated tooling, write the response back as the matching result file, and\nre-run the identical command. Results are bound to the request hash and folded\ninto the study's provenance; resuming is idempotent.\n\n## Notes\n\n- Chaos Studio **V2** only: workspaces are the lifecycle root and resource\n  selection flows through workspace scopes. There is no V1 path or fallback.\n- Core is `az` \u002F Azure Resource Manager: no SRE Agent or MCP dependency.\n- Sealed studies are immutable — re-testing scopes a **new** study.\n- `-Parameters` (scenario), `-ActionParameters` (action), `-AcknowledgeFaultDurationSeconds` and `-AbortCriteria` (the customer's stop rule: `@{ statement; source = 'customer'; signal; condition; resourceCorrelation }`) all cross the front door.\n",{"name":5,"description":7},"main","6044a40e14cfafb6a7edb297a8a29e98cfe3c543","copilot-cli-plugin\u002Fskills\u002Fchaos-study",4,"https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fchaos-studio","https:\u002F\u002Fgithub.com\u002Fmicrosoft","https:\u002F\u002Fopensource.microsoft.com","Open source projects and samples from Microsoft","2026-10-28 22:44:24",{"source":8,"skillName":5,"commands":38,"prompt":48,"zipAvailable":49},[39,42,45],{"tool":40,"command":41},"npx","npx skills add microsoft\u002Fchaos-studio --skill chaos-study",{"tool":43,"command":44},"bunx","bunx skills add microsoft\u002Fchaos-studio --skill chaos-study",{"tool":46,"command":47},"pnpm","pnpm dlx skills add microsoft\u002Fchaos-studio --skill chaos-study","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fchaos-studio\u002Fblob\u002Fmain\u002Fcopilot-cli-plugin\u002Fskills\u002Fchaos-study\u002FSKILL.md",true,[51,57,61,65,69,73,77,81,85,89,93,97,103,107,111,115,119,123,127,131,135,139,143,147,151,155,159,163,167,171,175],{"path":52,"resourceType":53,"resourceTypeDesc":54,"sizeBytes":55,"sourceHtmlUrl":56},"backends\u002Fatr.md","REFERENCE","参考文档",751,"https:\u002F\u002Fgithub.com\u002Fapache\u002Fmagpie\u002Fblob\u002Fmain\u002Fplugins\u002Fmagpie-release-management\u002Fskills\u002Fkeys-sync\u002Fbackends\u002Fatr.md",{"path":58,"resourceType":53,"resourceTypeDesc":54,"sizeBytes":59,"sourceHtmlUrl":60},"backends\u002Fgit.md",1583,"https:\u002F\u002Fgithub.com\u002Fapache\u002Fmagpie\u002Fblob\u002Fmain\u002Fplugins\u002Fmagpie-release-management\u002Fskills\u002Fkeys-sync\u002Fbackends\u002Fgit.md",{"path":62,"resourceType":53,"resourceTypeDesc":54,"sizeBytes":63,"sourceHtmlUrl":64},"backends\u002Fsvnpubsub.md",875,"https:\u002F\u002Fgithub.com\u002Fapache\u002Fmagpie\u002Fblob\u002Fmain\u002Fplugins\u002Fmagpie-release-management\u002Fskills\u002Fkeys-sync\u002Fbackends\u002Fsvnpubsub.md",{"path":66,"resourceType":53,"resourceTypeDesc":54,"sizeBytes":67,"sourceHtmlUrl":68},"outcomes\u002Fblocked-key.md",1212,"https:\u002F\u002Fgithub.com\u002Fapache\u002Fmagpie\u002Fblob\u002Fmain\u002Fplugins\u002Fmagpie-release-management\u002Fskills\u002Fkeys-sync\u002Foutcomes\u002Fblocked-key.md",{"path":70,"resourceType":53,"resourceTypeDesc":54,"sizeBytes":71,"sourceHtmlUrl":72},"outcomes\u002Fkey-rolled.md",921,"https:\u002F\u002Fgithub.com\u002Fapache\u002Fmagpie\u002Fblob\u002Fmain\u002Fplugins\u002Fmagpie-release-management\u002Fskills\u002Fkeys-sync\u002Foutcomes\u002Fkey-rolled.md",{"path":74,"resourceType":53,"resourceTypeDesc":54,"sizeBytes":75,"sourceHtmlUrl":76},"outcomes\u002Fnoop.md",423,"https:\u002F\u002Fgithub.com\u002Fapache\u002Fmagpie\u002Fblob\u002Fmain\u002Fplugins\u002Fmagpie-release-management\u002Fskills\u002Fkeys-sync\u002Foutcomes\u002Fnoop.md",{"path":78,"resourceType":53,"resourceTypeDesc":54,"sizeBytes":79,"sourceHtmlUrl":80},"prerequisites.md",1191,"https:\u002F\u002Fgithub.com\u002Fapache\u002Fmagpie\u002Fblob\u002Fmain\u002Fplugins\u002Fmagpie-release-management\u002Fskills\u002Fkeys-sync\u002Fprerequisites.md",{"path":82,"resourceType":53,"resourceTypeDesc":54,"sizeBytes":83,"sourceHtmlUrl":84},"references.md",1480,"https:\u002F\u002Fgithub.com\u002Fapache\u002Fmagpie\u002Fblob\u002Fmain\u002Fplugins\u002Fmagpie-release-management\u002Fskills\u002Fkeys-sync\u002Freferences.md",{"path":86,"resourceType":53,"resourceTypeDesc":54,"sizeBytes":87,"sourceHtmlUrl":88},"references\u002Fdesign-method.md",7328,"https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fchaos-studio\u002Fblob\u002Fmain\u002Fcopilot-cli-plugin\u002Fskills\u002Fchaos-study\u002Freferences\u002Fdesign-method.md",{"path":90,"resourceType":53,"resourceTypeDesc":54,"sizeBytes":91,"sourceHtmlUrl":92},"references\u002Freport-contract.md",4743,"https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fchaos-studio\u002Fblob\u002Fmain\u002Fcopilot-cli-plugin\u002Fskills\u002Fchaos-study\u002Freferences\u002Freport-contract.md",{"path":94,"resourceType":53,"resourceTypeDesc":54,"sizeBytes":95,"sourceHtmlUrl":96},"references\u002Fstudy-method.md",9784,"https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fchaos-studio\u002Fblob\u002Fmain\u002Fcopilot-cli-plugin\u002Fskills\u002Fchaos-study\u002Freferences\u002Fstudy-method.md",{"path":98,"resourceType":99,"resourceTypeDesc":100,"sizeBytes":101,"sourceHtmlUrl":102},"scripts\u002FInvoke-ChaosStudy.ps1","SCRIPT","脚本",38488,"https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fchaos-studio\u002Fblob\u002Fmain\u002Fcopilot-cli-plugin\u002Fskills\u002Fchaos-study\u002Fscripts\u002FInvoke-ChaosStudy.ps1",{"path":104,"resourceType":99,"resourceTypeDesc":100,"sizeBytes":105,"sourceHtmlUrl":106},"scripts\u002Flib\u002FAbortCriteria.ps1",15679,"https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fchaos-studio\u002Fblob\u002Fmain\u002Fcopilot-cli-plugin\u002Fskills\u002Fchaos-study\u002Fscripts\u002Flib\u002FAbortCriteria.ps1",{"path":108,"resourceType":99,"resourceTypeDesc":100,"sizeBytes":109,"sourceHtmlUrl":110},"scripts\u002Flib\u002FAdapters.ps1",34921,"https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fchaos-studio\u002Fblob\u002Fmain\u002Fcopilot-cli-plugin\u002Fskills\u002Fchaos-study\u002Fscripts\u002Flib\u002FAdapters.ps1",{"path":112,"resourceType":99,"resourceTypeDesc":100,"sizeBytes":113,"sourceHtmlUrl":114},"scripts\u002Flib\u002FApiVersions.ps1",3514,"https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fchaos-studio\u002Fblob\u002Fmain\u002Fcopilot-cli-plugin\u002Fskills\u002Fchaos-study\u002Fscripts\u002Flib\u002FApiVersions.ps1",{"path":116,"resourceType":99,"resourceTypeDesc":100,"sizeBytes":117,"sourceHtmlUrl":118},"scripts\u002Flib\u002FAzCli.ps1",17923,"https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fchaos-studio\u002Fblob\u002Fmain\u002Fcopilot-cli-plugin\u002Fskills\u002Fchaos-study\u002Fscripts\u002Flib\u002FAzCli.ps1",{"path":120,"resourceType":99,"resourceTypeDesc":100,"sizeBytes":121,"sourceHtmlUrl":122},"scripts\u002Flib\u002FCommon.ps1",29154,"https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fchaos-studio\u002Fblob\u002Fmain\u002Fcopilot-cli-plugin\u002Fskills\u002Fchaos-study\u002Fscripts\u002Flib\u002FCommon.ps1",{"path":124,"resourceType":99,"resourceTypeDesc":100,"sizeBytes":125,"sourceHtmlUrl":126},"scripts\u002Flib\u002FConfigurationPayload.ps1",8768,"https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fchaos-studio\u002Fblob\u002Fmain\u002Fcopilot-cli-plugin\u002Fskills\u002Fchaos-study\u002Fscripts\u002Flib\u002FConfigurationPayload.ps1",{"path":128,"resourceType":99,"resourceTypeDesc":100,"sizeBytes":129,"sourceHtmlUrl":130},"scripts\u002Flib\u002FExecutionPlan.ps1",28464,"https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fchaos-studio\u002Fblob\u002Fmain\u002Fcopilot-cli-plugin\u002Fskills\u002Fchaos-study\u002Fscripts\u002Flib\u002FExecutionPlan.ps1",{"path":132,"resourceType":99,"resourceTypeDesc":100,"sizeBytes":133,"sourceHtmlUrl":134},"scripts\u002Flib\u002FExercise.ps1",14931,"https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fchaos-studio\u002Fblob\u002Fmain\u002Fcopilot-cli-plugin\u002Fskills\u002Fchaos-study\u002Fscripts\u002Flib\u002FExercise.ps1",{"path":136,"resourceType":99,"resourceTypeDesc":100,"sizeBytes":137,"sourceHtmlUrl":138},"scripts\u002Flib\u002FFaultDuration.ps1",11470,"https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fchaos-studio\u002Fblob\u002Fmain\u002Fcopilot-cli-plugin\u002Fskills\u002Fchaos-study\u002Fscripts\u002Flib\u002FFaultDuration.ps1",{"path":140,"resourceType":99,"resourceTypeDesc":100,"sizeBytes":141,"sourceHtmlUrl":142},"scripts\u002Flib\u002FOperation.ps1",22650,"https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fchaos-studio\u002Fblob\u002Fmain\u002Fcopilot-cli-plugin\u002Fskills\u002Fchaos-study\u002Fscripts\u002Flib\u002FOperation.ps1",{"path":144,"resourceType":99,"resourceTypeDesc":100,"sizeBytes":145,"sourceHtmlUrl":146},"scripts\u002Flib\u002FProvenance.ps1",8016,"https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fchaos-studio\u002Fblob\u002Fmain\u002Fcopilot-cli-plugin\u002Fskills\u002Fchaos-study\u002Fscripts\u002Flib\u002FProvenance.ps1",{"path":148,"resourceType":99,"resourceTypeDesc":100,"sizeBytes":149,"sourceHtmlUrl":150},"scripts\u002Flib\u002FRender.ps1",15347,"https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fchaos-studio\u002Fblob\u002Fmain\u002Fcopilot-cli-plugin\u002Fskills\u002Fchaos-study\u002Fscripts\u002Flib\u002FRender.ps1",{"path":152,"resourceType":99,"resourceTypeDesc":100,"sizeBytes":153,"sourceHtmlUrl":154},"scripts\u002Flib\u002FResidue.ps1",39727,"https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fchaos-studio\u002Fblob\u002Fmain\u002Fcopilot-cli-plugin\u002Fskills\u002Fchaos-study\u002Fscripts\u002Flib\u002FResidue.ps1",{"path":156,"resourceType":99,"resourceTypeDesc":100,"sizeBytes":157,"sourceHtmlUrl":158},"scripts\u002Flib\u002FSignalIdentity.ps1",17283,"https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fchaos-studio\u002Fblob\u002Fmain\u002Fcopilot-cli-plugin\u002Fskills\u002Fchaos-study\u002Fscripts\u002Flib\u002FSignalIdentity.ps1",{"path":160,"resourceType":99,"resourceTypeDesc":100,"sizeBytes":161,"sourceHtmlUrl":162},"scripts\u002Flib\u002FStudy.ps1",30871,"https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fchaos-studio\u002Fblob\u002Fmain\u002Fcopilot-cli-plugin\u002Fskills\u002Fchaos-study\u002Fscripts\u002Flib\u002FStudy.ps1",{"path":164,"resourceType":99,"resourceTypeDesc":100,"sizeBytes":165,"sourceHtmlUrl":166},"scripts\u002Flib\u002FVersioning.ps1",6555,"https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fchaos-studio\u002Fblob\u002Fmain\u002Fcopilot-cli-plugin\u002Fskills\u002Fchaos-study\u002Fscripts\u002Flib\u002FVersioning.ps1",{"path":168,"resourceType":53,"resourceTypeDesc":54,"sizeBytes":169,"sourceHtmlUrl":170},"tests\u002Ffixtures\u002Flive-run-list-35f214c.json",104240,"https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fchaos-studio\u002Fblob\u002Fmain\u002Fcopilot-cli-plugin\u002Fskills\u002Fchaos-study\u002Ftests\u002Ffixtures\u002Flive-run-list-35f214c.json",{"path":172,"resourceType":99,"resourceTypeDesc":100,"sizeBytes":173,"sourceHtmlUrl":174},"tests\u002FPollCadence.Tests.ps1",13687,"https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fchaos-studio\u002Fblob\u002Fmain\u002Fcopilot-cli-plugin\u002Fskills\u002Fchaos-study\u002Ftests\u002FPollCadence.Tests.ps1",{"path":176,"resourceType":99,"resourceTypeDesc":100,"sizeBytes":177,"sourceHtmlUrl":178},"tests\u002FRunPhaseFixes.Tests.ps1",19987,"https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fchaos-studio\u002Fblob\u002Fmain\u002Fcopilot-cli-plugin\u002Fskills\u002Fchaos-study\u002Ftests\u002FRunPhaseFixes.Tests.ps1",[180,195,206,217,229,237],{"skillNo":181,"name":182,"nameZh":6,"description":183,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":184,"categories":185,"keywords":191,"publishedAt":23},"01M69BDNKG9SA6RN6T9MA63EP8","start-chaos","Orchestrate the full Chaos Studio v2 workflow: auth → workspace → scenario → run. Shares a single state file and supports resume.","https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fchaos-studio\u002Fblob\u002Fmain\u002Fcopilot-cli-plugin\u002Fskills\u002Fstart-chaos\u002FSKILL.md",[186],{"categoryNo":187,"categoryName":188,"taxonomyType":189,"taxonomyTypeDesc":190},"01JBSEEDCAT000000000000023","Skill与智能体开发","SCENARIO","使用场景",[192,193,194,22],"start chaos","workflow","auth",{"skillNo":196,"name":197,"nameZh":6,"description":198,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":199,"categories":200,"keywords":202,"publishedAt":23},"01M69BDKWYZC4DFX9YC08310S6","setup-scenario","Discover recommended scenarios, build and validate a ScenarioConfiguration, and auto-fix resource permissions.","https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fchaos-studio\u002Fblob\u002Fmain\u002Fcopilot-cli-plugin\u002Fskills\u002Fsetup-scenario\u002FSKILL.md",[201],{"categoryNo":187,"categoryName":188,"taxonomyType":189,"taxonomyTypeDesc":190},[203,204,205],"setup scenario","setup","validate",{"skillNo":207,"name":208,"nameZh":6,"description":209,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":210,"categories":211,"keywords":213,"publishedAt":23},"01M69BDJ2ZYXPETM69X6SKP6YT","run-scenario","Execute a ScenarioConfiguration and stream ScenarioRun status with per-action breakdown until terminal state.","https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fchaos-studio\u002Fblob\u002Fmain\u002Fcopilot-cli-plugin\u002Fskills\u002Frun-scenario\u002FSKILL.md",[212],{"categoryNo":187,"categoryName":188,"taxonomyType":189,"taxonomyTypeDesc":190},[214,22,215,216],"run scenario","terminal","status",{"skillNo":218,"name":219,"nameZh":6,"description":220,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":221,"categories":222,"keywords":224,"publishedAt":23},"01M69BDGAJDDM8HM02YH8HSBDE","create-workspace","Provision a Microsoft.Chaos\u002Fworkspaces resource (v2), bind a managed identity, set scopes, and grant Reader RBAC on the scope.","https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fchaos-studio\u002Fblob\u002Fmain\u002Fcopilot-cli-plugin\u002Fskills\u002Fcreate-workspace\u002FSKILL.md",[223],{"categoryNo":187,"categoryName":188,"taxonomyType":189,"taxonomyTypeDesc":190},[225,226,9,227,228],"create workspace","create","identity","rbac",{"skillNo":230,"name":231,"nameZh":6,"description":232,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":233,"categories":234,"keywords":235,"publishedAt":23},"01M69BDAY9AM9RXZKCW8653V5C","chaos-study-scope","Frame a Chaos study before anything runs: resolve or create a Chaos Studio V2 workspace, read the resources the service discovered inside its scopes, discover live which actions and scenarios are available for that region, express a falsifiable steady-state predicate, verify the action fits what is in scope, then freeze and hash the study plan.","https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fchaos-studio\u002Fblob\u002Fmain\u002Fcopilot-cli-plugin\u002Fskills\u002Fchaos-study-scope\u002FSKILL.md",[],[236,226],"chaos study scope",{"skillNo":238,"name":239,"nameZh":6,"description":240,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":241,"categories":242,"keywords":243,"publishedAt":23},"01M69BD9106S43A040QY767T2B","chaos-study-run","Execute a frozen Chaos study plan against Azure using Chaos Studio V2: verify plan integrity, create and validate a scenario configuration, require an explicit typed consent phrase, collect baseline evidence, execute the scenario run, collect during and post-recovery evidence, and always clean up the configuration. Dry-run by default.","https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fchaos-studio\u002Fblob\u002Fmain\u002Fcopilot-cli-plugin\u002Fskills\u002Fchaos-study-run\u002FSKILL.md",[],[244,22,205,226,21],"chaos study run"]