[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill-azure-kars-spawn":3,"mdc--rcn699-key":54,"related-repo-azure-kars-spawn":485,"related-org-azure-kars-spawn":599},{"slug":4,"name":4,"fn":5,"description":6,"org":7,"tags":12,"stars":26,"repoUrl":27,"updatedAt":28,"license":29,"forks":30,"topics":31,"repo":49,"sourceUrl":52,"mdContent":53},"kars-spawn","spawn and manage isolated sub-agent sandboxes","Spawn secure isolated sub-agent sandboxes, delegate tasks via AGT mesh, receive results, and destroy sub-agents. Uses the kars_spawn, kars_mesh_send, kars_mesh_inbox, and kars_spawn_destroy tools.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},"azure","Azure (Microsoft)","https:\u002F\u002Fpexgzepcugksgbtrxkhf.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Forg-logos\u002Fazure.png","Azure",[13,17,20,23],{"name":14,"slug":15,"type":16},"Automation","automation","tag",{"name":18,"slug":19,"type":16},"Sandboxing","sandboxing",{"name":21,"slug":22,"type":16},"Agents","agents",{"name":24,"slug":25,"type":16},"Multi-Agent","multi-agent",28,"https:\u002F\u002Fgithub.com\u002FAzure\u002Fkars","2026-07-12T08:18:29.1475",null,2,[32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48],"agent","agentic-framework","agentmesh","ai-agent","ai-agents-automation","ai-sandbox","end-to-end-encryption","hermes","hermes-agent","kubernetes","kubernetes-sandbox","microsoft","openclaw","reference","runtime-security","sandbox","zero-trust",{"repoUrl":27,"stars":26,"forks":30,"topics":50,"description":51},[32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48],"Agent Reference Stack for Kubernetes (kars) - an open source stack from Microsoft for running AI agents safely on Kubernetes. Multi-runtime, Foundry-aware, hardened per-agent sandboxes, governed egress, end-to-end encrypted inter-agent mesh.","https:\u002F\u002Fgithub.com\u002FAzure\u002Fkars\u002Ftree\u002FHEAD\u002Fruntimes\u002Fopenclaw\u002Fskills\u002Fkars-spawn","---\nname: kars-spawn\ndescription: Spawn secure isolated sub-agent sandboxes, delegate tasks via AGT mesh, receive results, and destroy sub-agents. Uses the kars_spawn, kars_mesh_send, kars_mesh_inbox, and kars_spawn_destroy tools.\nmetadata: {\"openclaw\": {\"always\": true}}\n---\n\n# kars Sub-Agent Spawn\n\nSpawn secure isolated sub-agent sandboxes on AKS. Each sub-agent runs in its own K8s namespace with full security controls (NetworkPolicy, seccomp, read-only rootfs). Use the kars tools for ALL operations — do NOT use curl.\n\n## Available tools\n\n- **kars_spawn** — create a new sub-agent ({name, model, governance})\n- **kars_spawn_status** — check sub-agent phase ({name})\n- **kars_spawn_list** — list all spawned sub-agents\n- **kars_mesh_send** — send a task via AGT mesh ({to_agent, content})\n- **kars_mesh_inbox** — check inbox for sub-agent responses\n- **kars_spawn_destroy** — tear down a sub-agent ({name})\n\n## When to use\n\n- **Delegation**: spawn a specialist sub-agent for a focused task (security audit, code review, data analysis)\n- **Parallel work**: run multiple sub-agents on different tasks simultaneously\n- **Isolation**: give untrusted code its own sandboxed environment\n\n## Workflow\n\n1. **kars_spawn** with `name`, `model` (default: parent's), `governance: true`, and **`role`** (the persona — e.g. `\"data analyst\"`, `\"visualization engineer\"`, `\"technical writer\"`). The platform tracks the role in a Peer roster and prepends it to every mesh_send so siblings can resolve role references to canonical names.\n2. **kars_spawn_status** — poll every 5s until phase is \"Running\"\n3. **kars_mesh_send** — send the task to the sub-agent\n4. Wait 30-60 seconds for the sub-agent to process and auto-reply\n5. **kars_mesh_inbox** — read the sub-agent's response\n6. **kars_spawn_destroy** — clean up when done\n\n### Multi-agent peer roster (CRITICAL for pipelines)\n\nWhen you spawn multiple sub-agents that will hand work to each other (e.g. `analyst → viz → writer`), always pass a `role` to every spawn. The platform automatically prepends a `Peer roster:` block to every mesh_send:\n\n```\nPeer roster (use these EXACT agent names with mesh_send \u002F mesh_transfer_file ...):\n  - analyst — data analyst\n  - viz — visualization engineer\n  - writer — technical writer\n```\n\nThis eliminates LLM guessing when sub-agents reference each other by role (\"hand to the writer\"). Without `role`, the roster lists names only and your sub-agents may misroute artifacts.\n\n## File sharing between agents — CRITICAL\n\nSub-agents are **completely isolated containers** with separate filesystems.\nThey CANNOT read each other's files or yours. The AGT mesh is the ONLY channel.\n\n**To share files between agents, you MUST pass file contents inside mesh messages.**\n\n### Multi-agent collaboration pattern (e.g. Red\u002FBlue\u002FJudge):\n\n```\n1. Spawn blue-team, red-team, judge\n2. mesh_send to blue-team: \"Write a proposal. Return the FULL TEXT in your reply.\"\n3. Read blue-team's reply → extract the proposal text\n4. mesh_send to red-team: \"Here is the proposal:\\n\\n\u003Cpaste full text>\\n\\nFind 3 flaws.\"\n5. Read red-team's reply → extract the critique text\n6. mesh_send to blue-team: \"Here is the critique:\\n\\n\u003Cpaste full text>\\n\\nRevise your proposal.\"\n7. Read blue-team's reply → extract revised proposal\n8. mesh_send to judge: \"Proposal:\\n\\n\u003Cpaste>\\n\\nCritique:\\n\\n\u003Cpaste>\\n\\nRevisions:\\n\\n\u003Cpaste>\\n\\nScore it.\"\n9. Read judge's verdict\n10. Destroy all sub-agents\n```\n\n### Key rules:\n- **Always ask sub-agents to return file contents in their reply** — not just \"save to file\"\n- **You (the parent) are the relay** — extract text from each reply and forward to the next agent\n- **Never tell a sub-agent to read a file another agent wrote** — the file doesn't exist in its container\n- Sub-agents CAN write files locally for their own use, but other agents can't see them\n\n## Important notes\n\n- Each sub-agent runs in its own isolated namespace — it cannot see your files or other agents' files\n- AGT mesh is the only communication channel (E2E encrypted, trust-gated, audited)\n- Each sub-agent has its own token budget, blocklist, and Content Safety\n- Names must be DNS-safe: lowercase alphanumeric + hyphens, 1-63 chars\n- Only you (the parent) can destroy sub-agents you spawned\n",{"data":55,"body":59},{"name":4,"description":6,"metadata":56},{"openclaw":57},{"always":58},true,{"type":60,"children":61},"root",[62,71,77,84,150,156,189,195,307,314,342,354,366,372,384,392,398,407,413,451,457],{"type":63,"tag":64,"props":65,"children":67},"element","h1",{"id":66},"kars-sub-agent-spawn",[68],{"type":69,"value":70},"text","kars Sub-Agent Spawn",{"type":63,"tag":72,"props":73,"children":74},"p",{},[75],{"type":69,"value":76},"Spawn secure isolated sub-agent sandboxes on AKS. Each sub-agent runs in its own K8s namespace with full security controls (NetworkPolicy, seccomp, read-only rootfs). Use the kars tools for ALL operations — do NOT use curl.",{"type":63,"tag":78,"props":79,"children":81},"h2",{"id":80},"available-tools",[82],{"type":69,"value":83},"Available tools",{"type":63,"tag":85,"props":86,"children":87},"ul",{},[88,100,110,120,130,140],{"type":63,"tag":89,"props":90,"children":91},"li",{},[92,98],{"type":63,"tag":93,"props":94,"children":95},"strong",{},[96],{"type":69,"value":97},"kars_spawn",{"type":69,"value":99}," — create a new sub-agent ({name, model, governance})",{"type":63,"tag":89,"props":101,"children":102},{},[103,108],{"type":63,"tag":93,"props":104,"children":105},{},[106],{"type":69,"value":107},"kars_spawn_status",{"type":69,"value":109}," — check sub-agent phase ({name})",{"type":63,"tag":89,"props":111,"children":112},{},[113,118],{"type":63,"tag":93,"props":114,"children":115},{},[116],{"type":69,"value":117},"kars_spawn_list",{"type":69,"value":119}," — list all spawned sub-agents",{"type":63,"tag":89,"props":121,"children":122},{},[123,128],{"type":63,"tag":93,"props":124,"children":125},{},[126],{"type":69,"value":127},"kars_mesh_send",{"type":69,"value":129}," — send a task via AGT mesh ({to_agent, content})",{"type":63,"tag":89,"props":131,"children":132},{},[133,138],{"type":63,"tag":93,"props":134,"children":135},{},[136],{"type":69,"value":137},"kars_mesh_inbox",{"type":69,"value":139}," — check inbox for sub-agent responses",{"type":63,"tag":89,"props":141,"children":142},{},[143,148],{"type":63,"tag":93,"props":144,"children":145},{},[146],{"type":69,"value":147},"kars_spawn_destroy",{"type":69,"value":149}," — tear down a sub-agent ({name})",{"type":63,"tag":78,"props":151,"children":153},{"id":152},"when-to-use",[154],{"type":69,"value":155},"When to use",{"type":63,"tag":85,"props":157,"children":158},{},[159,169,179],{"type":63,"tag":89,"props":160,"children":161},{},[162,167],{"type":63,"tag":93,"props":163,"children":164},{},[165],{"type":69,"value":166},"Delegation",{"type":69,"value":168},": spawn a specialist sub-agent for a focused task (security audit, code review, data analysis)",{"type":63,"tag":89,"props":170,"children":171},{},[172,177],{"type":63,"tag":93,"props":173,"children":174},{},[175],{"type":69,"value":176},"Parallel work",{"type":69,"value":178},": run multiple sub-agents on different tasks simultaneously",{"type":63,"tag":89,"props":180,"children":181},{},[182,187],{"type":63,"tag":93,"props":183,"children":184},{},[185],{"type":69,"value":186},"Isolation",{"type":69,"value":188},": give untrusted code its own sandboxed environment",{"type":63,"tag":78,"props":190,"children":192},{"id":191},"workflow",[193],{"type":69,"value":194},"Workflow",{"type":63,"tag":196,"props":197,"children":198},"ol",{},[199,266,275,284,289,298],{"type":63,"tag":89,"props":200,"children":201},{},[202,206,208,215,217,223,225,231,233,242,244,250,251,257,258,264],{"type":63,"tag":93,"props":203,"children":204},{},[205],{"type":69,"value":97},{"type":69,"value":207}," with ",{"type":63,"tag":209,"props":210,"children":212},"code",{"className":211},[],[213],{"type":69,"value":214},"name",{"type":69,"value":216},", ",{"type":63,"tag":209,"props":218,"children":220},{"className":219},[],[221],{"type":69,"value":222},"model",{"type":69,"value":224}," (default: parent's), ",{"type":63,"tag":209,"props":226,"children":228},{"className":227},[],[229],{"type":69,"value":230},"governance: true",{"type":69,"value":232},", and ",{"type":63,"tag":93,"props":234,"children":235},{},[236],{"type":63,"tag":209,"props":237,"children":239},{"className":238},[],[240],{"type":69,"value":241},"role",{"type":69,"value":243}," (the persona — e.g. ",{"type":63,"tag":209,"props":245,"children":247},{"className":246},[],[248],{"type":69,"value":249},"\"data analyst\"",{"type":69,"value":216},{"type":63,"tag":209,"props":252,"children":254},{"className":253},[],[255],{"type":69,"value":256},"\"visualization engineer\"",{"type":69,"value":216},{"type":63,"tag":209,"props":259,"children":261},{"className":260},[],[262],{"type":69,"value":263},"\"technical writer\"",{"type":69,"value":265},"). The platform tracks the role in a Peer roster and prepends it to every mesh_send so siblings can resolve role references to canonical names.",{"type":63,"tag":89,"props":267,"children":268},{},[269,273],{"type":63,"tag":93,"props":270,"children":271},{},[272],{"type":69,"value":107},{"type":69,"value":274}," — poll every 5s until phase is \"Running\"",{"type":63,"tag":89,"props":276,"children":277},{},[278,282],{"type":63,"tag":93,"props":279,"children":280},{},[281],{"type":69,"value":127},{"type":69,"value":283}," — send the task to the sub-agent",{"type":63,"tag":89,"props":285,"children":286},{},[287],{"type":69,"value":288},"Wait 30-60 seconds for the sub-agent to process and auto-reply",{"type":63,"tag":89,"props":290,"children":291},{},[292,296],{"type":63,"tag":93,"props":293,"children":294},{},[295],{"type":69,"value":137},{"type":69,"value":297}," — read the sub-agent's response",{"type":63,"tag":89,"props":299,"children":300},{},[301,305],{"type":63,"tag":93,"props":302,"children":303},{},[304],{"type":69,"value":147},{"type":69,"value":306}," — clean up when done",{"type":63,"tag":308,"props":309,"children":311},"h3",{"id":310},"multi-agent-peer-roster-critical-for-pipelines",[312],{"type":69,"value":313},"Multi-agent peer roster (CRITICAL for pipelines)",{"type":63,"tag":72,"props":315,"children":316},{},[317,319,325,327,332,334,340],{"type":69,"value":318},"When you spawn multiple sub-agents that will hand work to each other (e.g. ",{"type":63,"tag":209,"props":320,"children":322},{"className":321},[],[323],{"type":69,"value":324},"analyst → viz → writer",{"type":69,"value":326},"), always pass a ",{"type":63,"tag":209,"props":328,"children":330},{"className":329},[],[331],{"type":69,"value":241},{"type":69,"value":333}," to every spawn. The platform automatically prepends a ",{"type":63,"tag":209,"props":335,"children":337},{"className":336},[],[338],{"type":69,"value":339},"Peer roster:",{"type":69,"value":341}," block to every mesh_send:",{"type":63,"tag":343,"props":344,"children":348},"pre",{"className":345,"code":347,"language":69},[346],"language-text","Peer roster (use these EXACT agent names with mesh_send \u002F mesh_transfer_file ...):\n  - analyst — data analyst\n  - viz — visualization engineer\n  - writer — technical writer\n",[349],{"type":63,"tag":209,"props":350,"children":352},{"__ignoreMap":351},"",[353],{"type":69,"value":347},{"type":63,"tag":72,"props":355,"children":356},{},[357,359,364],{"type":69,"value":358},"This eliminates LLM guessing when sub-agents reference each other by role (\"hand to the writer\"). Without ",{"type":63,"tag":209,"props":360,"children":362},{"className":361},[],[363],{"type":69,"value":241},{"type":69,"value":365},", the roster lists names only and your sub-agents may misroute artifacts.",{"type":63,"tag":78,"props":367,"children":369},{"id":368},"file-sharing-between-agents-critical",[370],{"type":69,"value":371},"File sharing between agents — CRITICAL",{"type":63,"tag":72,"props":373,"children":374},{},[375,377,382],{"type":69,"value":376},"Sub-agents are ",{"type":63,"tag":93,"props":378,"children":379},{},[380],{"type":69,"value":381},"completely isolated containers",{"type":69,"value":383}," with separate filesystems.\nThey CANNOT read each other's files or yours. The AGT mesh is the ONLY channel.",{"type":63,"tag":72,"props":385,"children":386},{},[387],{"type":63,"tag":93,"props":388,"children":389},{},[390],{"type":69,"value":391},"To share files between agents, you MUST pass file contents inside mesh messages.",{"type":63,"tag":308,"props":393,"children":395},{"id":394},"multi-agent-collaboration-pattern-eg-redbluejudge",[396],{"type":69,"value":397},"Multi-agent collaboration pattern (e.g. Red\u002FBlue\u002FJudge):",{"type":63,"tag":343,"props":399,"children":402},{"className":400,"code":401,"language":69},[346],"1. Spawn blue-team, red-team, judge\n2. mesh_send to blue-team: \"Write a proposal. Return the FULL TEXT in your reply.\"\n3. Read blue-team's reply → extract the proposal text\n4. mesh_send to red-team: \"Here is the proposal:\\n\\n\u003Cpaste full text>\\n\\nFind 3 flaws.\"\n5. Read red-team's reply → extract the critique text\n6. mesh_send to blue-team: \"Here is the critique:\\n\\n\u003Cpaste full text>\\n\\nRevise your proposal.\"\n7. Read blue-team's reply → extract revised proposal\n8. mesh_send to judge: \"Proposal:\\n\\n\u003Cpaste>\\n\\nCritique:\\n\\n\u003Cpaste>\\n\\nRevisions:\\n\\n\u003Cpaste>\\n\\nScore it.\"\n9. Read judge's verdict\n10. Destroy all sub-agents\n",[403],{"type":63,"tag":209,"props":404,"children":405},{"__ignoreMap":351},[406],{"type":69,"value":401},{"type":63,"tag":308,"props":408,"children":410},{"id":409},"key-rules",[411],{"type":69,"value":412},"Key rules:",{"type":63,"tag":85,"props":414,"children":415},{},[416,426,436,446],{"type":63,"tag":89,"props":417,"children":418},{},[419,424],{"type":63,"tag":93,"props":420,"children":421},{},[422],{"type":69,"value":423},"Always ask sub-agents to return file contents in their reply",{"type":69,"value":425}," — not just \"save to file\"",{"type":63,"tag":89,"props":427,"children":428},{},[429,434],{"type":63,"tag":93,"props":430,"children":431},{},[432],{"type":69,"value":433},"You (the parent) are the relay",{"type":69,"value":435}," — extract text from each reply and forward to the next agent",{"type":63,"tag":89,"props":437,"children":438},{},[439,444],{"type":63,"tag":93,"props":440,"children":441},{},[442],{"type":69,"value":443},"Never tell a sub-agent to read a file another agent wrote",{"type":69,"value":445}," — the file doesn't exist in its container",{"type":63,"tag":89,"props":447,"children":448},{},[449],{"type":69,"value":450},"Sub-agents CAN write files locally for their own use, but other agents can't see them",{"type":63,"tag":78,"props":452,"children":454},{"id":453},"important-notes",[455],{"type":69,"value":456},"Important notes",{"type":63,"tag":85,"props":458,"children":459},{},[460,465,470,475,480],{"type":63,"tag":89,"props":461,"children":462},{},[463],{"type":69,"value":464},"Each sub-agent runs in its own isolated namespace — it cannot see your files or other agents' files",{"type":63,"tag":89,"props":466,"children":467},{},[468],{"type":69,"value":469},"AGT mesh is the only communication channel (E2E encrypted, trust-gated, audited)",{"type":63,"tag":89,"props":471,"children":472},{},[473],{"type":69,"value":474},"Each sub-agent has its own token budget, blocklist, and Content Safety",{"type":63,"tag":89,"props":476,"children":477},{},[478],{"type":69,"value":479},"Names must be DNS-safe: lowercase alphanumeric + hyphens, 1-63 chars",{"type":63,"tag":89,"props":481,"children":482},{},[483],{"type":69,"value":484},"Only you (the parent) can destroy sub-agents you spawned",{"items":486,"total":598},[487,505,517,540,552,566,581],{"slug":488,"name":488,"fn":489,"description":490,"org":491,"tags":492,"stars":26,"repoUrl":27,"updatedAt":504},"agt-governance","enforce behavioral governance for agents","Behavioral governance for OpenClaw agents via AGT — tool-level policy, inter-agent trust, audit logging.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[493,494,497,498,501],{"name":21,"slug":22,"type":16},{"name":495,"slug":496,"type":16},"Audit","audit",{"name":11,"slug":8,"type":16},{"name":499,"slug":500,"type":16},"Governance","governance",{"name":502,"slug":503,"type":16},"Policy","policy","2026-07-12T08:18:34.446985",{"slug":506,"name":506,"fn":507,"description":508,"org":509,"tags":510,"stars":26,"repoUrl":27,"updatedAt":516},"foundry-agents","inspect and invoke Foundry prompt agents","Query and inspect Foundry prompt agents and invoke Foundry tools via the Responses API. OpenClaw is the orchestrator — Foundry provides managed AI services.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[511,512,515],{"name":21,"slug":22,"type":16},{"name":513,"slug":514,"type":16},"API Development","api-development",{"name":11,"slug":8,"type":16},"2026-07-12T08:18:31.798351",{"slug":518,"name":518,"fn":519,"description":520,"org":521,"tags":522,"stars":26,"repoUrl":27,"updatedAt":539},"foundry-code","execute Python code in Azure Foundry","Python code execution via Azure AI Foundry Responses API with code_interpreter tool. Data analysis, charts, and math in a managed sandbox.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[523,524,527,530,533,536],{"name":11,"slug":8,"type":16},{"name":525,"slug":526,"type":16},"Charts","charts",{"name":528,"slug":529,"type":16},"Data Analysis","data-analysis",{"name":531,"slug":532,"type":16},"Data Visualization","data-visualization",{"name":534,"slug":535,"type":16},"Mathematics","mathematics",{"name":537,"slug":538,"type":16},"Python","python","2026-07-12T08:18:30.4206",{"slug":541,"name":541,"fn":542,"description":543,"org":544,"tags":545,"stars":26,"repoUrl":27,"updatedAt":551},"foundry-conversations","manage persistent conversations via Foundry","Manage persistent conversations via Foundry Conversations API. Create conversations, add messages, and maintain history across sessions.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[546,547,548],{"name":21,"slug":22,"type":16},{"name":11,"slug":8,"type":16},{"name":549,"slug":550,"type":16},"Memory","memory","2026-07-12T08:18:38.407887",{"slug":553,"name":553,"fn":554,"description":555,"org":556,"tags":557,"stars":26,"repoUrl":27,"updatedAt":565},"foundry-deployments","query Foundry model deployments and infrastructure","Query model deployments, connections, and indexes in the Foundry project. Discover available models and infrastructure.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[558,559,562],{"name":11,"slug":8,"type":16},{"name":560,"slug":561,"type":16},"Infrastructure","infrastructure",{"name":563,"slug":564,"type":16},"LLM","llm","2026-07-12T08:18:39.662409",{"slug":567,"name":567,"fn":568,"description":569,"org":570,"tags":571,"stars":26,"repoUrl":27,"updatedAt":580},"foundry-evaluations","evaluate agent quality with Foundry Evals","Evaluate agent quality using Foundry OpenAI Evals API. Create evaluations, run them against models, and analyze results.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[572,573,576,577],{"name":11,"slug":8,"type":16},{"name":574,"slug":575,"type":16},"Evals","evals",{"name":563,"slug":564,"type":16},{"name":578,"slug":579,"type":16},"OpenAI","openai","2026-07-12T08:18:40.889909",{"slug":582,"name":582,"fn":583,"description":584,"org":585,"tags":586,"stars":26,"repoUrl":27,"updatedAt":597},"foundry-knowledge","retrieve knowledge via Foundry RAG","Knowledge retrieval (RAG) via Foundry file_search and azure_ai_search tools. Agentic retrieval with citations — uses Responses API.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[587,588,591,594],{"name":11,"slug":8,"type":16},{"name":589,"slug":590,"type":16},"Knowledge Management","knowledge-management",{"name":592,"slug":593,"type":16},"Research","research",{"name":595,"slug":596,"type":16},"Search","search","2026-07-12T08:18:35.894301",11,{"items":600,"total":771},[601,616,635,652,667,682,695,710,721,733,746,759],{"slug":602,"name":602,"fn":603,"description":604,"org":605,"tags":606,"stars":613,"repoUrl":614,"updatedAt":615},"azure-arg-external-evaluation-policy-author","author and test Azure Resource Graph policies","Use when the user wants to author, design, or test an Azure Policy that queries Azure Resource Graph (ARG) at request-time — i.e. a policy whose deny\u002Faudit decision depends on data from elsewhere in the subscription (sibling\u002Fparent resource state, RG-wide invariants, multi-hop relationships, etc.). Formally called Azure Policy External Evaluation; sometimes referred to colloquially as \"Invoke\". Drives an iterative KQL co-design loop against the user's real subscription via `az graph query`, then emits a policy definition, assignment, `.http` test flow, and an `EXPLANATION.md` companion. Read-only; never provisions anything.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[607,608,611,612],{"name":11,"slug":8,"type":16},{"name":609,"slug":610,"type":16},"Compliance","compliance",{"name":499,"slug":500,"type":16},{"name":502,"slug":503,"type":16},1686,"https:\u002F\u002Fgithub.com\u002FAzure\u002Fazure-policy","2026-07-12T08:17:48.378432",{"slug":617,"name":617,"fn":618,"description":619,"org":620,"tags":621,"stars":632,"repoUrl":633,"updatedAt":634},"azure-blueprints-migration","migrate Azure Blueprints to Template Specs","Use when a user needs to migrate off Azure Blueprints (definitions and\u002For assignments) to Template Specs and Deployment Stacks before the January 31, 2027 retirement. Covers inventory, export, conversion to Bicep, policy decoupling, Template Spec publishing, Deployment Stack deployment with deny-settings, validation, and cutover.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[622,623,626,629],{"name":11,"slug":8,"type":16},{"name":624,"slug":625,"type":16},"Deployment","deployment",{"name":627,"slug":628,"type":16},"Infrastructure as Code","infrastructure-as-code",{"name":630,"slug":631,"type":16},"Migration","migration",260,"https:\u002F\u002Fgithub.com\u002FAzure\u002Fazure-blueprints","2026-07-12T08:17:49.646405",{"slug":636,"name":636,"fn":637,"description":638,"org":639,"tags":640,"stars":649,"repoUrl":650,"updatedAt":651},"apiview-feedback-resolution","resolve APIView feedback on Azure SDKs","Analyze and resolve APIView review feedback on Azure SDK PRs. **UTILITY SKILL**. USE FOR: APIView comments, API review feedback, SDK API surface changes. DO NOT USE FOR: general code review, non-APIView feedback. INVOKES: azure-sdk-mcp:azsdk_apiview_get_comments, azure-sdk-mcp:azsdk_typespec_customized_code_update.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[641,642,643,646],{"name":513,"slug":514,"type":16},{"name":11,"slug":8,"type":16},{"name":644,"slug":645,"type":16},"Code Review","code-review",{"name":647,"slug":648,"type":16},"Documentation","documentation",133,"https:\u002F\u002Fgithub.com\u002FAzure\u002Fazure-sdk-tools","2026-07-12T08:17:43.350876",{"slug":653,"name":653,"fn":654,"description":655,"org":656,"tags":657,"stars":649,"repoUrl":650,"updatedAt":666},"azsdk-common-live-and-recorded-tests","deploy resources and run Azure SDK tests","Deploy test resources and run Azure SDK tests in live, record, or playback mode. WHEN: \"run live tests\", \"run recorded tests\", \"deploy test resources\", \"record tests\", \"run tests in record mode\", \"clean up test resources\", \"run tests against live resources\". DO NOT USE FOR: writing new tests, authoring Bicep templates, playback-only test runs without resource deployment. INVOKES: azure-sdk-mcp:azsdk_package_run_tests.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[658,659,660,663],{"name":11,"slug":8,"type":16},{"name":624,"slug":625,"type":16},{"name":661,"slug":662,"type":16},"SDK","sdk",{"name":664,"slug":665,"type":16},"Testing","testing","2026-07-12T08:17:44.718943",{"slug":668,"name":668,"fn":669,"description":670,"org":671,"tags":672,"stars":649,"repoUrl":650,"updatedAt":681},"azsdk-common-prepare-release-plan","manage Azure SDK release plan work items","Create, get, update, abandon, and link SDK PRs to release plan work items for Azure SDK releases. **UTILITY SKILL**. USE FOR: \"create release plan\", \"get release plan\", \"update release plan\", \"update API spec in release plan\", \"update SDK details in release plan\", \"abandon release plan\", \"link SDK PR to plan\", \"namespace approval\", \"check release plan status\". DO NOT USE FOR: SDK code generation, pipeline troubleshooting, API review feedback. INVOKES: azure-sdk-mcp:azsdk_create_release_plan, azure-sdk-mcp:azsdk_get_release_plan, azure-sdk-mcp:azsdk_get_release_plan_for_spec_pr, azure-sdk-mcp:azsdk_update_release_plan, azure-sdk-mcp:azsdk_update_api_spec_pull_request_in_release_plan, azure-sdk-mcp:azsdk_update_sdk_details_in_release_plan, azure-sdk-mcp:azsdk_abandon_release_plan, azure-sdk-mcp:azsdk_link_sdk_pull_request_to_release_plan, azure-sdk-mcp:azsdk_link_namespace_approval_issue.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[673,674,677,680],{"name":11,"slug":8,"type":16},{"name":675,"slug":676,"type":16},"GitHub","github",{"name":678,"slug":679,"type":16},"Project Management","project-management",{"name":661,"slug":662,"type":16},"2026-07-12T08:17:38.345387",{"slug":683,"name":683,"fn":684,"description":685,"org":686,"tags":687,"stars":649,"repoUrl":650,"updatedAt":694},"azsdk-common-sdk-release","release Azure SDK packages","Check release readiness and trigger the release pipeline for Azure SDK packages. **UTILITY SKILL**. USE FOR: \"release SDK\", \"trigger release\", \"check release readiness\", \"release pipeline\", \"publish package\", \"ship SDK\". DO NOT USE FOR: SDK development, code generation, pipeline debugging, release plan creation. INVOKES: azure-sdk-mcp:azsdk_release_sdk.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[688,689,692,693],{"name":11,"slug":8,"type":16},{"name":690,"slug":691,"type":16},"CI\u002FCD","ci-cd",{"name":624,"slug":625,"type":16},{"name":661,"slug":662,"type":16},"2026-07-12T08:17:34.27607",{"slug":696,"name":696,"fn":697,"description":698,"org":699,"tags":700,"stars":649,"repoUrl":650,"updatedAt":709},"azure-typespec-author","author and modify Azure TypeSpec API specifications","Authors and modifies Azure TypeSpec (.tsp) API specifications. USE FOR: any TypeSpec\u002Ftsp change — api versions (add, bump, preview, stable, promote), resources, operations, models, properties, decorators, visibility, constraints, breaking changes, LRO, suppressions, operationId, spread model. Covers ARM resource-manager and data-plane services. DO NOT USE FOR: SDK generation, releasing SDK packages, or single MCP tool calls. INVOKES: azure-sdk-mcp:azsdk_typespec_generate_authoring_plan, azure-sdk-mcp:azsdk_run_typespec_validation.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[701,702,703,706],{"name":513,"slug":514,"type":16},{"name":11,"slug":8,"type":16},{"name":704,"slug":705,"type":16},"OpenAPI","openapi",{"name":707,"slug":708,"type":16},"Technical Writing","technical-writing","2026-07-12T08:17:39.603232",{"slug":711,"name":711,"fn":712,"description":713,"org":714,"tags":715,"stars":649,"repoUrl":650,"updatedAt":720},"generate-sdk-locally","generate and test Azure SDKs locally","Generate, build, and test Azure SDKs locally from TypeSpec with automatic customization. WHEN: \"generate SDK locally\", \"build SDK\", \"run SDK tests\", \"run CI checks\", \"validate package\", \"run checks\", \"update changelog\", \"fix SDK build errors\", \"fix breaking changes\", \"resolve SDK generation errors\", \"customize TypeSpec\", \"rename SDK client\", \"rename SDK model\", \"hide operation from SDK\", \"fix analyzer errors\", \"resolve customization drift\", \"create subclient\", \"update metadata\", \"update version\". DO NOT USE FOR: publishing to package registries, CI pipeline configuration, API design review. INVOKES: azsdk_verify_setup, azsdk_package_generate_code, azsdk_package_build_code, azsdk_package_run_check, azsdk_package_run_tests, azsdk_customized_code_update, azsdk_package_update_changelog_content, azsdk_package_update_metadata, azsdk_package_update_version.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[716,717,718,719],{"name":11,"slug":8,"type":16},{"name":690,"slug":691,"type":16},{"name":661,"slug":662,"type":16},{"name":664,"slug":665,"type":16},"2026-07-12T08:17:37.08523",{"slug":722,"name":722,"fn":723,"description":724,"org":725,"tags":726,"stars":649,"repoUrl":650,"updatedAt":732},"markdown-token-optimizer","optimize markdown files for token efficiency","Analyze markdown files for token efficiency and reduce context-window bloat. **UTILITY SKILL**. DO NOT USE FOR: code optimization, general file editing, non-markdown files. TRIGGERS: optimize markdown, reduce tokens, token count, token bloat, too many tokens, make concise, shrink file, file too large, optimize for AI, token efficiency, verbose markdown, reduce file size. INVOKES: waza CLI.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[727,728,731],{"name":563,"slug":564,"type":16},{"name":729,"slug":730,"type":16},"Performance","performance",{"name":707,"slug":708,"type":16},"2026-07-12T08:17:42.080413",{"slug":734,"name":734,"fn":735,"description":736,"org":737,"tags":738,"stars":649,"repoUrl":650,"updatedAt":745},"pipeline-troubleshooting","troubleshoot Azure SDK CI pipelines","Diagnose and resolve failures in Azure SDK CI and generation pipelines. **UTILITY SKILL**. USE FOR: \"pipeline failed\", \"build failure\", \"CI check failing\", \"SDK generation error\", \"reproduce pipeline locally\", \"debug SDK pipeline\". DO NOT USE FOR: local build issues without pipeline context, API design review, SDK publishing. INVOKES: azure-sdk-mcp:azsdk_analyze_pipeline, azure-sdk-mcp:azsdk_package_build_code, azure-sdk-mcp:azsdk_package_run_check.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[739,740,741,744],{"name":11,"slug":8,"type":16},{"name":690,"slug":691,"type":16},{"name":742,"slug":743,"type":16},"Debugging","debugging",{"name":661,"slug":662,"type":16},"2026-07-12T08:17:40.821512",{"slug":747,"name":747,"fn":748,"description":749,"org":750,"tags":751,"stars":649,"repoUrl":650,"updatedAt":758},"sensei","improve skill frontmatter compliance","**WORKFLOW SKILL** — Iteratively improve skill frontmatter compliance using the Ralph loop pattern. WHEN: \"run sensei\", \"sensei help\", \"improve skill\", \"fix frontmatter\", \"skill compliance\", \"frontmatter audit\", \"score skill\", \"check skill tokens\". INVOKES: token counting tools, test runners, git commands. FOR SINGLE OPERATIONS: use token CLI directly for counts\u002Fchecks.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[752,753,754,757],{"name":11,"slug":8,"type":16},{"name":609,"slug":610,"type":16},{"name":755,"slug":756,"type":16},"Process Optimization","process-optimization",{"name":707,"slug":708,"type":16},"2026-07-12T08:17:32.970921",{"slug":760,"name":760,"fn":761,"description":762,"org":763,"tags":764,"stars":649,"repoUrl":650,"updatedAt":770},"skill-authoring","author agent skills for agentskills.io","Write Agent Skills that comply with the agentskills.io specification. WHEN: \"create a skill\", \"new skill\", \"write a skill\", \"skill template\", \"skill structure\", \"review skill\", \"skill PR\", \"skill compliance\", \"SKILL.md format\", \"skill frontmatter\", \"skill best practices\". DO NOT USE FOR: improving existing skills (use sensei), general documentation. INVOKES: waza CLI.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[765,766,769],{"name":647,"slug":648,"type":16},{"name":767,"slug":768,"type":16},"Plugin Development","plugin-development",{"name":707,"slug":708,"type":16},"2026-07-12T08:17:35.873862",109]