[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M69BD9106S43A040QY767T2B":3,"skill-same:01M69BD9106S43A040QY767T2B":108},{"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":24,"linkOnly":25,"markdownContent":26,"markdownBody":27,"frontMatter":28,"refName":29,"commitSha":30,"skillDirectory":31,"forksCount":32,"repositoryHtmlUrl":33,"repositoryHomepageUrl":6,"ownerHtmlUrl":34,"ownerBlogUrl":35,"ownerDescription":36,"updatedAt":37,"install":38,"resources":51},"01M69BD9106S43A040QY767T2B","chaos-study-run",null,"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.","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-run\u002FSKILL.md",[],[19,20,21,22,23],"run","chaos study run","validate","azure","create","2026-11-10 10:12:30",false,"---\nname: chaos-study-run\ndescription: \"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.\"\n---\n\n# chaos-study-run — the only phase that changes Azure resources\n\nEverything before this is analysis. This skill executes a Chaos Studio\nscenario against your Azure resources, so it is deliberately the most\nconservative script in the suite.\n\n## Principles\n\n**Dry run is the default.** `-DryRun` defaults to `$true`. A dry run walks the\nentire sequence — plan verification, the configuration it would create, the\nconsent phrase — and executes nothing. Read it before arming anything.\n\n**Consent is typed, not flagged.** Arming requires `-DryRun:$false` **and**\n`-Consent` matching the phrase the plan prints, exactly. A wrong phrase exits\n`11`. There is no `-Yes`, no `-Force` that bypasses it, and no environment\nvariable that pre-approves it.\n\n**Granting access is a second, separate consent.** Injecting a fault is\nreversible and bounded; a role assignment is neither. So `-ApprovePermissions`\ntakes its own phrase, derived from the grants the service actually previewed,\nand one consent can never stand in for the other.\n\n**The plan is verified, not trusted.** The plan hash is recomputed before\nexecution. Any drift since scoping exits `12`. If the blast radius changed after\nyou approved it, this refuses rather than running something you did not review.\n\n**Validation is a gate, not advice.** The scenario configuration is created and\nvalidated *before* any evidence is collected or anything is injected, and the run\nis refused unless the service reports `Succeeded`. If validation fails, the\nmissing role assignments are previewed with `--what-if` and shown — and then the\nstudy stops. Granting them is not covered by the injection consent: a role\nassignment outlives the study and changes who can reach those resources\nafterwards, whereas the injection ends when the run ends. So it takes its own\nphrase, bound to the exact grant set the service previewed. Neither consent\nimplies the other, and a preview that widens invalidates the phrase.\n\n**Evidence is collected around the run, not just during it.** Baseline, during,\nand post windows are all collected. Without baseline there is no \"normal\" to\ncompare to; without post there is no way to tell degradation from damage.\n\n**Missing evidence stays missing.** A collector that fails records `null` and a\nreason. Nothing is interpolated, averaged, or inferred to fill a gap — the report\nwould rather say *not measured* than mislead.\n\n**Cleanup is unconditional, and removal is observed rather than assumed.** The\nscenario configuration is deleted in a `finally` block, so a crash, a Ctrl-C, or\na failed collection still tears it down. A run still in flight is cancelled. The\ndelete returning 0 only means the control plane accepted the request, so each\nremoval is followed by a read-back: only an observed absence is reported as\nremoved, and anything still present stays in the residue ledger with the exact\ncommand to finish the job. Use `-KeepConfiguration` only when you intend to\ninspect it afterwards.\n\n**Sealed studies are immutable.** Re-running a sealed study exits `13`. Re-test\nby scoping a new one, so history stays comparable.\n\n## Usage\n\n**Preview** (this is the default, and executes nothing):\n\n```powershell\n.\u002Fscripts\u002FInvoke-ChaosStudyRun.ps1 -StudyId latest\n```\n\n**Arm it**, using the phrase the preview printed:\n\n```powershell\n.\u002Fscripts\u002FInvoke-ChaosStudyRun.ps1 -StudyId latest -DryRun:$false `\n    -Consent '\u003Cphrase from the preview>'\n```\n\nSwitches worth knowing:\n\n- `-SignalSource` — add evidence sources beyond the plan's (`metrics:` \u002F `logs:`)\n- `-PollSeconds` — scenario run status poll interval\n- `-KeepConfiguration` — leave the configuration for inspection\n\n## What actually happens\n\n1. Load the plan and re-verify its hash\n2. Render the consent phrase and require it back verbatim\n3. Create the scenario configuration on the workspace, carrying the plan's\n   frozen parameters, filters and exclusions\n4. Validate it; on failure, preview the missing grants with `--what-if` and stop\n   with exit 19 unless `-ApprovePermissions '\u003Cphrase>'` matches that preview —\n   only then apply the grants, record them, and validate again\n5. Refuse to continue unless validation now says `Succeeded`\n6. Collect the **baseline** window\n7. Execute the configuration; poll the scenario run until it reaches a terminal\n   state or the injection window closes\n8. Collect the **during** window while the run is live\n9. Wait out recovery, then collect the **post** window\n10. Cancel any in-flight run and delete the configuration (always), then write\n    `run-record.v3.json` and evidence\n\nSteps 3–5 run before step 6 deliberately. A configuration that cannot validate\nis a run that fails within seconds, and finding that out after a baseline window\nhas already elapsed wastes the window and the study. `Start-ChaosStudyScenarioRun`\ntakes the validation result as a required argument and asserts it again as the\nlast thing before `run start`, so the guarantee belongs to the function that\nstarts the run rather than to the order of statements in one caller.\n\nThe configuration is built from what the plan captured live at scope time: the\nscenario, its parameters as the `{key,value}` pairs the service takes, and the\nblast radius as `filters` and `exclusions`. An empty blast-radius member is\nomitted rather than sent — `{\"locations\":[]}` means *no locations*, which would\nsilently turn a real study into a no-op that still reports success.\n\n## Implementation note\n\nExecution uses Chaos Studio **V2** exclusively: workspaces are the lifecycle\nroot, resource selection comes from workspace scopes and the resources the\nservice discovered inside them, and the run is a scenario run against a\nvalidated scenario configuration. There is no V1 path and no fallback to one.\nApi-versions are pinned centrally so a service-side default change cannot\nsilently alter behaviour.\n\n## Exit codes\n\n| Code | Meaning |\n| --- | --- |\n| `0` | Success (including a completed dry run) |\n| `1` | Error |\n| `11` | Consent declined or phrase mismatch |\n| `12` | Plan changed after it was frozen |\n| `13` | Study already sealed |\n| `17` | Scenario configuration failed validation |\n| `19` | Role assignments needed; re-run with `-ApprovePermissions` |\n| `18` | Paused: an Azure operation must be executed by the host, then resumed |\n| `22` | No adapter can reach Azure — pass `-Adapter` explicitly |\n| `23` | Study was written by an older contract version |\n\n`18` is a pause, not a failure, and it is the one exit where the temptation to\nimprovise is most dangerous. Under `-Adapter external` this phase cannot call\nAzure itself, so it wrote the exact request — including the scenario run start or\nstatus poll — into the study's `operations\u002F` directory. Execute that call with\nyour own authenticated tooling, write the response back as the matching result\nfile, and re-run the identical command. The run resumes from where it stopped;\nconsent and the frozen plan hash still hold, so nothing is re-armed and no fault\nis injected twice.\n\nDo not fabricate a run outcome or hand-write evidence to get past this. Results\nare hashed against their request and recorded in the study's provenance, and a\nreport built on unprovenanced evidence cannot be sealed as compliant.\n\n## If it fails partway\n\nThe study state on disk reflects exactly how far it got, and whatever evidence\nwas collected is kept. Because reporting is a pure read of that evidence, a\npartial run can still be reported — the report will simply carry the gaps as\nexplicit limitations rather than hiding them.\n\n## Next\n\n```powershell\nchaos-study-report -StudyId \u003CstudyId>\n```\n\nSee `..\u002Fchaos-study\u002Freferences\u002Fstudy-method.md` for how the windows are chosen.\nPer-action blast radius is not documented locally: it comes from the live action\nmetadata frozen into the plan, which the dry run prints in full.\n","# chaos-study-run — the only phase that changes Azure resources\n\nEverything before this is analysis. This skill executes a Chaos Studio\nscenario against your Azure resources, so it is deliberately the most\nconservative script in the suite.\n\n## Principles\n\n**Dry run is the default.** `-DryRun` defaults to `$true`. A dry run walks the\nentire sequence — plan verification, the configuration it would create, the\nconsent phrase — and executes nothing. Read it before arming anything.\n\n**Consent is typed, not flagged.** Arming requires `-DryRun:$false` **and**\n`-Consent` matching the phrase the plan prints, exactly. A wrong phrase exits\n`11`. There is no `-Yes`, no `-Force` that bypasses it, and no environment\nvariable that pre-approves it.\n\n**Granting access is a second, separate consent.** Injecting a fault is\nreversible and bounded; a role assignment is neither. So `-ApprovePermissions`\ntakes its own phrase, derived from the grants the service actually previewed,\nand one consent can never stand in for the other.\n\n**The plan is verified, not trusted.** The plan hash is recomputed before\nexecution. Any drift since scoping exits `12`. If the blast radius changed after\nyou approved it, this refuses rather than running something you did not review.\n\n**Validation is a gate, not advice.** The scenario configuration is created and\nvalidated *before* any evidence is collected or anything is injected, and the run\nis refused unless the service reports `Succeeded`. If validation fails, the\nmissing role assignments are previewed with `--what-if` and shown — and then the\nstudy stops. Granting them is not covered by the injection consent: a role\nassignment outlives the study and changes who can reach those resources\nafterwards, whereas the injection ends when the run ends. So it takes its own\nphrase, bound to the exact grant set the service previewed. Neither consent\nimplies the other, and a preview that widens invalidates the phrase.\n\n**Evidence is collected around the run, not just during it.** Baseline, during,\nand post windows are all collected. Without baseline there is no \"normal\" to\ncompare to; without post there is no way to tell degradation from damage.\n\n**Missing evidence stays missing.** A collector that fails records `null` and a\nreason. Nothing is interpolated, averaged, or inferred to fill a gap — the report\nwould rather say *not measured* than mislead.\n\n**Cleanup is unconditional, and removal is observed rather than assumed.** The\nscenario configuration is deleted in a `finally` block, so a crash, a Ctrl-C, or\na failed collection still tears it down. A run still in flight is cancelled. The\ndelete returning 0 only means the control plane accepted the request, so each\nremoval is followed by a read-back: only an observed absence is reported as\nremoved, and anything still present stays in the residue ledger with the exact\ncommand to finish the job. Use `-KeepConfiguration` only when you intend to\ninspect it afterwards.\n\n**Sealed studies are immutable.** Re-running a sealed study exits `13`. Re-test\nby scoping a new one, so history stays comparable.\n\n## Usage\n\n**Preview** (this is the default, and executes nothing):\n\n```powershell\n.\u002Fscripts\u002FInvoke-ChaosStudyRun.ps1 -StudyId latest\n```\n\n**Arm it**, using the phrase the preview printed:\n\n```powershell\n.\u002Fscripts\u002FInvoke-ChaosStudyRun.ps1 -StudyId latest -DryRun:$false `\n    -Consent '\u003Cphrase from the preview>'\n```\n\nSwitches worth knowing:\n\n- `-SignalSource` — add evidence sources beyond the plan's (`metrics:` \u002F `logs:`)\n- `-PollSeconds` — scenario run status poll interval\n- `-KeepConfiguration` — leave the configuration for inspection\n\n## What actually happens\n\n1. Load the plan and re-verify its hash\n2. Render the consent phrase and require it back verbatim\n3. Create the scenario configuration on the workspace, carrying the plan's\n   frozen parameters, filters and exclusions\n4. Validate it; on failure, preview the missing grants with `--what-if` and stop\n   with exit 19 unless `-ApprovePermissions '\u003Cphrase>'` matches that preview —\n   only then apply the grants, record them, and validate again\n5. Refuse to continue unless validation now says `Succeeded`\n6. Collect the **baseline** window\n7. Execute the configuration; poll the scenario run until it reaches a terminal\n   state or the injection window closes\n8. Collect the **during** window while the run is live\n9. Wait out recovery, then collect the **post** window\n10. Cancel any in-flight run and delete the configuration (always), then write\n    `run-record.v3.json` and evidence\n\nSteps 3–5 run before step 6 deliberately. A configuration that cannot validate\nis a run that fails within seconds, and finding that out after a baseline window\nhas already elapsed wastes the window and the study. `Start-ChaosStudyScenarioRun`\ntakes the validation result as a required argument and asserts it again as the\nlast thing before `run start`, so the guarantee belongs to the function that\nstarts the run rather than to the order of statements in one caller.\n\nThe configuration is built from what the plan captured live at scope time: the\nscenario, its parameters as the `{key,value}` pairs the service takes, and the\nblast radius as `filters` and `exclusions`. An empty blast-radius member is\nomitted rather than sent — `{\"locations\":[]}` means *no locations*, which would\nsilently turn a real study into a no-op that still reports success.\n\n## Implementation note\n\nExecution uses Chaos Studio **V2** exclusively: workspaces are the lifecycle\nroot, resource selection comes from workspace scopes and the resources the\nservice discovered inside them, and the run is a scenario run against a\nvalidated scenario configuration. There is no V1 path and no fallback to one.\nApi-versions are pinned centrally so a service-side default change cannot\nsilently alter behaviour.\n\n## Exit codes\n\n| Code | Meaning |\n| --- | --- |\n| `0` | Success (including a completed dry run) |\n| `1` | Error |\n| `11` | Consent declined or phrase mismatch |\n| `12` | Plan changed after it was frozen |\n| `13` | Study already sealed |\n| `17` | Scenario configuration failed validation |\n| `19` | Role assignments needed; re-run with `-ApprovePermissions` |\n| `18` | Paused: an Azure operation must be executed by the host, then resumed |\n| `22` | No adapter can reach Azure — pass `-Adapter` explicitly |\n| `23` | Study was written by an older contract version |\n\n`18` is a pause, not a failure, and it is the one exit where the temptation to\nimprovise is most dangerous. Under `-Adapter external` this phase cannot call\nAzure itself, so it wrote the exact request — including the scenario run start or\nstatus poll — into the study's `operations\u002F` directory. Execute that call with\nyour own authenticated tooling, write the response back as the matching result\nfile, and re-run the identical command. The run resumes from where it stopped;\nconsent and the frozen plan hash still hold, so nothing is re-armed and no fault\nis injected twice.\n\nDo not fabricate a run outcome or hand-write evidence to get past this. Results\nare hashed against their request and recorded in the study's provenance, and a\nreport built on unprovenanced evidence cannot be sealed as compliant.\n\n## If it fails partway\n\nThe study state on disk reflects exactly how far it got, and whatever evidence\nwas collected is kept. Because reporting is a pure read of that evidence, a\npartial run can still be reported — the report will simply carry the gaps as\nexplicit limitations rather than hiding them.\n\n## Next\n\n```powershell\nchaos-study-report -StudyId \u003CstudyId>\n```\n\nSee `..\u002Fchaos-study\u002Freferences\u002Fstudy-method.md` for how the windows are chosen.\nPer-action blast radius is not documented locally: it comes from the live action\nmetadata frozen into the plan, which the dry run prints in full.\n",{"name":5,"description":7},"main","6044a40e14cfafb6a7edb297a8a29e98cfe3c543","copilot-cli-plugin\u002Fskills\u002Fchaos-study-run",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":39,"prompt":49,"zipAvailable":50},[40,43,46],{"tool":41,"command":42},"npx","npx skills add microsoft\u002Fchaos-studio --skill chaos-study-run",{"tool":44,"command":45},"bunx","bunx skills add microsoft\u002Fchaos-studio --skill chaos-study-run",{"tool":47,"command":48},"pnpm","pnpm dlx skills add microsoft\u002Fchaos-studio --skill chaos-study-run","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fchaos-studio\u002Fblob\u002Fmain\u002Fcopilot-cli-plugin\u002Fskills\u002Fchaos-study-run\u002FSKILL.md",true,[52,58,62,66,70,74,78,82,86,90,94,100,104],{"path":53,"resourceType":54,"resourceTypeDesc":55,"sizeBytes":56,"sourceHtmlUrl":57},"backends\u002Fatr.md","REFERENCE","参考文档",1518,"https:\u002F\u002Fgithub.com\u002Fapache\u002Fmagpie\u002Fblob\u002Fmain\u002Fplugins\u002Fmagpie-release-management\u002Fskills\u002Farchive-sweep\u002Fbackends\u002Fatr.md",{"path":59,"resourceType":54,"resourceTypeDesc":55,"sizeBytes":60,"sourceHtmlUrl":61},"backends\u002Fgithub-releases.md",656,"https:\u002F\u002Fgithub.com\u002Fapache\u002Fmagpie\u002Fblob\u002Fmain\u002Fplugins\u002Fmagpie-release-management\u002Fskills\u002Farchive-sweep\u002Fbackends\u002Fgithub-releases.md",{"path":63,"resourceType":54,"resourceTypeDesc":55,"sizeBytes":64,"sourceHtmlUrl":65},"backends\u002Fs3.md",627,"https:\u002F\u002Fgithub.com\u002Fapache\u002Fmagpie\u002Fblob\u002Fmain\u002Fplugins\u002Fmagpie-release-management\u002Fskills\u002Farchive-sweep\u002Fbackends\u002Fs3.md",{"path":67,"resourceType":54,"resourceTypeDesc":55,"sizeBytes":68,"sourceHtmlUrl":69},"backends\u002Fself-hosted.md",700,"https:\u002F\u002Fgithub.com\u002Fapache\u002Fmagpie\u002Fblob\u002Fmain\u002Fplugins\u002Fmagpie-release-management\u002Fskills\u002Farchive-sweep\u002Fbackends\u002Fself-hosted.md",{"path":71,"resourceType":54,"resourceTypeDesc":55,"sizeBytes":72,"sourceHtmlUrl":73},"backends\u002Fsvnpubsub.md",901,"https:\u002F\u002Fgithub.com\u002Fapache\u002Fmagpie\u002Fblob\u002Fmain\u002Fplugins\u002Fmagpie-release-management\u002Fskills\u002Farchive-sweep\u002Fbackends\u002Fsvnpubsub.md",{"path":75,"resourceType":54,"resourceTypeDesc":55,"sizeBytes":76,"sourceHtmlUrl":77},"outcomes\u002Forphans.md",810,"https:\u002F\u002Fgithub.com\u002Fapache\u002Fmagpie\u002Fblob\u002Fmain\u002Fplugins\u002Fmagpie-release-management\u002Fskills\u002Farchive-sweep\u002Foutcomes\u002Forphans.md",{"path":79,"resourceType":54,"resourceTypeDesc":55,"sizeBytes":80,"sourceHtmlUrl":81},"outcomes\u002Fretention-rule-error.md",728,"https:\u002F\u002Fgithub.com\u002Fapache\u002Fmagpie\u002Fblob\u002Fmain\u002Fplugins\u002Fmagpie-release-management\u002Fskills\u002Farchive-sweep\u002Foutcomes\u002Fretention-rule-error.md",{"path":83,"resourceType":54,"resourceTypeDesc":55,"sizeBytes":84,"sourceHtmlUrl":85},"outcomes\u002Funmapped.md",547,"https:\u002F\u002Fgithub.com\u002Fapache\u002Fmagpie\u002Fblob\u002Fmain\u002Fplugins\u002Fmagpie-release-management\u002Fskills\u002Farchive-sweep\u002Foutcomes\u002Funmapped.md",{"path":87,"resourceType":54,"resourceTypeDesc":55,"sizeBytes":88,"sourceHtmlUrl":89},"prerequisites.md",908,"https:\u002F\u002Fgithub.com\u002Fapache\u002Fmagpie\u002Fblob\u002Fmain\u002Fplugins\u002Fmagpie-release-management\u002Fskills\u002Farchive-sweep\u002Fprerequisites.md",{"path":91,"resourceType":54,"resourceTypeDesc":55,"sizeBytes":92,"sourceHtmlUrl":93},"references.md",1334,"https:\u002F\u002Fgithub.com\u002Fapache\u002Fmagpie\u002Fblob\u002Fmain\u002Fplugins\u002Fmagpie-release-management\u002Fskills\u002Farchive-sweep\u002Freferences.md",{"path":95,"resourceType":96,"resourceTypeDesc":97,"sizeBytes":98,"sourceHtmlUrl":99},"scripts\u002FInvoke-ChaosStudyRun.ps1","SCRIPT","脚本",51293,"https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fchaos-studio\u002Fblob\u002Fmain\u002Fcopilot-cli-plugin\u002Fskills\u002Fchaos-study-run\u002Fscripts\u002FInvoke-ChaosStudyRun.ps1",{"path":101,"resourceType":96,"resourceTypeDesc":97,"sizeBytes":102,"sourceHtmlUrl":103},"scripts\u002Flib\u002FExecute.ps1",97935,"https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fchaos-studio\u002Fblob\u002Fmain\u002Fcopilot-cli-plugin\u002Fskills\u002Fchaos-study-run\u002Fscripts\u002Flib\u002FExecute.ps1",{"path":105,"resourceType":96,"resourceTypeDesc":97,"sizeBytes":106,"sourceHtmlUrl":107},"scripts\u002Flib\u002FSignals.ps1",15880,"https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fchaos-studio\u002Fblob\u002Fmain\u002Fcopilot-cli-plugin\u002Fskills\u002Fchaos-study-run\u002Fscripts\u002Flib\u002FSignals.ps1",[109,124,134,145,156,165],{"skillNo":110,"name":111,"nameZh":6,"description":112,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":113,"categories":114,"keywords":120,"publishedAt":24},"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",[115],{"categoryNo":116,"categoryName":117,"taxonomyType":118,"taxonomyTypeDesc":119},"01JBSEEDCAT000000000000023","Skill与智能体开发","SCENARIO","使用场景",[121,122,123,19],"start chaos","workflow","auth",{"skillNo":125,"name":126,"nameZh":6,"description":127,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":128,"categories":129,"keywords":131,"publishedAt":24},"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",[130],{"categoryNo":116,"categoryName":117,"taxonomyType":118,"taxonomyTypeDesc":119},[132,133,21],"setup scenario","setup",{"skillNo":135,"name":136,"nameZh":6,"description":137,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":138,"categories":139,"keywords":141,"publishedAt":24},"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",[140],{"categoryNo":116,"categoryName":117,"taxonomyType":118,"taxonomyTypeDesc":119},[142,19,143,144],"run scenario","terminal","status",{"skillNo":146,"name":147,"nameZh":6,"description":148,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":149,"categories":150,"keywords":152,"publishedAt":24},"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",[151],{"categoryNo":116,"categoryName":117,"taxonomyType":118,"taxonomyTypeDesc":119},[153,23,154,155,9],"create workspace","rbac","identity",{"skillNo":157,"name":158,"nameZh":6,"description":159,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":160,"categories":161,"keywords":162,"publishedAt":24},"01M69BDCS0CFAC1240TJKR91FY","chaos-study","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?'.","https:\u002F\u002Fgithub.com\u002Fmicrosoft\u002Fchaos-studio\u002Fblob\u002Fmain\u002Fcopilot-cli-plugin\u002Fskills\u002Fchaos-study\u002FSKILL.md",[],[163,19,22,164],"chaos study","review",{"skillNo":166,"name":167,"nameZh":6,"description":168,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":169,"categories":170,"keywords":171,"publishedAt":24},"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",[],[172,23],"chaos study scope"]