[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill-nvidia-nemo-relay-get-started":3,"mdc--5e11rm-key":52,"related-repo-nvidia-nemo-relay-get-started":461,"related-org-nvidia-nemo-relay-get-started":563},{"slug":4,"name":4,"fn":5,"description":6,"org":7,"tags":11,"stars":23,"repoUrl":24,"updatedAt":25,"license":26,"forks":27,"topics":28,"repo":47,"sourceUrl":50,"mdContent":51},"nemo-relay-get-started","get started with NeMo Relay","Use this skill when first-time NeMo Relay users want to try Relay, choose the least-complex supported quick start, or verify initial value through the CLI, a maintained integration, or direct Python, Node.js, or Rust instrumentation before production setup.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},"nvidia","NVIDIA","https:\u002F\u002Fpexgzepcugksgbtrxkhf.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Forg-logos\u002Fnvidia.png",[12,16,19,20],{"name":13,"slug":14,"type":15},"Onboarding","onboarding","tag",{"name":17,"slug":18,"type":15},"CLI","cli",{"name":9,"slug":8,"type":15},{"name":21,"slug":22,"type":15},"Python","python",75,"https:\u002F\u002Fgithub.com\u002FNVIDIA\u002FNeMo-Relay","2026-07-23T05:43:39.488827","Apache-2.0",41,[29,30,31,32,33,34,35,36,37,38,39,8,40,41,42,43,44,45,46],"agents","ai","atif","atof","claude-code","codex","deepagents","hermes-agent","langchain","langgraph","middleware","observability","openclaw","openinference","optimization","otel","runtime","security",{"repoUrl":24,"stars":23,"forks":27,"topics":48,"description":49},[29,30,31,32,33,34,35,36,37,38,39,8,40,41,42,43,44,45,46],"Multi-language agent runtime and library for execution scope management, lifecycle events, and middleware on tool and LLM calls.","https:\u002F\u002Fgithub.com\u002FNVIDIA\u002FNeMo-Relay\u002Ftree\u002FHEAD\u002Fskills\u002Fnemo-relay-get-started","---\nname: nemo-relay-get-started\ndescription: Use this skill when first-time NeMo Relay users want to try Relay, choose the least-complex supported quick start, or verify initial value through the CLI, a maintained integration, or direct Python, Node.js, or Rust instrumentation before production setup.\nlicense: Apache-2.0\nmetadata:\n  author: NVIDIA Corporation and Affiliates\n---\n\n# Get Started With NeMo Relay\n\nGuide a new user to visible Relay value with the least complicated applicable\ntrial. Do not begin with production deployment or Relay's full architecture.\nKeep the first run focused on one observable success path.\n\n## Choose A Try-Now Path\n\nEvaluate these paths in order. Use the first one that fits the user's stated\ngoal and existing environment.\n\n1. **CLI try-now (default)**: choose this for a generic \"try Relay\" request or\n   when the user wants value without modifying application code. Run Codex,\n   Claude Code, or Hermes through the local CLI wrapper. Read\n   [CLI Try-Now](references\u002Fcli-try-now.md).\n2. **Built-in integrations try-now**: choose this when an existing LangChain,\n   LangGraph, Deep Agents, or OpenClaw application owns the execution boundary.\n   Prefer the maintained supported integration over manual wrapping. Read\n   [Built-In Integrations Try-Now](references\u002Fbuilt-in-integrations-try-now.md).\n3. **Language-specific manual try-now**: choose this when the user's Python,\n   Node.js, or Rust application directly owns its tool or LLM call sites and no\n   maintained integration is the better boundary. Read\n   [Manual Language Try-Now](references\u002Fmanual-language-try-now.md).\n\nDo not ask the user to choose among all three when their request, manifest, or\nframework already identifies the boundary. For an unspecified request, use the\nCLI path. When more than one CLI agent is available, ask one concise question\nto select the agent.\n\n## Resolve Installation Without Looping\n\nSelect the try-now path before choosing an install package.\n\n- CLI path -> verify `nemo-relay --version`; if missing, use\n  `nemo-relay-install` for the CLI outcome.\n- Built-in integration path -> use `nemo-relay-install` for the named\n  framework or harness package.\n- Manual language path -> use `nemo-relay-install` for the detected language\n  package.\n\nIf installation already succeeded, preserve the chosen path and continue from\nits next step. Do not ask the install-path question again or bounce between the\ninstall and get-started skills.\n\n## Apply The Common First-Value Contract\n\nFollow the selected reference, then:\n\n1. Inspect the target environment and existing Relay configuration before\n   proposing changes.\n2. Explain the attachment boundary and show the exact minimal change.\n3. Obtain confirmation before writing configuration, modifying application\n   code, or launching a model-consuming run.\n4. Use a non-sensitive, read-only trial and exercise one representative tool\n   and LLM path when the selected surface exposes both.\n5. Verify observable evidence from the selected path rather than treating a\n   successful application result as proof that Relay is active.\n6. Summarize the captured root, tool, and model relationships without dumping\n   prompts, credentials, or complete event payloads.\n\nExplain only the concepts visible in the result: the chosen attachment\nboundary, scopes and parentage, captured lifecycle events, and how subscribers\nor the Observability plugin make those events inspectable. Keep instrumentation\nand export distinct.\n\n## Continue With One Plugin\n\nStop the initial try-now workflow when the selected path's success checks pass.\nThen make one additional built-in plugin the primary suggested next step.\n\nExplain Relay's core progression: instrument an execution boundary once, then\nchange or extend behavior through plugin configuration without repeatedly\nrewriting those call sites. Easy plugin configuration and reconfiguration is\nthe main value to demonstrate after the first observable proof.\n\nIf the selected path did not use plugin-managed Observability, add it first to\nestablish the reusable plugin path. If Observability already produced the\nproof, ask what outcome matters next and recommend exactly one plugin:\n\n- Adaptive -> adaptive runtime behavior and optimization\n- NeMo Guardrails -> policy checks around managed execution\n- PII Redaction -> sanitization of sensitive observability payloads\n- Model Pricing -> cost estimates for managed LLM responses\n\nUse the\n[plugin overview](https:\u002F\u002Fdocs.nvidia.com\u002Fnemo\u002Frelay\u002Fdev\u002Fconfigure-plugins\u002Fabout) to\nselect the next component. Preview its smallest configuration, obtain\nconfirmation, and verify its behavior before layering in another plugin.\n\nUse another handoff only after the user accepts or declines plugin progression,\nor when the demonstrated boundary does not yet cover the intended workflow:\n\n- Direct application expansion -> `nemo-relay-instrument-calls`\n- Additional exporters or durable observability configuration ->\n  `nemo-relay-plugin-observability`\n- A different package or supported integration -> `nemo-relay-install`\n- Persistent Claude Code or Codex loading -> `nemo-relay-install`; for Codex\n  Desktop, complete its recovery-note safety gate before changing global config\n- Missing hooks, gateway traffic, or events -> `nemo-relay doctor`,\n  `nemo-relay doctor --json`, or `nemo-relay-debug-runtime-integration`\n\nDo not configure production OTLP backends, model pricing, guardrails, adaptive\ntuning, custom plugins, Go or FFI examples during the quick start. Mention\noptional plugins only after the initial proof and add only the one the user\nselects.\n\n## Public Entry Points\n\nUse these public entry points for current product documentation:\n\n- [CLI overview](https:\u002F\u002Fdocs.nvidia.com\u002Fnemo\u002Frelay\u002Fdev\u002Fnemo-relay-cli\u002Fabout)\n- [Maintained integrations](https:\u002F\u002Fdocs.nvidia.com\u002Fnemo\u002Frelay\u002Fdev\u002Fsupported-integrations\u002Fabout)\n- [Language quick starts](https:\u002F\u002Fdocs.nvidia.com\u002Fnemo\u002Frelay\u002Fdev\u002Fgetting-started\u002Fquick-start)\n- [Plugin selection](https:\u002F\u002Fdocs.nvidia.com\u002Fnemo\u002Frelay\u002Fdev\u002Fconfigure-plugins\u002Fabout)\n",{"data":53,"body":56},{"name":4,"description":6,"license":26,"metadata":54},{"author":55},"NVIDIA Corporation and Affiliates",{"type":57,"children":58},"root",[59,68,74,81,86,145,150,156,161,211,216,222,227,260,265,271,276,281,286,309,324,329,403,408,414,419],{"type":60,"tag":61,"props":62,"children":64},"element","h1",{"id":63},"get-started-with-nemo-relay",[65],{"type":66,"value":67},"text","Get Started With NeMo Relay",{"type":60,"tag":69,"props":70,"children":71},"p",{},[72],{"type":66,"value":73},"Guide a new user to visible Relay value with the least complicated applicable\ntrial. Do not begin with production deployment or Relay's full architecture.\nKeep the first run focused on one observable success path.",{"type":60,"tag":75,"props":76,"children":78},"h2",{"id":77},"choose-a-try-now-path",[79],{"type":66,"value":80},"Choose A Try-Now Path",{"type":60,"tag":69,"props":82,"children":83},{},[84],{"type":66,"value":85},"Evaluate these paths in order. Use the first one that fits the user's stated\ngoal and existing environment.",{"type":60,"tag":87,"props":88,"children":89},"ol",{},[90,111,128],{"type":60,"tag":91,"props":92,"children":93},"li",{},[94,100,102,109],{"type":60,"tag":95,"props":96,"children":97},"strong",{},[98],{"type":66,"value":99},"CLI try-now (default)",{"type":66,"value":101},": choose this for a generic \"try Relay\" request or\nwhen the user wants value without modifying application code. Run Codex,\nClaude Code, or Hermes through the local CLI wrapper. Read\n",{"type":60,"tag":103,"props":104,"children":106},"a",{"href":105},"references\u002Fcli-try-now.md",[107],{"type":66,"value":108},"CLI Try-Now",{"type":66,"value":110},".",{"type":60,"tag":91,"props":112,"children":113},{},[114,119,121,127],{"type":60,"tag":95,"props":115,"children":116},{},[117],{"type":66,"value":118},"Built-in integrations try-now",{"type":66,"value":120},": choose this when an existing LangChain,\nLangGraph, Deep Agents, or OpenClaw application owns the execution boundary.\nPrefer the maintained supported integration over manual wrapping. Read\n",{"type":60,"tag":103,"props":122,"children":124},{"href":123},"references\u002Fbuilt-in-integrations-try-now.md",[125],{"type":66,"value":126},"Built-In Integrations Try-Now",{"type":66,"value":110},{"type":60,"tag":91,"props":129,"children":130},{},[131,136,138,144],{"type":60,"tag":95,"props":132,"children":133},{},[134],{"type":66,"value":135},"Language-specific manual try-now",{"type":66,"value":137},": choose this when the user's Python,\nNode.js, or Rust application directly owns its tool or LLM call sites and no\nmaintained integration is the better boundary. Read\n",{"type":60,"tag":103,"props":139,"children":141},{"href":140},"references\u002Fmanual-language-try-now.md",[142],{"type":66,"value":143},"Manual Language Try-Now",{"type":66,"value":110},{"type":60,"tag":69,"props":146,"children":147},{},[148],{"type":66,"value":149},"Do not ask the user to choose among all three when their request, manifest, or\nframework already identifies the boundary. For an unspecified request, use the\nCLI path. When more than one CLI agent is available, ask one concise question\nto select the agent.",{"type":60,"tag":75,"props":151,"children":153},{"id":152},"resolve-installation-without-looping",[154],{"type":66,"value":155},"Resolve Installation Without Looping",{"type":60,"tag":69,"props":157,"children":158},{},[159],{"type":66,"value":160},"Select the try-now path before choosing an install package.",{"type":60,"tag":162,"props":163,"children":164},"ul",{},[165,187,199],{"type":60,"tag":91,"props":166,"children":167},{},[168,170,177,179,185],{"type":66,"value":169},"CLI path -> verify ",{"type":60,"tag":171,"props":172,"children":174},"code",{"className":173},[],[175],{"type":66,"value":176},"nemo-relay --version",{"type":66,"value":178},"; if missing, use\n",{"type":60,"tag":171,"props":180,"children":182},{"className":181},[],[183],{"type":66,"value":184},"nemo-relay-install",{"type":66,"value":186}," for the CLI outcome.",{"type":60,"tag":91,"props":188,"children":189},{},[190,192,197],{"type":66,"value":191},"Built-in integration path -> use ",{"type":60,"tag":171,"props":193,"children":195},{"className":194},[],[196],{"type":66,"value":184},{"type":66,"value":198}," for the named\nframework or harness package.",{"type":60,"tag":91,"props":200,"children":201},{},[202,204,209],{"type":66,"value":203},"Manual language path -> use ",{"type":60,"tag":171,"props":205,"children":207},{"className":206},[],[208],{"type":66,"value":184},{"type":66,"value":210}," for the detected language\npackage.",{"type":60,"tag":69,"props":212,"children":213},{},[214],{"type":66,"value":215},"If installation already succeeded, preserve the chosen path and continue from\nits next step. Do not ask the install-path question again or bounce between the\ninstall and get-started skills.",{"type":60,"tag":75,"props":217,"children":219},{"id":218},"apply-the-common-first-value-contract",[220],{"type":66,"value":221},"Apply The Common First-Value Contract",{"type":60,"tag":69,"props":223,"children":224},{},[225],{"type":66,"value":226},"Follow the selected reference, then:",{"type":60,"tag":87,"props":228,"children":229},{},[230,235,240,245,250,255],{"type":60,"tag":91,"props":231,"children":232},{},[233],{"type":66,"value":234},"Inspect the target environment and existing Relay configuration before\nproposing changes.",{"type":60,"tag":91,"props":236,"children":237},{},[238],{"type":66,"value":239},"Explain the attachment boundary and show the exact minimal change.",{"type":60,"tag":91,"props":241,"children":242},{},[243],{"type":66,"value":244},"Obtain confirmation before writing configuration, modifying application\ncode, or launching a model-consuming run.",{"type":60,"tag":91,"props":246,"children":247},{},[248],{"type":66,"value":249},"Use a non-sensitive, read-only trial and exercise one representative tool\nand LLM path when the selected surface exposes both.",{"type":60,"tag":91,"props":251,"children":252},{},[253],{"type":66,"value":254},"Verify observable evidence from the selected path rather than treating a\nsuccessful application result as proof that Relay is active.",{"type":60,"tag":91,"props":256,"children":257},{},[258],{"type":66,"value":259},"Summarize the captured root, tool, and model relationships without dumping\nprompts, credentials, or complete event payloads.",{"type":60,"tag":69,"props":261,"children":262},{},[263],{"type":66,"value":264},"Explain only the concepts visible in the result: the chosen attachment\nboundary, scopes and parentage, captured lifecycle events, and how subscribers\nor the Observability plugin make those events inspectable. Keep instrumentation\nand export distinct.",{"type":60,"tag":75,"props":266,"children":268},{"id":267},"continue-with-one-plugin",[269],{"type":66,"value":270},"Continue With One Plugin",{"type":60,"tag":69,"props":272,"children":273},{},[274],{"type":66,"value":275},"Stop the initial try-now workflow when the selected path's success checks pass.\nThen make one additional built-in plugin the primary suggested next step.",{"type":60,"tag":69,"props":277,"children":278},{},[279],{"type":66,"value":280},"Explain Relay's core progression: instrument an execution boundary once, then\nchange or extend behavior through plugin configuration without repeatedly\nrewriting those call sites. Easy plugin configuration and reconfiguration is\nthe main value to demonstrate after the first observable proof.",{"type":60,"tag":69,"props":282,"children":283},{},[284],{"type":66,"value":285},"If the selected path did not use plugin-managed Observability, add it first to\nestablish the reusable plugin path. If Observability already produced the\nproof, ask what outcome matters next and recommend exactly one plugin:",{"type":60,"tag":162,"props":287,"children":288},{},[289,294,299,304],{"type":60,"tag":91,"props":290,"children":291},{},[292],{"type":66,"value":293},"Adaptive -> adaptive runtime behavior and optimization",{"type":60,"tag":91,"props":295,"children":296},{},[297],{"type":66,"value":298},"NeMo Guardrails -> policy checks around managed execution",{"type":60,"tag":91,"props":300,"children":301},{},[302],{"type":66,"value":303},"PII Redaction -> sanitization of sensitive observability payloads",{"type":60,"tag":91,"props":305,"children":306},{},[307],{"type":66,"value":308},"Model Pricing -> cost estimates for managed LLM responses",{"type":60,"tag":69,"props":310,"children":311},{},[312,314,322],{"type":66,"value":313},"Use the\n",{"type":60,"tag":103,"props":315,"children":319},{"href":316,"rel":317},"https:\u002F\u002Fdocs.nvidia.com\u002Fnemo\u002Frelay\u002Fdev\u002Fconfigure-plugins\u002Fabout",[318],"nofollow",[320],{"type":66,"value":321},"plugin overview",{"type":66,"value":323}," to\nselect the next component. Preview its smallest configuration, obtain\nconfirmation, and verify its behavior before layering in another plugin.",{"type":60,"tag":69,"props":325,"children":326},{},[327],{"type":66,"value":328},"Use another handoff only after the user accepts or declines plugin progression,\nor when the demonstrated boundary does not yet cover the intended workflow:",{"type":60,"tag":162,"props":330,"children":331},{},[332,343,354,364,376],{"type":60,"tag":91,"props":333,"children":334},{},[335,337],{"type":66,"value":336},"Direct application expansion -> ",{"type":60,"tag":171,"props":338,"children":340},{"className":339},[],[341],{"type":66,"value":342},"nemo-relay-instrument-calls",{"type":60,"tag":91,"props":344,"children":345},{},[346,348],{"type":66,"value":347},"Additional exporters or durable observability configuration ->\n",{"type":60,"tag":171,"props":349,"children":351},{"className":350},[],[352],{"type":66,"value":353},"nemo-relay-plugin-observability",{"type":60,"tag":91,"props":355,"children":356},{},[357,359],{"type":66,"value":358},"A different package or supported integration -> ",{"type":60,"tag":171,"props":360,"children":362},{"className":361},[],[363],{"type":66,"value":184},{"type":60,"tag":91,"props":365,"children":366},{},[367,369,374],{"type":66,"value":368},"Persistent Claude Code or Codex loading -> ",{"type":60,"tag":171,"props":370,"children":372},{"className":371},[],[373],{"type":66,"value":184},{"type":66,"value":375},"; for Codex\nDesktop, complete its recovery-note safety gate before changing global config",{"type":60,"tag":91,"props":377,"children":378},{},[379,381,387,389,395,397],{"type":66,"value":380},"Missing hooks, gateway traffic, or events -> ",{"type":60,"tag":171,"props":382,"children":384},{"className":383},[],[385],{"type":66,"value":386},"nemo-relay doctor",{"type":66,"value":388},",\n",{"type":60,"tag":171,"props":390,"children":392},{"className":391},[],[393],{"type":66,"value":394},"nemo-relay doctor --json",{"type":66,"value":396},", or ",{"type":60,"tag":171,"props":398,"children":400},{"className":399},[],[401],{"type":66,"value":402},"nemo-relay-debug-runtime-integration",{"type":60,"tag":69,"props":404,"children":405},{},[406],{"type":66,"value":407},"Do not configure production OTLP backends, model pricing, guardrails, adaptive\ntuning, custom plugins, Go or FFI examples during the quick start. Mention\noptional plugins only after the initial proof and add only the one the user\nselects.",{"type":60,"tag":75,"props":409,"children":411},{"id":410},"public-entry-points",[412],{"type":66,"value":413},"Public Entry Points",{"type":60,"tag":69,"props":415,"children":416},{},[417],{"type":66,"value":418},"Use these public entry points for current product documentation:",{"type":60,"tag":162,"props":420,"children":421},{},[422,432,442,452],{"type":60,"tag":91,"props":423,"children":424},{},[425],{"type":60,"tag":103,"props":426,"children":429},{"href":427,"rel":428},"https:\u002F\u002Fdocs.nvidia.com\u002Fnemo\u002Frelay\u002Fdev\u002Fnemo-relay-cli\u002Fabout",[318],[430],{"type":66,"value":431},"CLI overview",{"type":60,"tag":91,"props":433,"children":434},{},[435],{"type":60,"tag":103,"props":436,"children":439},{"href":437,"rel":438},"https:\u002F\u002Fdocs.nvidia.com\u002Fnemo\u002Frelay\u002Fdev\u002Fsupported-integrations\u002Fabout",[318],[440],{"type":66,"value":441},"Maintained integrations",{"type":60,"tag":91,"props":443,"children":444},{},[445],{"type":60,"tag":103,"props":446,"children":449},{"href":447,"rel":448},"https:\u002F\u002Fdocs.nvidia.com\u002Fnemo\u002Frelay\u002Fdev\u002Fgetting-started\u002Fquick-start",[318],[450],{"type":66,"value":451},"Language quick starts",{"type":60,"tag":91,"props":453,"children":454},{},[455],{"type":60,"tag":103,"props":456,"children":458},{"href":316,"rel":457},[318],[459],{"type":66,"value":460},"Plugin selection",{"items":462,"total":562},[463,478,485,503,517,530,545],{"slug":402,"name":402,"fn":464,"description":465,"org":466,"tags":467,"stars":23,"repoUrl":24,"updatedAt":477},"debug NeMo Relay runtime integrations","Use this skill when NeMo Relay is installed or imported but application-side runtime behavior is missing or incorrect, including load failures, inactive scopes, missing events, and plugin or adaptive wiring problems.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[468,471,474,476],{"name":469,"slug":470,"type":15},"Debugging","debugging",{"name":472,"slug":473,"type":15},"Engineering","engineering",{"name":475,"slug":39,"type":15},"Middleware",{"name":9,"slug":8,"type":15},"2026-07-17T05:30:27.286964",{"slug":4,"name":4,"fn":5,"description":6,"org":479,"tags":480,"stars":23,"repoUrl":24,"updatedAt":25},{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[481,482,483,484],{"name":17,"slug":18,"type":15},{"name":9,"slug":8,"type":15},{"name":13,"slug":14,"type":15},{"name":21,"slug":22,"type":15},{"slug":184,"name":184,"fn":486,"description":487,"org":488,"tags":489,"stars":23,"repoUrl":24,"updatedAt":502},"install NeMo Relay for various environments","Use this skill when choosing or running NeMo Relay installation for the CLI, Python, Node.js, Rust, OpenClaw, Hermes, or maintained framework integrations before runtime configuration or quick-start setup.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[490,491,494,497,498,499],{"name":17,"slug":18,"type":15},{"name":492,"slug":493,"type":15},"Deployment","deployment",{"name":495,"slug":496,"type":15},"Node.js","nodejs",{"name":9,"slug":8,"type":15},{"name":21,"slug":22,"type":15},{"name":500,"slug":501,"type":15},"Rust","rust","2026-07-17T05:30:31.443244",{"slug":342,"name":342,"fn":504,"description":505,"org":506,"tags":507,"stars":23,"repoUrl":24,"updatedAt":516},"instrument LLM calls with NeMo Relay","Use this skill when an application owns tool or LLM\u002Fprovider call sites and needs to wrap them with NeMo Relay scopes and managed execution APIs for lifecycle events, middleware, or guardrails.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[508,511,514,515],{"name":509,"slug":510,"type":15},"Automation","automation",{"name":512,"slug":513,"type":15},"LLM","llm",{"name":475,"slug":39,"type":15},{"name":9,"slug":8,"type":15},"2026-07-17T05:30:29.31214",{"slug":518,"name":518,"fn":519,"description":520,"org":521,"tags":522,"stars":23,"repoUrl":24,"updatedAt":529},"nemo-relay-instrument-context-isolation","isolate NeMo Relay execution contexts","Use this skill when concurrent requests, async tasks, threads, workers, goroutines, or agents need independent NeMo Relay scope stacks and correct ancestry propagation.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[523,526,527,528],{"name":524,"slug":525,"type":15},"Concurrency","concurrency",{"name":472,"slug":473,"type":15},{"name":475,"slug":39,"type":15},{"name":9,"slug":8,"type":15},"2026-07-17T05:30:33.080543",{"slug":531,"name":531,"fn":532,"description":533,"org":534,"tags":535,"stars":23,"repoUrl":24,"updatedAt":544},"nemo-relay-instrument-typed-wrappers","instrument NeMo Relay typed wrappers","Use this skill when adding NeMo Relay typed wrappers, domain types, or provider codecs while preserving JSON middleware semantics and caller-visible behavior.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[536,539,542,543],{"name":537,"slug":538,"type":15},"API Development","api-development",{"name":540,"slug":541,"type":15},"JSON","json",{"name":475,"slug":39,"type":15},{"name":9,"slug":8,"type":15},"2026-07-23T05:43:40.456621",{"slug":546,"name":546,"fn":547,"description":548,"org":549,"tags":550,"stars":23,"repoUrl":24,"updatedAt":561},"nemo-relay-migrate-from-flow","migrate applications from NeMo Flow to Relay","Use this skill when migrating applications, examples, integrations, documentation, manifests, or repository code from NeMo Flow to NeMo Relay across Python, Rust, Node.js, Go, C FFI, CLI, configuration, and observability surfaces.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[551,554,557,558,559,560],{"name":552,"slug":553,"type":15},"Go","go",{"name":555,"slug":556,"type":15},"Migration","migration",{"name":495,"slug":496,"type":15},{"name":9,"slug":8,"type":15},{"name":21,"slug":22,"type":15},{"name":500,"slug":501,"type":15},"2026-07-17T05:30:28.259456",10,{"items":564,"total":713},[565,583,597,608,620,632,645,657,670,681,695,704],{"slug":566,"name":566,"fn":567,"description":568,"org":569,"tags":570,"stars":580,"repoUrl":581,"updatedAt":582},"nemoclaw-user-guide","retrieve NemoClaw documentation and configuration","Guides human users' AI agents to the NemoClaw docs MCP server and canonical Fern documentation in Markdown form. Use when users ask how to install, configure, operate, troubleshoot, secure, or learn NemoClaw with an AI coding assistant. Trigger keywords - nemoclaw docs, use nemoclaw with ai agent, nemoclaw mcp docs, nemoclaw install help, nemoclaw quickstart, nemoclaw markdown docs, llms.txt, agent skills.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[571,574,577],{"name":572,"slug":573,"type":15},"Documentation","documentation",{"name":575,"slug":576,"type":15},"MCP","mcp",{"name":578,"slug":579,"type":15},"Search","search",21777,"https:\u002F\u002Fgithub.com\u002FNVIDIA\u002FNemoClaw","2026-07-20T06:00:01.461044",{"slug":584,"name":584,"fn":585,"description":586,"org":587,"tags":588,"stars":594,"repoUrl":595,"updatedAt":596},"mcore-build-and-dependency","manage Megatron-LM development environments","Container-based dev environment setup and dependency management for Megatron-LM. Covers acquiring and launching the CI container, uv package management, and updating uv.lock.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[589,592,593],{"name":590,"slug":591,"type":15},"Containers","containers",{"name":492,"slug":493,"type":15},{"name":21,"slug":22,"type":15},17049,"https:\u002F\u002Fgithub.com\u002FNVIDIA\u002FMegatron-LM","2026-07-27T06:06:11.249662",{"slug":598,"name":598,"fn":599,"description":600,"org":601,"tags":602,"stars":594,"repoUrl":595,"updatedAt":607},"mcore-bump-base-image","update NVIDIA PyTorch base images","Bump the NVIDIA PyTorch base image (`nvcr.io\u002Fnvidia\u002Fpytorch:YY.MM-py3`) used by Megatron-LM CI. Covers the two pin sites (GitHub CI in `docker\u002F.ngc_version.dev` and GitLab CI in `.gitlab\u002Fstages\u002F01.build.yml`), the post-bump CI loop (re-run functional tests, refresh golden values, mark broken tests), and the gotchas that bit PRs",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[603,606],{"name":604,"slug":605,"type":15},"CI\u002FCD","ci-cd",{"name":492,"slug":493,"type":15},"2026-07-14T05:25:59.97109",{"slug":609,"name":609,"fn":610,"description":611,"org":612,"tags":613,"stars":594,"repoUrl":595,"updatedAt":619},"mcore-cicd","manage CI\u002FCD pipelines for Megatron-LM","CI\u002FCD reference for Megatron-LM. Covers CI pipeline structure, PR scope labels, triggering internal GitLab CI (which force-pushes the current branch to a pull-request\u002FBRANCH ref — always dry-run and verify the destination first; never run against shared or protected branches), and CI failure investigation.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[614,615,616],{"name":604,"slug":605,"type":15},{"name":492,"slug":493,"type":15},{"name":617,"slug":618,"type":15},"GitHub","github","2026-07-27T06:06:12.278222",{"slug":621,"name":621,"fn":622,"description":623,"org":624,"tags":625,"stars":594,"repoUrl":595,"updatedAt":631},"mcore-create-issue","investigate CI failures and create issues","Investigate a failing GitHub Actions run or job and create a GitHub issue for the failure.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[626,627,628],{"name":469,"slug":470,"type":15},{"name":617,"slug":618,"type":15},{"name":629,"slug":630,"type":15},"Triage","triage","2026-07-14T05:25:57.442089",{"slug":633,"name":633,"fn":634,"description":635,"org":636,"tags":637,"stars":594,"repoUrl":595,"updatedAt":644},"mcore-linting-and-formatting","lint and format Megatron-LM code","Linting and formatting for Megatron-LM. Covers running autoformat.sh, tools (ruff, black, isort, pylint, mypy), and code style rules.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[638,641],{"name":639,"slug":640,"type":15},"Best Practices","best-practices",{"name":642,"slug":643,"type":15},"Code Analysis","code-analysis","2026-07-14T05:25:56.18433",{"slug":646,"name":646,"fn":647,"description":648,"org":649,"tags":650,"stars":594,"repoUrl":595,"updatedAt":656},"mcore-migrate-gpt-to-hybrid","migrate Megatron-LM models to HybridModel","Migration guide for moving Megatron Core GPTModel checkpoints, model providers, training commands, and layer mappings to HybridModel.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[651,654,655],{"name":652,"slug":653,"type":15},"Machine Learning","machine-learning",{"name":555,"slug":556,"type":15},{"name":9,"slug":8,"type":15},"2026-07-17T06:07:11.777011",{"slug":658,"name":658,"fn":659,"description":660,"org":661,"tags":662,"stars":594,"repoUrl":595,"updatedAt":669},"mcore-onboard-gb200-1node-tests","onboard functional tests for GB200","Onboard 1-node GitHub MR functional tests for GB200 from existing mr-scoped 2-node tests.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[663,666],{"name":664,"slug":665,"type":15},"QA","qa",{"name":667,"slug":668,"type":15},"Testing","testing","2026-07-14T05:25:53.673039",{"slug":671,"name":671,"fn":672,"description":673,"org":674,"tags":675,"stars":594,"repoUrl":595,"updatedAt":680},"mcore-run-on-slurm","launch distributed training jobs on SLURM","How to launch distributed Megatron-LM training jobs on a SLURM cluster. Covers a minimal sbatch skeleton, environment-variable setup for torch.distributed.run, CUDA_DEVICE_MAX_CONNECTIONS rules across hardware and parallelism modes, container conventions, monitoring, and per-rank failure diagnosis.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[676,677],{"name":492,"slug":493,"type":15},{"name":678,"slug":679,"type":15},"Infrastructure","infrastructure","2026-07-14T05:25:49.362534",{"slug":682,"name":682,"fn":683,"description":684,"org":685,"tags":686,"stars":594,"repoUrl":595,"updatedAt":694},"mcore-split-pr","split pull requests to reduce review load","Split a PR into multiple PRs to reduce the number of required CODEOWNERS reviewer groups.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[687,690,691],{"name":688,"slug":689,"type":15},"Code Review","code-review",{"name":617,"slug":618,"type":15},{"name":692,"slug":693,"type":15},"Pull Requests","pull-requests","2026-07-14T05:26:01.226578",{"slug":696,"name":696,"fn":697,"description":698,"org":699,"tags":700,"stars":594,"repoUrl":595,"updatedAt":703},"mcore-testing","run and manage Megatron-LM tests","Test system for Megatron-LM. Covers test layout, recipe YAML structure, adding and running unit and functional tests, golden values, marker filters, and CI parity.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[701,702],{"name":664,"slug":665,"type":15},{"name":667,"slug":668,"type":15},"2026-07-14T05:25:54.928983",{"slug":705,"name":705,"fn":706,"description":707,"org":708,"tags":709,"stars":594,"repoUrl":595,"updatedAt":712},"nightly-sync","manage nightly main-to-dev sync workflows","Domain knowledge for the nightly main-to-dev sync workflow. Covers merge strategy, CI architecture, failure investigation, and known issues.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[710,711],{"name":509,"slug":510,"type":15},{"name":604,"slug":605,"type":15},"2026-07-30T05:29:03.275638",496]