[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill-azure-mesh-federation":3,"mdc--seu7t5-key":52,"related-repo-azure-mesh-federation":1453,"related-org-azure-mesh-federation":1567},{"slug":4,"name":4,"fn":5,"description":6,"org":7,"tags":12,"stars":24,"repoUrl":25,"updatedAt":26,"license":27,"forks":28,"topics":29,"repo":47,"sourceUrl":50,"mdContent":51},"mesh-federation","offload tasks to cloud sandboxes","Pair with a kars cluster and offload heavy tasks to governed cloud sandboxes with GPU \u002F foundation-model inference \u002F Azure AI services, or communicate with other agents over end-to-end encrypted AgentMesh. Triggers on natural-language intents like \"offload to the cloud\", \"run this on Azure\", \"ask my cluster to…\", \"send a message to agent X\", \"who is on the mesh\", \"check my inbox\", \"is my offload done\".",{"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",[13,17,18,21],{"name":14,"slug":15,"type":16},"Security","security","tag",{"name":11,"slug":8,"type":16},{"name":19,"slug":20,"type":16},"Agents","agents",{"name":22,"slug":23,"type":16},"AI Infrastructure","ai-infrastructure",28,"https:\u002F\u002Fgithub.com\u002FAzure\u002Fkars","2026-07-12T08:18:27.878411",null,2,[30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46],"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":25,"stars":24,"forks":28,"topics":48,"description":49},[30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46],"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\u002Fmesh-plugin\u002Fskills\u002Fmesh-federation","---\nname: mesh-federation\ndescription: Pair with a kars cluster and offload heavy tasks to governed cloud sandboxes with GPU \u002F foundation-model inference \u002F Azure AI services, or communicate with other agents over end-to-end encrypted AgentMesh. Triggers on natural-language intents like \"offload to the cloud\", \"run this on Azure\", \"ask my cluster to…\", \"send a message to agent X\", \"who is on the mesh\", \"check my inbox\", \"is my offload done\".\nmetadata: {\"openclaw\": {\"always\": true}}\n---\n\n# kars Mesh Federation — Cloud Offload & Inter-Agent Messaging\n\nYou have the **kars Mesh** plugin installed. It federates this OpenClaw agent with a kars cluster and other agents via end-to-end encrypted AgentMesh (Signal Protocol). You can:\n\n1. Delegate heavy\u002Flong-running tasks to a governed cloud sandbox that has GPU, Azure AI Foundry models, Content Safety, and AGT governance.\n2. Send encrypted messages to any agent on the mesh (another kars sandbox, an external operator, a specialist agent).\n3. Discover agents by capability or name.\n4. Read your encrypted inbox.\n\n---\n\n## When to pick this skill\n\nUse this skill when the user's intent matches any of:\n\n- **\"offload this to the cloud\"**, \"run this on Azure\", \"run it on my cluster\", \"do it in the cloud sandbox\", \"delegate to the cluster\"\n- \"too heavy for local\" \u002F \"I need GPU\" \u002F \"this will take too long here\" \u002F \"use the big model\"\n- \"pair with the cluster\", \"connect to kars\", \"set up federation\"\n- \"send a message to \u003Cagent>\", \"tell \u003Cagent> to …\", \"notify \u003Cagent>\"\n- \"check my inbox\", \"any new messages?\", \"did \u003Cagent> reply?\"\n- \"who's on the mesh?\", \"find agents that can \u003CX>\", \"discover \u003Ccapability>\"\n- \"how's the offload going?\", \"is the task done?\", \"status of the job\"\n\nAny of the above should prompt you to reach for the tools below instead of attempting the work locally (especially if local tools would be slow, require a large model, or lack the capability entirely).\n\n---\n\n## Tool reference\n\n### `mesh_pair` — one-time pairing\n**Signature:** `mesh_pair(token: string)`\n\n**When:** Before any other tool. If the user gives you a string starting with `azcp_1_…`, immediately call `mesh_pair` with it. Pairing state is persisted to `~\u002F.kars\u002Fpairings.json` — you only pair once per cluster.\n\n**Returns:** A success block with your AMID, token budget, and relay RTT.\n\n---\n\n### `cloud_offload` — delegate a task to the cluster\n**Signature:** `cloud_offload(task: string, files?: string[], model?: string, timeout_minutes?: number)`\n\n**Parameter aliases accepted** (use whichever feels natural; all resolve to `task`): `task | prompt | description | request | content | instruction | query`. For files: `files | file_paths`. For timeout: `timeout_minutes | timeout`.\n\n**Behaviour:**\n- Returns immediately (typically in \u003C1 s) with a confirmation and a `request_id`. The actual work happens asynchronously on the cluster.\n- The cluster spawns a fresh, isolated sandbox pod for this task. The sandbox gets the task via env vars and **proactively announces itself** to you with an `offload_hello` message the moment it's online.\n- After announcing, the sandbox streams progress heartbeats (every ~20 s), output files, and a final summary — all via E2E-encrypted mesh messages. You do **not** need to poll for these; just call `offload_status`.\n\n**When to use:**\n- The user asks for deep research, large refactors, security audits, data processing, or anything GPU-heavy.\n- Any request that would take more than a minute or two locally.\n- Any request that needs a capability you don't have here (e.g., a larger model, Azure Content Safety, Foundry memory store).\n\n**Examples:**\n```\ncloud_offload(task: \"Analyze this codebase for OWASP Top 10 vulns; focus on SQLi and XSS. Return a markdown report.\", files: [\"src\u002Fauth.ts\",\"src\u002Fdb.ts\"], timeout_minutes: 15)\ncloud_offload(task: \"Research the latest community trends in AI agent sandboxing: GitHub repos, blog posts, open issues. Produce a markdown summary with links.\", timeout_minutes: 30)\n```\n\n**Limits:**\n- 30 MB per file. Total task size is bounded by your paired budget (default 500 K tokens).\n- One offload in-flight per pairing. If one is active, finish or cancel it before starting another.\n\n---\n\n### `offload_status` — watch progress\n**Signature:** `offload_status()`  *(no arguments)*\n\n**Returns:** A human-readable snapshot of the current offload with the most recent stages. Phases progress through:\n\n| Phase | Icon | Meaning |\n|---|---|---|\n| `submitted` | 📤 | Request packaged locally |\n| `validating` | 🔎 | Cluster checking budget \u002F pairing |\n| `spawning` | 🚀 | Controller creating the sandbox CRD |\n| `scheduled` | 📅 | Kubernetes scheduling the pod |\n| `ready` | 🟢 | Sandbox pod running |\n| `acknowledged` | 👋 | Sandbox sent `offload_hello` — task accepted and running |\n| `verifying` | 🏓 | (legacy fallback) pinging sandbox over mesh |\n| `uploading` | 📦 | Transferring input files |\n| `running` | ⚙️ | Task executing; heartbeats arriving every ~20 s |\n| `returning` | 📥 | Sandbox streaming output files back |\n| `done` | ✅ | Task complete — summary + files ready |\n| `error` | ❌ | Something failed — error message shown |\n\n**When:** Call every ~5–15 s while waiting for a task. Stop as soon as phase is `done` or `error`. Do **not** tight-loop — one call per conversational turn is plenty.\n\n---\n\n### `mesh_send` — send an encrypted message to another agent\n**Signature:** `mesh_send(to: string, message: string)`\n\n**Parameter aliases:** `to | to_amid | to_agent | target | amid | recipient` for the recipient; `message | content | body | text` for the body. You can pass a display name (e.g. `\"offload-parent\"`) or a raw AMID (base58, 20+ chars) — display names are resolved via the registry.\n\n**When:**\n- Relaying a message to another agent (\"ask the security-auditor to re-check PR #42\").\n- Notifying an offload sandbox you spawned (e.g., \"stop, I got the answer locally\").\n- Any peer-to-peer coordination.\n\n**Example:**\n```\nmesh_send(to: \"offload-parent\", message: \"I've finished the initial draft — see files in your inbox\")\n```\n\n---\n\n### `mesh_inbox` — read incoming messages\n**Signature:** `mesh_inbox(limit?: number)`\n\n**When:** The user asks \"any updates?\", \"did X reply?\", or to check on an offload that you started a while ago. Also use this as the primary way to pick up progress\u002Fresults from a cloud offload (offload_status is a cleaner formatted view of the same messages).\n\n---\n\n### `discover` — list agents on the mesh\n**Signature:** `discover(capability?: string, limit?: number)`\n\n**When:** The user asks who is on the mesh, or you want to find an agent with a specific capability (e.g., `\"security-audit\"`, `\"offload\"`, `\"assistant\"`). Without a capability, `discover` aggregates across well-known seed capabilities (kars-agent, task-execution, cluster-controller, offload, mesh-peer, etc.) and returns a deduped list.\n\n---\n\n## Typical conversation flows\n\n### Pairing then offloading\n```\nUser: here's a pairing token azcp_1_eyJ...\n  → mesh_pair(token: \"azcp_1_eyJ...\")\nUser: great — now research latest trends in AI agent sandboxing\n  → cloud_offload(task: \"research latest trends in AI agent sandboxing ...\")\n  → (brief wait, then) offload_status()\n  → Continue calling offload_status each turn until phase=done\n  → Report summary + files back to the user\n```\n\n### Checking on a running offload\n```\nUser: how's that research going?\n  → offload_status()\n  → Report current phase + elapsed time + last heartbeat\n```\n\n### Sending a peer message\n```\nUser: tell the cluster sandbox I'm done\n  → mesh_send(to: \"offload-sandbox-name\", message: \"...\")\n```\n\n---\n\n## Important operational notes\n\n- **Pairing is pre-req for everything except `mesh_pair` itself.** If a tool returns \"Not paired\", stop and ask the user for a pairing token.\n- **Proactive offload:** Modern kars sandboxes *self-announce* the moment they're ready and begin work immediately. You'll see phase `acknowledged` (👋) within ~20 s of `ready`. If you don't, the cluster may be running an older sandbox image and will fall back to the ping\u002Fupload\u002Fdispatch flow (still works, just slower).\n- **Tokens are metered.** `offload_status` surfaces budget consumption. If an offload fails with \"budget exceeded\", ask the admin for a new pairing token with a larger budget.\n- **No secrets in tasks.** Never put API keys or passwords in `cloud_offload` task text or files. The sandbox has its own managed identity for Azure access.\n- **One offload at a time** per pairing. Cancel (or let it finish) before starting another.\n- **Files are transferred via E2E-encrypted chunked transfer** with SHA-256 integrity. Each file ≤ 30 MB.\n- **Timeouts:** default 30 min; max configurable via `timeout_minutes`. The sandbox auto-terminates on timeout.\n\n## Common errors\n\n| Message | Cause | Fix |\n|---|---|---|\n| `❌ cloud_offload requires a task parameter…` | No recognizable task field passed | Re-issue with `task: \"...\"` |\n| `❌ Not paired with any kars cluster` | No pairing persisted | Call `mesh_pair` with a valid token |\n| `❌ Mesh connection is not live` | WebSocket to relay dropped | Wait a few seconds — plugin auto-reconnects; retry |\n| `❌ Offload already in progress` | Prior offload still running | `offload_status` to check \u002F wait for `done` |\n| `❌ Sandbox '...' is registered but not responding to mesh pings` | Legacy flow, sandbox came up but E2E not ready | Usually self-heals on retry; modern sandboxes skip this via `offload_hello` |\n| `No agents found on the mesh.` | No agents registered under queried capability | Try without filter, or use a different capability name |\n\n",{"data":53,"body":57},{"name":4,"description":6,"metadata":54},{"openclaw":55},{"always":56},true,{"type":58,"children":59},"root",[60,69,83,108,112,119,124,199,204,207,213,227,243,276,286,289,301,315,357,365,420,428,446,454,466,474,487,490,501,523,532,836,865,868,880,894,927,934,952,960,969,972,984,998,1007,1010,1022,1036,1075,1078,1084,1090,1099,1105,1114,1120,1129,1132,1138,1260,1266],{"type":61,"tag":62,"props":63,"children":65},"element","h1",{"id":64},"kars-mesh-federation-cloud-offload-inter-agent-messaging",[66],{"type":67,"value":68},"text","kars Mesh Federation — Cloud Offload & Inter-Agent Messaging",{"type":61,"tag":70,"props":71,"children":72},"p",{},[73,75,81],{"type":67,"value":74},"You have the ",{"type":61,"tag":76,"props":77,"children":78},"strong",{},[79],{"type":67,"value":80},"kars Mesh",{"type":67,"value":82}," plugin installed. It federates this OpenClaw agent with a kars cluster and other agents via end-to-end encrypted AgentMesh (Signal Protocol). You can:",{"type":61,"tag":84,"props":85,"children":86},"ol",{},[87,93,98,103],{"type":61,"tag":88,"props":89,"children":90},"li",{},[91],{"type":67,"value":92},"Delegate heavy\u002Flong-running tasks to a governed cloud sandbox that has GPU, Azure AI Foundry models, Content Safety, and AGT governance.",{"type":61,"tag":88,"props":94,"children":95},{},[96],{"type":67,"value":97},"Send encrypted messages to any agent on the mesh (another kars sandbox, an external operator, a specialist agent).",{"type":61,"tag":88,"props":99,"children":100},{},[101],{"type":67,"value":102},"Discover agents by capability or name.",{"type":61,"tag":88,"props":104,"children":105},{},[106],{"type":67,"value":107},"Read your encrypted inbox.",{"type":61,"tag":109,"props":110,"children":111},"hr",{},[],{"type":61,"tag":113,"props":114,"children":116},"h2",{"id":115},"when-to-pick-this-skill",[117],{"type":67,"value":118},"When to pick this skill",{"type":61,"tag":70,"props":120,"children":121},{},[122],{"type":67,"value":123},"Use this skill when the user's intent matches any of:",{"type":61,"tag":125,"props":126,"children":127},"ul",{},[128,138,143,148,168,178,194],{"type":61,"tag":88,"props":129,"children":130},{},[131,136],{"type":61,"tag":76,"props":132,"children":133},{},[134],{"type":67,"value":135},"\"offload this to the cloud\"",{"type":67,"value":137},", \"run this on Azure\", \"run it on my cluster\", \"do it in the cloud sandbox\", \"delegate to the cluster\"",{"type":61,"tag":88,"props":139,"children":140},{},[141],{"type":67,"value":142},"\"too heavy for local\" \u002F \"I need GPU\" \u002F \"this will take too long here\" \u002F \"use the big model\"",{"type":61,"tag":88,"props":144,"children":145},{},[146],{"type":67,"value":147},"\"pair with the cluster\", \"connect to kars\", \"set up federation\"",{"type":61,"tag":88,"props":149,"children":150},{},[151,153],{"type":67,"value":152},"\"send a message to ",{"type":61,"tag":30,"props":154,"children":155},{},[156,158],{"type":67,"value":157},"\", \"tell ",{"type":61,"tag":30,"props":159,"children":160},{},[161,163],{"type":67,"value":162}," to …\", \"notify ",{"type":61,"tag":30,"props":164,"children":165},{},[166],{"type":67,"value":167},"\"",{"type":61,"tag":88,"props":169,"children":170},{},[171,173],{"type":67,"value":172},"\"check my inbox\", \"any new messages?\", \"did ",{"type":61,"tag":30,"props":174,"children":175},{},[176],{"type":67,"value":177}," reply?\"",{"type":61,"tag":88,"props":179,"children":180},{},[181,183],{"type":67,"value":182},"\"who's on the mesh?\", \"find agents that can ",{"type":61,"tag":184,"props":185,"children":186},"x",{},[187,189],{"type":67,"value":188},"\", \"discover ",{"type":61,"tag":190,"props":191,"children":192},"capability",{},[193],{"type":67,"value":167},{"type":61,"tag":88,"props":195,"children":196},{},[197],{"type":67,"value":198},"\"how's the offload going?\", \"is the task done?\", \"status of the job\"",{"type":61,"tag":70,"props":200,"children":201},{},[202],{"type":67,"value":203},"Any of the above should prompt you to reach for the tools below instead of attempting the work locally (especially if local tools would be slow, require a large model, or lack the capability entirely).",{"type":61,"tag":109,"props":205,"children":206},{},[],{"type":61,"tag":113,"props":208,"children":210},{"id":209},"tool-reference",[211],{"type":67,"value":212},"Tool reference",{"type":61,"tag":214,"props":215,"children":217},"h3",{"id":216},"mesh_pair-one-time-pairing",[218,225],{"type":61,"tag":219,"props":220,"children":222},"code",{"className":221},[],[223],{"type":67,"value":224},"mesh_pair",{"type":67,"value":226}," — one-time pairing",{"type":61,"tag":70,"props":228,"children":229},{},[230,235,237],{"type":61,"tag":76,"props":231,"children":232},{},[233],{"type":67,"value":234},"Signature:",{"type":67,"value":236}," ",{"type":61,"tag":219,"props":238,"children":240},{"className":239},[],[241],{"type":67,"value":242},"mesh_pair(token: string)",{"type":61,"tag":70,"props":244,"children":245},{},[246,251,253,259,261,266,268,274],{"type":61,"tag":76,"props":247,"children":248},{},[249],{"type":67,"value":250},"When:",{"type":67,"value":252}," Before any other tool. If the user gives you a string starting with ",{"type":61,"tag":219,"props":254,"children":256},{"className":255},[],[257],{"type":67,"value":258},"azcp_1_…",{"type":67,"value":260},", immediately call ",{"type":61,"tag":219,"props":262,"children":264},{"className":263},[],[265],{"type":67,"value":224},{"type":67,"value":267}," with it. Pairing state is persisted to ",{"type":61,"tag":219,"props":269,"children":271},{"className":270},[],[272],{"type":67,"value":273},"~\u002F.kars\u002Fpairings.json",{"type":67,"value":275}," — you only pair once per cluster.",{"type":61,"tag":70,"props":277,"children":278},{},[279,284],{"type":61,"tag":76,"props":280,"children":281},{},[282],{"type":67,"value":283},"Returns:",{"type":67,"value":285}," A success block with your AMID, token budget, and relay RTT.",{"type":61,"tag":109,"props":287,"children":288},{},[],{"type":61,"tag":214,"props":290,"children":292},{"id":291},"cloud_offload-delegate-a-task-to-the-cluster",[293,299],{"type":61,"tag":219,"props":294,"children":296},{"className":295},[],[297],{"type":67,"value":298},"cloud_offload",{"type":67,"value":300}," — delegate a task to the cluster",{"type":61,"tag":70,"props":302,"children":303},{},[304,308,309],{"type":61,"tag":76,"props":305,"children":306},{},[307],{"type":67,"value":234},{"type":67,"value":236},{"type":61,"tag":219,"props":310,"children":312},{"className":311},[],[313],{"type":67,"value":314},"cloud_offload(task: string, files?: string[], model?: string, timeout_minutes?: number)",{"type":61,"tag":70,"props":316,"children":317},{},[318,323,325,331,333,339,341,347,349,355],{"type":61,"tag":76,"props":319,"children":320},{},[321],{"type":67,"value":322},"Parameter aliases accepted",{"type":67,"value":324}," (use whichever feels natural; all resolve to ",{"type":61,"tag":219,"props":326,"children":328},{"className":327},[],[329],{"type":67,"value":330},"task",{"type":67,"value":332},"): ",{"type":61,"tag":219,"props":334,"children":336},{"className":335},[],[337],{"type":67,"value":338},"task | prompt | description | request | content | instruction | query",{"type":67,"value":340},". For files: ",{"type":61,"tag":219,"props":342,"children":344},{"className":343},[],[345],{"type":67,"value":346},"files | file_paths",{"type":67,"value":348},". For timeout: ",{"type":61,"tag":219,"props":350,"children":352},{"className":351},[],[353],{"type":67,"value":354},"timeout_minutes | timeout",{"type":67,"value":356},".",{"type":61,"tag":70,"props":358,"children":359},{},[360],{"type":61,"tag":76,"props":361,"children":362},{},[363],{"type":67,"value":364},"Behaviour:",{"type":61,"tag":125,"props":366,"children":367},{},[368,381,401],{"type":61,"tag":88,"props":369,"children":370},{},[371,373,379],{"type":67,"value":372},"Returns immediately (typically in \u003C1 s) with a confirmation and a ",{"type":61,"tag":219,"props":374,"children":376},{"className":375},[],[377],{"type":67,"value":378},"request_id",{"type":67,"value":380},". The actual work happens asynchronously on the cluster.",{"type":61,"tag":88,"props":382,"children":383},{},[384,386,391,393,399],{"type":67,"value":385},"The cluster spawns a fresh, isolated sandbox pod for this task. The sandbox gets the task via env vars and ",{"type":61,"tag":76,"props":387,"children":388},{},[389],{"type":67,"value":390},"proactively announces itself",{"type":67,"value":392}," to you with an ",{"type":61,"tag":219,"props":394,"children":396},{"className":395},[],[397],{"type":67,"value":398},"offload_hello",{"type":67,"value":400}," message the moment it's online.",{"type":61,"tag":88,"props":402,"children":403},{},[404,406,411,413,419],{"type":67,"value":405},"After announcing, the sandbox streams progress heartbeats (every ~20 s), output files, and a final summary — all via E2E-encrypted mesh messages. You do ",{"type":61,"tag":76,"props":407,"children":408},{},[409],{"type":67,"value":410},"not",{"type":67,"value":412}," need to poll for these; just call ",{"type":61,"tag":219,"props":414,"children":416},{"className":415},[],[417],{"type":67,"value":418},"offload_status",{"type":67,"value":356},{"type":61,"tag":70,"props":421,"children":422},{},[423],{"type":61,"tag":76,"props":424,"children":425},{},[426],{"type":67,"value":427},"When to use:",{"type":61,"tag":125,"props":429,"children":430},{},[431,436,441],{"type":61,"tag":88,"props":432,"children":433},{},[434],{"type":67,"value":435},"The user asks for deep research, large refactors, security audits, data processing, or anything GPU-heavy.",{"type":61,"tag":88,"props":437,"children":438},{},[439],{"type":67,"value":440},"Any request that would take more than a minute or two locally.",{"type":61,"tag":88,"props":442,"children":443},{},[444],{"type":67,"value":445},"Any request that needs a capability you don't have here (e.g., a larger model, Azure Content Safety, Foundry memory store).",{"type":61,"tag":70,"props":447,"children":448},{},[449],{"type":61,"tag":76,"props":450,"children":451},{},[452],{"type":67,"value":453},"Examples:",{"type":61,"tag":455,"props":456,"children":460},"pre",{"className":457,"code":459,"language":67},[458],"language-text","cloud_offload(task: \"Analyze this codebase for OWASP Top 10 vulns; focus on SQLi and XSS. Return a markdown report.\", files: [\"src\u002Fauth.ts\",\"src\u002Fdb.ts\"], timeout_minutes: 15)\ncloud_offload(task: \"Research the latest community trends in AI agent sandboxing: GitHub repos, blog posts, open issues. Produce a markdown summary with links.\", timeout_minutes: 30)\n",[461],{"type":61,"tag":219,"props":462,"children":464},{"__ignoreMap":463},"",[465],{"type":67,"value":459},{"type":61,"tag":70,"props":467,"children":468},{},[469],{"type":61,"tag":76,"props":470,"children":471},{},[472],{"type":67,"value":473},"Limits:",{"type":61,"tag":125,"props":475,"children":476},{},[477,482],{"type":61,"tag":88,"props":478,"children":479},{},[480],{"type":67,"value":481},"30 MB per file. Total task size is bounded by your paired budget (default 500 K tokens).",{"type":61,"tag":88,"props":483,"children":484},{},[485],{"type":67,"value":486},"One offload in-flight per pairing. If one is active, finish or cancel it before starting another.",{"type":61,"tag":109,"props":488,"children":489},{},[],{"type":61,"tag":214,"props":491,"children":493},{"id":492},"offload_status-watch-progress",[494,499],{"type":61,"tag":219,"props":495,"children":497},{"className":496},[],[498],{"type":67,"value":418},{"type":67,"value":500}," — watch progress",{"type":61,"tag":70,"props":502,"children":503},{},[504,508,509,515,517],{"type":61,"tag":76,"props":505,"children":506},{},[507],{"type":67,"value":234},{"type":67,"value":236},{"type":61,"tag":219,"props":510,"children":512},{"className":511},[],[513],{"type":67,"value":514},"offload_status()",{"type":67,"value":516},"  ",{"type":61,"tag":518,"props":519,"children":520},"em",{},[521],{"type":67,"value":522},"(no arguments)",{"type":61,"tag":70,"props":524,"children":525},{},[526,530],{"type":61,"tag":76,"props":527,"children":528},{},[529],{"type":67,"value":283},{"type":67,"value":531}," A human-readable snapshot of the current offload with the most recent stages. Phases progress through:",{"type":61,"tag":533,"props":534,"children":535},"table",{},[536,560],{"type":61,"tag":537,"props":538,"children":539},"thead",{},[540],{"type":61,"tag":541,"props":542,"children":543},"tr",{},[544,550,555],{"type":61,"tag":545,"props":546,"children":547},"th",{},[548],{"type":67,"value":549},"Phase",{"type":61,"tag":545,"props":551,"children":552},{},[553],{"type":67,"value":554},"Icon",{"type":61,"tag":545,"props":556,"children":557},{},[558],{"type":67,"value":559},"Meaning",{"type":61,"tag":561,"props":562,"children":563},"tbody",{},[564,587,609,631,653,675,704,726,748,770,792,814],{"type":61,"tag":541,"props":565,"children":566},{},[567,577,582],{"type":61,"tag":568,"props":569,"children":570},"td",{},[571],{"type":61,"tag":219,"props":572,"children":574},{"className":573},[],[575],{"type":67,"value":576},"submitted",{"type":61,"tag":568,"props":578,"children":579},{},[580],{"type":67,"value":581},"📤",{"type":61,"tag":568,"props":583,"children":584},{},[585],{"type":67,"value":586},"Request packaged locally",{"type":61,"tag":541,"props":588,"children":589},{},[590,599,604],{"type":61,"tag":568,"props":591,"children":592},{},[593],{"type":61,"tag":219,"props":594,"children":596},{"className":595},[],[597],{"type":67,"value":598},"validating",{"type":61,"tag":568,"props":600,"children":601},{},[602],{"type":67,"value":603},"🔎",{"type":61,"tag":568,"props":605,"children":606},{},[607],{"type":67,"value":608},"Cluster checking budget \u002F pairing",{"type":61,"tag":541,"props":610,"children":611},{},[612,621,626],{"type":61,"tag":568,"props":613,"children":614},{},[615],{"type":61,"tag":219,"props":616,"children":618},{"className":617},[],[619],{"type":67,"value":620},"spawning",{"type":61,"tag":568,"props":622,"children":623},{},[624],{"type":67,"value":625},"🚀",{"type":61,"tag":568,"props":627,"children":628},{},[629],{"type":67,"value":630},"Controller creating the sandbox CRD",{"type":61,"tag":541,"props":632,"children":633},{},[634,643,648],{"type":61,"tag":568,"props":635,"children":636},{},[637],{"type":61,"tag":219,"props":638,"children":640},{"className":639},[],[641],{"type":67,"value":642},"scheduled",{"type":61,"tag":568,"props":644,"children":645},{},[646],{"type":67,"value":647},"📅",{"type":61,"tag":568,"props":649,"children":650},{},[651],{"type":67,"value":652},"Kubernetes scheduling the pod",{"type":61,"tag":541,"props":654,"children":655},{},[656,665,670],{"type":61,"tag":568,"props":657,"children":658},{},[659],{"type":61,"tag":219,"props":660,"children":662},{"className":661},[],[663],{"type":67,"value":664},"ready",{"type":61,"tag":568,"props":666,"children":667},{},[668],{"type":67,"value":669},"🟢",{"type":61,"tag":568,"props":671,"children":672},{},[673],{"type":67,"value":674},"Sandbox pod running",{"type":61,"tag":541,"props":676,"children":677},{},[678,687,692],{"type":61,"tag":568,"props":679,"children":680},{},[681],{"type":61,"tag":219,"props":682,"children":684},{"className":683},[],[685],{"type":67,"value":686},"acknowledged",{"type":61,"tag":568,"props":688,"children":689},{},[690],{"type":67,"value":691},"👋",{"type":61,"tag":568,"props":693,"children":694},{},[695,697,702],{"type":67,"value":696},"Sandbox sent ",{"type":61,"tag":219,"props":698,"children":700},{"className":699},[],[701],{"type":67,"value":398},{"type":67,"value":703}," — task accepted and running",{"type":61,"tag":541,"props":705,"children":706},{},[707,716,721],{"type":61,"tag":568,"props":708,"children":709},{},[710],{"type":61,"tag":219,"props":711,"children":713},{"className":712},[],[714],{"type":67,"value":715},"verifying",{"type":61,"tag":568,"props":717,"children":718},{},[719],{"type":67,"value":720},"🏓",{"type":61,"tag":568,"props":722,"children":723},{},[724],{"type":67,"value":725},"(legacy fallback) pinging sandbox over mesh",{"type":61,"tag":541,"props":727,"children":728},{},[729,738,743],{"type":61,"tag":568,"props":730,"children":731},{},[732],{"type":61,"tag":219,"props":733,"children":735},{"className":734},[],[736],{"type":67,"value":737},"uploading",{"type":61,"tag":568,"props":739,"children":740},{},[741],{"type":67,"value":742},"📦",{"type":61,"tag":568,"props":744,"children":745},{},[746],{"type":67,"value":747},"Transferring input files",{"type":61,"tag":541,"props":749,"children":750},{},[751,760,765],{"type":61,"tag":568,"props":752,"children":753},{},[754],{"type":61,"tag":219,"props":755,"children":757},{"className":756},[],[758],{"type":67,"value":759},"running",{"type":61,"tag":568,"props":761,"children":762},{},[763],{"type":67,"value":764},"⚙️",{"type":61,"tag":568,"props":766,"children":767},{},[768],{"type":67,"value":769},"Task executing; heartbeats arriving every ~20 s",{"type":61,"tag":541,"props":771,"children":772},{},[773,782,787],{"type":61,"tag":568,"props":774,"children":775},{},[776],{"type":61,"tag":219,"props":777,"children":779},{"className":778},[],[780],{"type":67,"value":781},"returning",{"type":61,"tag":568,"props":783,"children":784},{},[785],{"type":67,"value":786},"📥",{"type":61,"tag":568,"props":788,"children":789},{},[790],{"type":67,"value":791},"Sandbox streaming output files back",{"type":61,"tag":541,"props":793,"children":794},{},[795,804,809],{"type":61,"tag":568,"props":796,"children":797},{},[798],{"type":61,"tag":219,"props":799,"children":801},{"className":800},[],[802],{"type":67,"value":803},"done",{"type":61,"tag":568,"props":805,"children":806},{},[807],{"type":67,"value":808},"✅",{"type":61,"tag":568,"props":810,"children":811},{},[812],{"type":67,"value":813},"Task complete — summary + files ready",{"type":61,"tag":541,"props":815,"children":816},{},[817,826,831],{"type":61,"tag":568,"props":818,"children":819},{},[820],{"type":61,"tag":219,"props":821,"children":823},{"className":822},[],[824],{"type":67,"value":825},"error",{"type":61,"tag":568,"props":827,"children":828},{},[829],{"type":67,"value":830},"❌",{"type":61,"tag":568,"props":832,"children":833},{},[834],{"type":67,"value":835},"Something failed — error message shown",{"type":61,"tag":70,"props":837,"children":838},{},[839,843,845,850,852,857,859,863],{"type":61,"tag":76,"props":840,"children":841},{},[842],{"type":67,"value":250},{"type":67,"value":844}," Call every ~5–15 s while waiting for a task. Stop as soon as phase is ",{"type":61,"tag":219,"props":846,"children":848},{"className":847},[],[849],{"type":67,"value":803},{"type":67,"value":851}," or ",{"type":61,"tag":219,"props":853,"children":855},{"className":854},[],[856],{"type":67,"value":825},{"type":67,"value":858},". Do ",{"type":61,"tag":76,"props":860,"children":861},{},[862],{"type":67,"value":410},{"type":67,"value":864}," tight-loop — one call per conversational turn is plenty.",{"type":61,"tag":109,"props":866,"children":867},{},[],{"type":61,"tag":214,"props":869,"children":871},{"id":870},"mesh_send-send-an-encrypted-message-to-another-agent",[872,878],{"type":61,"tag":219,"props":873,"children":875},{"className":874},[],[876],{"type":67,"value":877},"mesh_send",{"type":67,"value":879}," — send an encrypted message to another agent",{"type":61,"tag":70,"props":881,"children":882},{},[883,887,888],{"type":61,"tag":76,"props":884,"children":885},{},[886],{"type":67,"value":234},{"type":67,"value":236},{"type":61,"tag":219,"props":889,"children":891},{"className":890},[],[892],{"type":67,"value":893},"mesh_send(to: string, message: string)",{"type":61,"tag":70,"props":895,"children":896},{},[897,902,903,909,911,917,919,925],{"type":61,"tag":76,"props":898,"children":899},{},[900],{"type":67,"value":901},"Parameter aliases:",{"type":67,"value":236},{"type":61,"tag":219,"props":904,"children":906},{"className":905},[],[907],{"type":67,"value":908},"to | to_amid | to_agent | target | amid | recipient",{"type":67,"value":910}," for the recipient; ",{"type":61,"tag":219,"props":912,"children":914},{"className":913},[],[915],{"type":67,"value":916},"message | content | body | text",{"type":67,"value":918}," for the body. You can pass a display name (e.g. ",{"type":61,"tag":219,"props":920,"children":922},{"className":921},[],[923],{"type":67,"value":924},"\"offload-parent\"",{"type":67,"value":926},") or a raw AMID (base58, 20+ chars) — display names are resolved via the registry.",{"type":61,"tag":70,"props":928,"children":929},{},[930],{"type":61,"tag":76,"props":931,"children":932},{},[933],{"type":67,"value":250},{"type":61,"tag":125,"props":935,"children":936},{},[937,942,947],{"type":61,"tag":88,"props":938,"children":939},{},[940],{"type":67,"value":941},"Relaying a message to another agent (\"ask the security-auditor to re-check PR #42\").",{"type":61,"tag":88,"props":943,"children":944},{},[945],{"type":67,"value":946},"Notifying an offload sandbox you spawned (e.g., \"stop, I got the answer locally\").",{"type":61,"tag":88,"props":948,"children":949},{},[950],{"type":67,"value":951},"Any peer-to-peer coordination.",{"type":61,"tag":70,"props":953,"children":954},{},[955],{"type":61,"tag":76,"props":956,"children":957},{},[958],{"type":67,"value":959},"Example:",{"type":61,"tag":455,"props":961,"children":964},{"className":962,"code":963,"language":67},[458],"mesh_send(to: \"offload-parent\", message: \"I've finished the initial draft — see files in your inbox\")\n",[965],{"type":61,"tag":219,"props":966,"children":967},{"__ignoreMap":463},[968],{"type":67,"value":963},{"type":61,"tag":109,"props":970,"children":971},{},[],{"type":61,"tag":214,"props":973,"children":975},{"id":974},"mesh_inbox-read-incoming-messages",[976,982],{"type":61,"tag":219,"props":977,"children":979},{"className":978},[],[980],{"type":67,"value":981},"mesh_inbox",{"type":67,"value":983}," — read incoming messages",{"type":61,"tag":70,"props":985,"children":986},{},[987,991,992],{"type":61,"tag":76,"props":988,"children":989},{},[990],{"type":67,"value":234},{"type":67,"value":236},{"type":61,"tag":219,"props":993,"children":995},{"className":994},[],[996],{"type":67,"value":997},"mesh_inbox(limit?: number)",{"type":61,"tag":70,"props":999,"children":1000},{},[1001,1005],{"type":61,"tag":76,"props":1002,"children":1003},{},[1004],{"type":67,"value":250},{"type":67,"value":1006}," The user asks \"any updates?\", \"did X reply?\", or to check on an offload that you started a while ago. Also use this as the primary way to pick up progress\u002Fresults from a cloud offload (offload_status is a cleaner formatted view of the same messages).",{"type":61,"tag":109,"props":1008,"children":1009},{},[],{"type":61,"tag":214,"props":1011,"children":1013},{"id":1012},"discover-list-agents-on-the-mesh",[1014,1020],{"type":61,"tag":219,"props":1015,"children":1017},{"className":1016},[],[1018],{"type":67,"value":1019},"discover",{"type":67,"value":1021}," — list agents on the mesh",{"type":61,"tag":70,"props":1023,"children":1024},{},[1025,1029,1030],{"type":61,"tag":76,"props":1026,"children":1027},{},[1028],{"type":67,"value":234},{"type":67,"value":236},{"type":61,"tag":219,"props":1031,"children":1033},{"className":1032},[],[1034],{"type":67,"value":1035},"discover(capability?: string, limit?: number)",{"type":61,"tag":70,"props":1037,"children":1038},{},[1039,1043,1045,1051,1053,1059,1060,1066,1068,1073],{"type":61,"tag":76,"props":1040,"children":1041},{},[1042],{"type":67,"value":250},{"type":67,"value":1044}," The user asks who is on the mesh, or you want to find an agent with a specific capability (e.g., ",{"type":61,"tag":219,"props":1046,"children":1048},{"className":1047},[],[1049],{"type":67,"value":1050},"\"security-audit\"",{"type":67,"value":1052},", ",{"type":61,"tag":219,"props":1054,"children":1056},{"className":1055},[],[1057],{"type":67,"value":1058},"\"offload\"",{"type":67,"value":1052},{"type":61,"tag":219,"props":1061,"children":1063},{"className":1062},[],[1064],{"type":67,"value":1065},"\"assistant\"",{"type":67,"value":1067},"). Without a capability, ",{"type":61,"tag":219,"props":1069,"children":1071},{"className":1070},[],[1072],{"type":67,"value":1019},{"type":67,"value":1074}," aggregates across well-known seed capabilities (kars-agent, task-execution, cluster-controller, offload, mesh-peer, etc.) and returns a deduped list.",{"type":61,"tag":109,"props":1076,"children":1077},{},[],{"type":61,"tag":113,"props":1079,"children":1081},{"id":1080},"typical-conversation-flows",[1082],{"type":67,"value":1083},"Typical conversation flows",{"type":61,"tag":214,"props":1085,"children":1087},{"id":1086},"pairing-then-offloading",[1088],{"type":67,"value":1089},"Pairing then offloading",{"type":61,"tag":455,"props":1091,"children":1094},{"className":1092,"code":1093,"language":67},[458],"User: here's a pairing token azcp_1_eyJ...\n  → mesh_pair(token: \"azcp_1_eyJ...\")\nUser: great — now research latest trends in AI agent sandboxing\n  → cloud_offload(task: \"research latest trends in AI agent sandboxing ...\")\n  → (brief wait, then) offload_status()\n  → Continue calling offload_status each turn until phase=done\n  → Report summary + files back to the user\n",[1095],{"type":61,"tag":219,"props":1096,"children":1097},{"__ignoreMap":463},[1098],{"type":67,"value":1093},{"type":61,"tag":214,"props":1100,"children":1102},{"id":1101},"checking-on-a-running-offload",[1103],{"type":67,"value":1104},"Checking on a running offload",{"type":61,"tag":455,"props":1106,"children":1109},{"className":1107,"code":1108,"language":67},[458],"User: how's that research going?\n  → offload_status()\n  → Report current phase + elapsed time + last heartbeat\n",[1110],{"type":61,"tag":219,"props":1111,"children":1112},{"__ignoreMap":463},[1113],{"type":67,"value":1108},{"type":61,"tag":214,"props":1115,"children":1117},{"id":1116},"sending-a-peer-message",[1118],{"type":67,"value":1119},"Sending a peer message",{"type":61,"tag":455,"props":1121,"children":1124},{"className":1122,"code":1123,"language":67},[458],"User: tell the cluster sandbox I'm done\n  → mesh_send(to: \"offload-sandbox-name\", message: \"...\")\n",[1125],{"type":61,"tag":219,"props":1126,"children":1127},{"__ignoreMap":463},[1128],{"type":67,"value":1123},{"type":61,"tag":109,"props":1130,"children":1131},{},[],{"type":61,"tag":113,"props":1133,"children":1135},{"id":1134},"important-operational-notes",[1136],{"type":67,"value":1137},"Important operational notes",{"type":61,"tag":125,"props":1139,"children":1140},{},[1141,1158,1189,1205,1222,1232,1242],{"type":61,"tag":88,"props":1142,"children":1143},{},[1144,1156],{"type":61,"tag":76,"props":1145,"children":1146},{},[1147,1149,1154],{"type":67,"value":1148},"Pairing is pre-req for everything except ",{"type":61,"tag":219,"props":1150,"children":1152},{"className":1151},[],[1153],{"type":67,"value":224},{"type":67,"value":1155}," itself.",{"type":67,"value":1157}," If a tool returns \"Not paired\", stop and ask the user for a pairing token.",{"type":61,"tag":88,"props":1159,"children":1160},{},[1161,1166,1168,1173,1175,1180,1182,1187],{"type":61,"tag":76,"props":1162,"children":1163},{},[1164],{"type":67,"value":1165},"Proactive offload:",{"type":67,"value":1167}," Modern kars sandboxes ",{"type":61,"tag":518,"props":1169,"children":1170},{},[1171],{"type":67,"value":1172},"self-announce",{"type":67,"value":1174}," the moment they're ready and begin work immediately. You'll see phase ",{"type":61,"tag":219,"props":1176,"children":1178},{"className":1177},[],[1179],{"type":67,"value":686},{"type":67,"value":1181}," (👋) within ~20 s of ",{"type":61,"tag":219,"props":1183,"children":1185},{"className":1184},[],[1186],{"type":67,"value":664},{"type":67,"value":1188},". If you don't, the cluster may be running an older sandbox image and will fall back to the ping\u002Fupload\u002Fdispatch flow (still works, just slower).",{"type":61,"tag":88,"props":1190,"children":1191},{},[1192,1197,1198,1203],{"type":61,"tag":76,"props":1193,"children":1194},{},[1195],{"type":67,"value":1196},"Tokens are metered.",{"type":67,"value":236},{"type":61,"tag":219,"props":1199,"children":1201},{"className":1200},[],[1202],{"type":67,"value":418},{"type":67,"value":1204}," surfaces budget consumption. If an offload fails with \"budget exceeded\", ask the admin for a new pairing token with a larger budget.",{"type":61,"tag":88,"props":1206,"children":1207},{},[1208,1213,1215,1220],{"type":61,"tag":76,"props":1209,"children":1210},{},[1211],{"type":67,"value":1212},"No secrets in tasks.",{"type":67,"value":1214}," Never put API keys or passwords in ",{"type":61,"tag":219,"props":1216,"children":1218},{"className":1217},[],[1219],{"type":67,"value":298},{"type":67,"value":1221}," task text or files. The sandbox has its own managed identity for Azure access.",{"type":61,"tag":88,"props":1223,"children":1224},{},[1225,1230],{"type":61,"tag":76,"props":1226,"children":1227},{},[1228],{"type":67,"value":1229},"One offload at a time",{"type":67,"value":1231}," per pairing. Cancel (or let it finish) before starting another.",{"type":61,"tag":88,"props":1233,"children":1234},{},[1235,1240],{"type":61,"tag":76,"props":1236,"children":1237},{},[1238],{"type":67,"value":1239},"Files are transferred via E2E-encrypted chunked transfer",{"type":67,"value":1241}," with SHA-256 integrity. Each file ≤ 30 MB.",{"type":61,"tag":88,"props":1243,"children":1244},{},[1245,1250,1252,1258],{"type":61,"tag":76,"props":1246,"children":1247},{},[1248],{"type":67,"value":1249},"Timeouts:",{"type":67,"value":1251}," default 30 min; max configurable via ",{"type":61,"tag":219,"props":1253,"children":1255},{"className":1254},[],[1256],{"type":67,"value":1257},"timeout_minutes",{"type":67,"value":1259},". The sandbox auto-terminates on timeout.",{"type":61,"tag":113,"props":1261,"children":1263},{"id":1262},"common-errors",[1264],{"type":67,"value":1265},"Common errors",{"type":61,"tag":533,"props":1267,"children":1268},{},[1269,1290],{"type":61,"tag":537,"props":1270,"children":1271},{},[1272],{"type":61,"tag":541,"props":1273,"children":1274},{},[1275,1280,1285],{"type":61,"tag":545,"props":1276,"children":1277},{},[1278],{"type":67,"value":1279},"Message",{"type":61,"tag":545,"props":1281,"children":1282},{},[1283],{"type":67,"value":1284},"Cause",{"type":61,"tag":545,"props":1286,"children":1287},{},[1288],{"type":67,"value":1289},"Fix",{"type":61,"tag":561,"props":1291,"children":1292},{},[1293,1321,1350,1372,1404,1431],{"type":61,"tag":541,"props":1294,"children":1295},{},[1296,1305,1310],{"type":61,"tag":568,"props":1297,"children":1298},{},[1299],{"type":61,"tag":219,"props":1300,"children":1302},{"className":1301},[],[1303],{"type":67,"value":1304},"❌ cloud_offload requires a task parameter…",{"type":61,"tag":568,"props":1306,"children":1307},{},[1308],{"type":67,"value":1309},"No recognizable task field passed",{"type":61,"tag":568,"props":1311,"children":1312},{},[1313,1315],{"type":67,"value":1314},"Re-issue with ",{"type":61,"tag":219,"props":1316,"children":1318},{"className":1317},[],[1319],{"type":67,"value":1320},"task: \"...\"",{"type":61,"tag":541,"props":1322,"children":1323},{},[1324,1333,1338],{"type":61,"tag":568,"props":1325,"children":1326},{},[1327],{"type":61,"tag":219,"props":1328,"children":1330},{"className":1329},[],[1331],{"type":67,"value":1332},"❌ Not paired with any kars cluster",{"type":61,"tag":568,"props":1334,"children":1335},{},[1336],{"type":67,"value":1337},"No pairing persisted",{"type":61,"tag":568,"props":1339,"children":1340},{},[1341,1343,1348],{"type":67,"value":1342},"Call ",{"type":61,"tag":219,"props":1344,"children":1346},{"className":1345},[],[1347],{"type":67,"value":224},{"type":67,"value":1349}," with a valid token",{"type":61,"tag":541,"props":1351,"children":1352},{},[1353,1362,1367],{"type":61,"tag":568,"props":1354,"children":1355},{},[1356],{"type":61,"tag":219,"props":1357,"children":1359},{"className":1358},[],[1360],{"type":67,"value":1361},"❌ Mesh connection is not live",{"type":61,"tag":568,"props":1363,"children":1364},{},[1365],{"type":67,"value":1366},"WebSocket to relay dropped",{"type":61,"tag":568,"props":1368,"children":1369},{},[1370],{"type":67,"value":1371},"Wait a few seconds — plugin auto-reconnects; retry",{"type":61,"tag":541,"props":1373,"children":1374},{},[1375,1384,1389],{"type":61,"tag":568,"props":1376,"children":1377},{},[1378],{"type":61,"tag":219,"props":1379,"children":1381},{"className":1380},[],[1382],{"type":67,"value":1383},"❌ Offload already in progress",{"type":61,"tag":568,"props":1385,"children":1386},{},[1387],{"type":67,"value":1388},"Prior offload still running",{"type":61,"tag":568,"props":1390,"children":1391},{},[1392,1397,1399],{"type":61,"tag":219,"props":1393,"children":1395},{"className":1394},[],[1396],{"type":67,"value":418},{"type":67,"value":1398}," to check \u002F wait for ",{"type":61,"tag":219,"props":1400,"children":1402},{"className":1401},[],[1403],{"type":67,"value":803},{"type":61,"tag":541,"props":1405,"children":1406},{},[1407,1416,1421],{"type":61,"tag":568,"props":1408,"children":1409},{},[1410],{"type":61,"tag":219,"props":1411,"children":1413},{"className":1412},[],[1414],{"type":67,"value":1415},"❌ Sandbox '...' is registered but not responding to mesh pings",{"type":61,"tag":568,"props":1417,"children":1418},{},[1419],{"type":67,"value":1420},"Legacy flow, sandbox came up but E2E not ready",{"type":61,"tag":568,"props":1422,"children":1423},{},[1424,1426],{"type":67,"value":1425},"Usually self-heals on retry; modern sandboxes skip this via ",{"type":61,"tag":219,"props":1427,"children":1429},{"className":1428},[],[1430],{"type":67,"value":398},{"type":61,"tag":541,"props":1432,"children":1433},{},[1434,1443,1448],{"type":61,"tag":568,"props":1435,"children":1436},{},[1437],{"type":61,"tag":219,"props":1438,"children":1440},{"className":1439},[],[1441],{"type":67,"value":1442},"No agents found on the mesh.",{"type":61,"tag":568,"props":1444,"children":1445},{},[1446],{"type":67,"value":1447},"No agents registered under queried capability",{"type":61,"tag":568,"props":1449,"children":1450},{},[1451],{"type":67,"value":1452},"Try without filter, or use a different capability name",{"items":1454,"total":1566},[1455,1473,1485,1508,1520,1534,1549],{"slug":1456,"name":1456,"fn":1457,"description":1458,"org":1459,"tags":1460,"stars":24,"repoUrl":25,"updatedAt":1472},"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},[1461,1462,1465,1466,1469],{"name":19,"slug":20,"type":16},{"name":1463,"slug":1464,"type":16},"Audit","audit",{"name":11,"slug":8,"type":16},{"name":1467,"slug":1468,"type":16},"Governance","governance",{"name":1470,"slug":1471,"type":16},"Policy","policy","2026-07-12T08:18:34.446985",{"slug":1474,"name":1474,"fn":1475,"description":1476,"org":1477,"tags":1478,"stars":24,"repoUrl":25,"updatedAt":1484},"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},[1479,1480,1483],{"name":19,"slug":20,"type":16},{"name":1481,"slug":1482,"type":16},"API Development","api-development",{"name":11,"slug":8,"type":16},"2026-07-12T08:18:31.798351",{"slug":1486,"name":1486,"fn":1487,"description":1488,"org":1489,"tags":1490,"stars":24,"repoUrl":25,"updatedAt":1507},"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},[1491,1492,1495,1498,1501,1504],{"name":11,"slug":8,"type":16},{"name":1493,"slug":1494,"type":16},"Charts","charts",{"name":1496,"slug":1497,"type":16},"Data Analysis","data-analysis",{"name":1499,"slug":1500,"type":16},"Data Visualization","data-visualization",{"name":1502,"slug":1503,"type":16},"Mathematics","mathematics",{"name":1505,"slug":1506,"type":16},"Python","python","2026-07-12T08:18:30.4206",{"slug":1509,"name":1509,"fn":1510,"description":1511,"org":1512,"tags":1513,"stars":24,"repoUrl":25,"updatedAt":1519},"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},[1514,1515,1516],{"name":19,"slug":20,"type":16},{"name":11,"slug":8,"type":16},{"name":1517,"slug":1518,"type":16},"Memory","memory","2026-07-12T08:18:38.407887",{"slug":1521,"name":1521,"fn":1522,"description":1523,"org":1524,"tags":1525,"stars":24,"repoUrl":25,"updatedAt":1533},"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},[1526,1527,1530],{"name":11,"slug":8,"type":16},{"name":1528,"slug":1529,"type":16},"Infrastructure","infrastructure",{"name":1531,"slug":1532,"type":16},"LLM","llm","2026-07-12T08:18:39.662409",{"slug":1535,"name":1535,"fn":1536,"description":1537,"org":1538,"tags":1539,"stars":24,"repoUrl":25,"updatedAt":1548},"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},[1540,1541,1544,1545],{"name":11,"slug":8,"type":16},{"name":1542,"slug":1543,"type":16},"Evals","evals",{"name":1531,"slug":1532,"type":16},{"name":1546,"slug":1547,"type":16},"OpenAI","openai","2026-07-12T08:18:40.889909",{"slug":1550,"name":1550,"fn":1551,"description":1552,"org":1553,"tags":1554,"stars":24,"repoUrl":25,"updatedAt":1565},"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},[1555,1556,1559,1562],{"name":11,"slug":8,"type":16},{"name":1557,"slug":1558,"type":16},"Knowledge Management","knowledge-management",{"name":1560,"slug":1561,"type":16},"Research","research",{"name":1563,"slug":1564,"type":16},"Search","search","2026-07-12T08:18:35.894301",11,{"items":1568,"total":1739},[1569,1584,1603,1620,1635,1650,1663,1678,1689,1701,1714,1727],{"slug":1570,"name":1570,"fn":1571,"description":1572,"org":1573,"tags":1574,"stars":1581,"repoUrl":1582,"updatedAt":1583},"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},[1575,1576,1579,1580],{"name":11,"slug":8,"type":16},{"name":1577,"slug":1578,"type":16},"Compliance","compliance",{"name":1467,"slug":1468,"type":16},{"name":1470,"slug":1471,"type":16},1686,"https:\u002F\u002Fgithub.com\u002FAzure\u002Fazure-policy","2026-07-12T08:17:48.378432",{"slug":1585,"name":1585,"fn":1586,"description":1587,"org":1588,"tags":1589,"stars":1600,"repoUrl":1601,"updatedAt":1602},"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},[1590,1591,1594,1597],{"name":11,"slug":8,"type":16},{"name":1592,"slug":1593,"type":16},"Deployment","deployment",{"name":1595,"slug":1596,"type":16},"Infrastructure as Code","infrastructure-as-code",{"name":1598,"slug":1599,"type":16},"Migration","migration",260,"https:\u002F\u002Fgithub.com\u002FAzure\u002Fazure-blueprints","2026-07-12T08:17:49.646405",{"slug":1604,"name":1604,"fn":1605,"description":1606,"org":1607,"tags":1608,"stars":1617,"repoUrl":1618,"updatedAt":1619},"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},[1609,1610,1611,1614],{"name":1481,"slug":1482,"type":16},{"name":11,"slug":8,"type":16},{"name":1612,"slug":1613,"type":16},"Code Review","code-review",{"name":1615,"slug":1616,"type":16},"Documentation","documentation",133,"https:\u002F\u002Fgithub.com\u002FAzure\u002Fazure-sdk-tools","2026-07-12T08:17:43.350876",{"slug":1621,"name":1621,"fn":1622,"description":1623,"org":1624,"tags":1625,"stars":1617,"repoUrl":1618,"updatedAt":1634},"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},[1626,1627,1628,1631],{"name":11,"slug":8,"type":16},{"name":1592,"slug":1593,"type":16},{"name":1629,"slug":1630,"type":16},"SDK","sdk",{"name":1632,"slug":1633,"type":16},"Testing","testing","2026-07-12T08:17:44.718943",{"slug":1636,"name":1636,"fn":1637,"description":1638,"org":1639,"tags":1640,"stars":1617,"repoUrl":1618,"updatedAt":1649},"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},[1641,1642,1645,1648],{"name":11,"slug":8,"type":16},{"name":1643,"slug":1644,"type":16},"GitHub","github",{"name":1646,"slug":1647,"type":16},"Project Management","project-management",{"name":1629,"slug":1630,"type":16},"2026-07-12T08:17:38.345387",{"slug":1651,"name":1651,"fn":1652,"description":1653,"org":1654,"tags":1655,"stars":1617,"repoUrl":1618,"updatedAt":1662},"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},[1656,1657,1660,1661],{"name":11,"slug":8,"type":16},{"name":1658,"slug":1659,"type":16},"CI\u002FCD","ci-cd",{"name":1592,"slug":1593,"type":16},{"name":1629,"slug":1630,"type":16},"2026-07-12T08:17:34.27607",{"slug":1664,"name":1664,"fn":1665,"description":1666,"org":1667,"tags":1668,"stars":1617,"repoUrl":1618,"updatedAt":1677},"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},[1669,1670,1671,1674],{"name":1481,"slug":1482,"type":16},{"name":11,"slug":8,"type":16},{"name":1672,"slug":1673,"type":16},"OpenAPI","openapi",{"name":1675,"slug":1676,"type":16},"Technical Writing","technical-writing","2026-07-12T08:17:39.603232",{"slug":1679,"name":1679,"fn":1680,"description":1681,"org":1682,"tags":1683,"stars":1617,"repoUrl":1618,"updatedAt":1688},"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},[1684,1685,1686,1687],{"name":11,"slug":8,"type":16},{"name":1658,"slug":1659,"type":16},{"name":1629,"slug":1630,"type":16},{"name":1632,"slug":1633,"type":16},"2026-07-12T08:17:37.08523",{"slug":1690,"name":1690,"fn":1691,"description":1692,"org":1693,"tags":1694,"stars":1617,"repoUrl":1618,"updatedAt":1700},"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},[1695,1696,1699],{"name":1531,"slug":1532,"type":16},{"name":1697,"slug":1698,"type":16},"Performance","performance",{"name":1675,"slug":1676,"type":16},"2026-07-12T08:17:42.080413",{"slug":1702,"name":1702,"fn":1703,"description":1704,"org":1705,"tags":1706,"stars":1617,"repoUrl":1618,"updatedAt":1713},"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},[1707,1708,1709,1712],{"name":11,"slug":8,"type":16},{"name":1658,"slug":1659,"type":16},{"name":1710,"slug":1711,"type":16},"Debugging","debugging",{"name":1629,"slug":1630,"type":16},"2026-07-12T08:17:40.821512",{"slug":1715,"name":1715,"fn":1716,"description":1717,"org":1718,"tags":1719,"stars":1617,"repoUrl":1618,"updatedAt":1726},"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},[1720,1721,1722,1725],{"name":11,"slug":8,"type":16},{"name":1577,"slug":1578,"type":16},{"name":1723,"slug":1724,"type":16},"Process Optimization","process-optimization",{"name":1675,"slug":1676,"type":16},"2026-07-12T08:17:32.970921",{"slug":1728,"name":1728,"fn":1729,"description":1730,"org":1731,"tags":1732,"stars":1617,"repoUrl":1618,"updatedAt":1738},"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},[1733,1734,1737],{"name":1615,"slug":1616,"type":16},{"name":1735,"slug":1736,"type":16},"Plugin Development","plugin-development",{"name":1675,"slug":1676,"type":16},"2026-07-12T08:17:35.873862",109]