[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill-nvidia-doca-debug":3,"mdc--lf906u-key":31,"related-org-nvidia-doca-debug":1060,"related-repo-nvidia-doca-debug":1219},{"slug":4,"name":4,"fn":5,"description":6,"org":7,"tags":11,"stars":20,"repoUrl":21,"updatedAt":22,"license":23,"forks":24,"topics":25,"repo":26,"sourceUrl":29,"mdContent":30},"doca-debug","debug NVIDIA DOCA applications","Use this skill when the user is debugging any DOCA symptom — a build that won't compile, a link step that can't resolve a doca_* symbol, a runtime call returning DOCA_ERROR_*, a silent service or tool, or a stack trace \u002F valgrind \u002F core dump — and needs the layered ladder (install → version → build → link → runtime → program → driver), verbosity controls (--sdk-log-level, DOCA_LOG_LEVEL, the doca-{lib}-trace flavor), container-debug constraints, or how to capture state for a Developer Forum post. Trigger even when the user does not say \"DOCA debug\" — implicit phrasings include \"undefined reference to doca_*\", \"how do I get more logs\", \"packets aren't reaching the wire\", \"doca_caps returned nothing\", or \"hugepages empty in the container\". Refuse and route elsewhere for library-specific debug (Flow pipe trace, RDMA QP, Comch stats), env-class pkg-config or hugepages symptoms, the DOCA_ERROR_* taxonomy and lifecycle interpretation, and performance or incident-response work — those belong to other skills.\n",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},"nvidia","NVIDIA","https:\u002F\u002Fpexgzepcugksgbtrxkhf.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Forg-logos\u002Fnvidia.png",[12,14,17],{"name":9,"slug":8,"type":13},"tag",{"name":15,"slug":16,"type":13},"Engineering","engineering",{"name":18,"slug":19,"type":13},"Debugging","debugging",2473,"https:\u002F\u002Fgithub.com\u002FNVIDIA\u002Fskills","2026-07-20T06:24:45.979363","Apache-2.0",281,[],{"repoUrl":21,"stars":20,"forks":24,"topics":27,"description":28},[],"AI agent skills published by NVIDIA","https:\u002F\u002Fgithub.com\u002FNVIDIA\u002Fskills\u002Ftree\u002FHEAD\u002Fskills\u002Fdoca-debug","---\nlicense: Apache-2.0\nname: doca-debug\ndescription: >\n  Use this skill when the user is debugging any DOCA symptom — a build\n  that won't compile, a link step that can't resolve a doca_* symbol,\n  a runtime call returning DOCA_ERROR_*, a silent service or tool, or\n  a stack trace \u002F valgrind \u002F core dump — and needs the layered ladder\n  (install → version → build → link → runtime → program → driver),\n  verbosity controls (--sdk-log-level, DOCA_LOG_LEVEL, the\n  doca-{lib}-trace flavor), container-debug constraints, or how to\n  capture state for a Developer Forum post. Trigger even when the user\n  does not say \"DOCA debug\" — implicit phrasings include \"undefined\n  reference to doca_*\", \"how do I get more logs\", \"packets aren't\n  reaching the wire\", \"doca_caps returned nothing\", or \"hugepages\n  empty in the container\". Refuse and route elsewhere for\n  library-specific debug (Flow pipe trace, RDMA QP, Comch stats),\n  env-class pkg-config or hugepages symptoms, the DOCA_ERROR_*\n  taxonomy and lifecycle interpretation, and performance or\n  incident-response work — those belong to other skills.\nmetadata:\n  kind: library\ncompatibility: >\n  No DOCA install required to read this skill (it is an overlay loaded\n  against any DOCA artifact skill); the validation steps within DO\n  require a live DOCA install at \u002Fopt\u002Fmellanox\u002Fdoca.\n---\n\n# DOCA debug\n\n**Where to start:** This skill is the canonical layered-ladder\nreference both [`doca-setup ## debug`](..\u002Fdoca-setup\u002FTASKS.md#debug)\nand [`doca-programming-guide ## debug`](..\u002Fdoca-programming-guide\u002FTASKS.md#debug)\nescalate to. If the symptom does not fit cleanly in env-class or\nprogram-class, start at [`TASKS.md ## debug`](TASKS.md#debug) — the\nfull layered ladder lives there.\n\n## Example questions this skill answers well\n\nThe CLASSES of debug questions this skill is built to answer, each\nwith one worked example.\n\n- **\"My DOCA build \u002F link \u002F runtime \u002F program failed — which layer is\n  it?\"** — worked example: *\"`ld` says `undefined reference to\n  doca_flow_init` — which layer?\"* Answered by the canonical layered\n  ladder in [`TASKS.md ## debug`](TASKS.md#debug) (layer 4 = Link).\n- **\"How do I turn up verbosity for any DOCA library or tool?\"** —\n  worked example: *\"My DOCA Flow program is silent — where do I get\n  more log output?\"* Answered by the trace-flavor \u002F log-level surface\n  in [`CAPABILITIES.md ## Observability`](CAPABILITIES.md#observability)\n  + the run-with-verbosity workflow in\n  [`TASKS.md ## run`](TASKS.md#run).\n- **\"How do I capture state for a forum question or bug report?\"** —\n  worked example: *\"I'd like to file a forum question with reproducible\n  context — what should I include?\"* Answered by the\n  capture-a-reproducible-state workflow in\n  [`TASKS.md ## test`](TASKS.md#test).\n- **\"What does this DOCA tool's output mean \u002F which tool answers\n  this?\"** — worked example: *\"`doca_caps` returned nothing for RDMA\n  — does that mean unsupported?\"* Answered by the\n  tool-vs-capability decision tree in\n  [`CAPABILITIES.md ## Capabilities and modes`](CAPABILITIES.md#capabilities-and-modes)\n  + cross-link to [doca-caps](..\u002Ftools\u002Fdoca-caps\u002FSKILL.md).\n- **\"I'm debugging inside the NGC container — what's\n  observable?\"** — worked example: *\"`hugepages` is empty inside the\n  container; is that a real problem or a container thing?\"*\n  Answered by the container-specific debug constraints in\n  [`CAPABILITIES.md ## Safety policy`](CAPABILITIES.md#safety-policy)\n  + [`TASKS.md ## debug`](TASKS.md#debug) layer 5 (Runtime).\n- **\"Where do I ask for help with this?\"** — worked example: *\"Where\n  is the customer-facing DOCA forum and what should the post\n  contain?\"* Answered by the Developer-Forum routing rule in\n  [`TASKS.md ## debug`](TASKS.md#debug) plus\n  [doca-public-knowledge-map](..\u002Fdoca-public-knowledge-map\u002FSKILL.md).\n\nIf the symptom is purely env-class, route to\n[`doca-setup ## debug`](..\u002Fdoca-setup\u002FTASKS.md#debug); if purely\nprogram-class, route to\n[`doca-programming-guide ## debug`](..\u002Fdoca-programming-guide\u002FTASKS.md#debug).\nThis skill is the cross-cutting layer both call into.\n\n## When to load this skill\n\nLoad this skill when the user is debugging anything DOCA-related — a build that won't compile, a link step that can't resolve a `doca_*` symbol, a runtime call that returns `DOCA_ERROR_*`, a packet that does not appear on the wire, a service that won't start, or a tool that returns no useful output. Concretely:\n\n- The user reports a symptom and needs to find the layer that caused it (install \u002F version \u002F build \u002F link \u002F runtime \u002F program).\n- The user asks \"how do I get more logs?\" or \"how do I turn up the verbosity?\" for any DOCA library or tool.\n- The user wants to capture state for a forum question or an internal bug report (the bundle does not own the internal-bug-report channel; it routes to the public DOCA Developer Forum).\n- The user is reading a stack trace, a `valgrind` output, or a core dump from a DOCA program and wants to know where to look first.\n- The user is debugging *inside* the NGC DOCA container and needs to know what is and is not observable from inside it.\n\nDo **not** load this skill for:\n\n- *\"What is `DOCA_ERROR_BAD_STATE`?\", \"what error codes does DOCA return?\"* — that is the cross-library *error taxonomy*, owned by [`doca-programming-guide CAPABILITIES.md ## Error taxonomy`](..\u002Fdoca-programming-guide\u002FCAPABILITIES.md#error-taxonomy). This skill consumes that taxonomy; it does not redefine it.\n- *\"My `pkg-config` cannot find `doca-flow`\", \"hugepages are not mounted\", \"my representor isn't visible\"* — those are env-class symptoms, owned by [`doca-setup ## debug`](..\u002Fdoca-setup\u002FTASKS.md#debug). This skill is the canonical pointer that env-class debug ladder redirects to once the symptom escalates beyond install \u002F version \u002F build prerequisites.\n- *Library-specific debugging* (Flow pipe trace, RDMA queue-pair state, Comch channel statistics) — those live in the matching library skill (e.g. [`doca-flow ## debug`](..\u002Flibs\u002Fdoca-flow\u002FTASKS.md#debug)). This skill provides the cross-cutting debug ladder; library skills layer their library-specific debug surface on top of it.\n\n## What this skill provides\n\nThis is a **thin loader**. The body keeps only the orientation needed to pick the right next file. The substantive debug material lives in two companion files:\n\n- [CAPABILITIES.md](CAPABILITIES.md) — what *kinds of debug surface* DOCA exposes: the layered debug model (install \u002F version \u002F build \u002F link \u002F runtime \u002F program), the read-only-first stance, version-availability of debug tools (e.g. `doca_caps` since DOCA 2.6.0), the cross-library error taxonomy (cross-link only — owned by `doca-programming-guide`), the observability primitives DOCA emits (`stderr` logs, `--sdk-log-level`, the `doca-\u003Clib>-trace` build flavor, library counters), and the safety constraints on debug actions (read-only first, don't mutate install tree mid-investigation).\n- [TASKS.md](TASKS.md) — the actual debug workflows: `## configure` (set up env for high-verbosity debug), `## test` (capture a reproducible state), `## debug` (the canonical layered ladder, the universal entry point that every library `## debug` redirects to), and the *Where to ask for help* routing (NVIDIA DOCA Developer Forum). Three other anchors (`build`, `modify`, `run`) exist for lint compliance and route to [`doca-programming-guide`](..\u002Fdoca-programming-guide\u002FSKILL.md), which owns those verbs after the env \u002F program split.\n\n## Loading order\n\n1. Read this `SKILL.md` first to confirm the user's symptom is *cross-cutting* debug (not env-class only, not program-class only, not library-internal only).\n2. **For the layered debug model, the read-only stance, the version-availability of debug tools, and the observability surface DOCA emits, see [CAPABILITIES.md](CAPABILITIES.md).**\n3. **For the canonical layered debug ladder and the *capture-and-report* workflow, see [TASKS.md](TASKS.md).** The `build`, `modify`, and `run` anchors in `TASKS.md` are stubs that route to [`doca-programming-guide`](..\u002Fdoca-programming-guide\u002FSKILL.md); their substance lives there.\n4. If the user's symptom turns out to be env-class (install \u002F build prerequisites), hand off to [`doca-setup ## debug`](..\u002Fdoca-setup\u002FTASKS.md#debug). If program-class, hand off to [`doca-programming-guide ## debug`](..\u002Fdoca-programming-guide\u002FTASKS.md#debug). If library-internal, hand off to the matching library skill's `## debug` (e.g. [`doca-flow ## debug`](..\u002Flibs\u002Fdoca-flow\u002FTASKS.md#debug)).\n\nThe two companion files cross-link to each other and to [`doca-public-knowledge-map`](..\u002Fdoca-public-knowledge-map\u002FSKILL.md) whenever the right answer is *\"look it up in the public docs or the installed package layout\"* rather than *\"debug-specific guidance\"*.\n\n## Related skills\n\n- [`doca-public-knowledge-map`](..\u002Fdoca-public-knowledge-map\u002FSKILL.md) — public DOCA documentation routing and the on-disk layout of an installed DOCA package. This skill defers all *\"where is X documented\"*, *\"where on disk is Y\"*, and *\"how do I check the installed version\"* questions to the knowledge-map.\n- [`doca-setup`](..\u002Fdoca-setup\u002FSKILL.md) — env-class debug (install \u002F build prerequisites): `pkg-config` failures, missing hugepages, representors not visible, header-vs-runtime version mismatches. `doca-setup ## debug` is the env-class layered ladder; this skill is the cross-cutting debug ladder both env and program ladders escalate to.\n- [`doca-programming-guide`](..\u002Fdoca-programming-guide\u002FSKILL.md) — program-class debug (lifecycle order, `DOCA_ERROR_*` interpretation, `doca_error_get_descr()` use, the validate-before-commit rule). `doca-programming-guide ## debug` is the program-class layered ladder; this skill picks up where it leaves off when the symptom involves cross-library tooling (`gdb`, `valgrind`, container introspection, core dumps).\n- Library skills (e.g. [`doca-flow`](..\u002Flibs\u002Fdoca-flow\u002FSKILL.md), [`doca-dms`](..\u002Fservices\u002Fdoca-dms\u002FSKILL.md), [`doca-caps`](..\u002Ftools\u002Fdoca-caps\u002FSKILL.md)) — library-specific debug overlays. Each library's `## debug` builds on the cross-cutting ladder defined here, then adds its own counters, traces, and inspector tools.\n",{"data":32,"body":36},{"license":23,"name":4,"description":6,"metadata":33,"compatibility":35},{"kind":34},"library","No DOCA install required to read this skill (it is an overlay loaded against any DOCA artifact skill); the validation steps within DO require a live DOCA install at \u002Fopt\u002Fmellanox\u002Fdoca.\n",{"type":37,"children":38},"root",[39,47,97,104,109,362,387,393,414,457,469,567,573,585,726,732,870,898,904],{"type":40,"tag":41,"props":42,"children":43},"element","h1",{"id":4},[44],{"type":45,"value":46},"text","DOCA debug",{"type":40,"tag":48,"props":49,"children":50},"p",{},[51,57,59,71,73,83,85,95],{"type":40,"tag":52,"props":53,"children":54},"strong",{},[55],{"type":45,"value":56},"Where to start:",{"type":45,"value":58}," This skill is the canonical layered-ladder\nreference both ",{"type":40,"tag":60,"props":61,"children":63},"a",{"href":62},"..\u002Fdoca-setup\u002FTASKS.md#debug",[64],{"type":40,"tag":65,"props":66,"children":68},"code",{"className":67},[],[69],{"type":45,"value":70},"doca-setup ## debug",{"type":45,"value":72},"\nand ",{"type":40,"tag":60,"props":74,"children":76},{"href":75},"..\u002Fdoca-programming-guide\u002FTASKS.md#debug",[77],{"type":40,"tag":65,"props":78,"children":80},{"className":79},[],[81],{"type":45,"value":82},"doca-programming-guide ## debug",{"type":45,"value":84},"\nescalate to. If the symptom does not fit cleanly in env-class or\nprogram-class, start at ",{"type":40,"tag":60,"props":86,"children":88},{"href":87},"TASKS.md#debug",[89],{"type":40,"tag":65,"props":90,"children":92},{"className":91},[],[93],{"type":45,"value":94},"TASKS.md ## debug",{"type":45,"value":96}," — the\nfull layered ladder lives there.",{"type":40,"tag":98,"props":99,"children":101},"h2",{"id":100},"example-questions-this-skill-answers-well",[102],{"type":45,"value":103},"Example questions this skill answers well",{"type":40,"tag":48,"props":105,"children":106},{},[107],{"type":45,"value":108},"The CLASSES of debug questions this skill is built to answer, each\nwith one worked example.",{"type":40,"tag":110,"props":111,"children":112},"ul",{},[113,158,205,232,280,329],{"type":40,"tag":114,"props":115,"children":116},"li",{},[117,122,124,146,148,156],{"type":40,"tag":52,"props":118,"children":119},{},[120],{"type":45,"value":121},"\"My DOCA build \u002F link \u002F runtime \u002F program failed — which layer is\nit?\"",{"type":45,"value":123}," — worked example: ",{"type":40,"tag":125,"props":126,"children":127},"em",{},[128,130,136,138,144],{"type":45,"value":129},"\"",{"type":40,"tag":65,"props":131,"children":133},{"className":132},[],[134],{"type":45,"value":135},"ld",{"type":45,"value":137}," says ",{"type":40,"tag":65,"props":139,"children":141},{"className":140},[],[142],{"type":45,"value":143},"undefined reference to doca_flow_init",{"type":45,"value":145}," — which layer?\"",{"type":45,"value":147}," Answered by the canonical layered\nladder in ",{"type":40,"tag":60,"props":149,"children":150},{"href":87},[151],{"type":40,"tag":65,"props":152,"children":154},{"className":153},[],[155],{"type":45,"value":94},{"type":45,"value":157}," (layer 4 = Link).",{"type":40,"tag":114,"props":159,"children":160},{},[161,166,168,173,175,185],{"type":40,"tag":52,"props":162,"children":163},{},[164],{"type":45,"value":165},"\"How do I turn up verbosity for any DOCA library or tool?\"",{"type":45,"value":167}," —\nworked example: ",{"type":40,"tag":125,"props":169,"children":170},{},[171],{"type":45,"value":172},"\"My DOCA Flow program is silent — where do I get\nmore log output?\"",{"type":45,"value":174}," Answered by the trace-flavor \u002F log-level surface\nin ",{"type":40,"tag":60,"props":176,"children":178},{"href":177},"CAPABILITIES.md#observability",[179],{"type":40,"tag":65,"props":180,"children":182},{"className":181},[],[183],{"type":45,"value":184},"CAPABILITIES.md ## Observability",{"type":40,"tag":110,"props":186,"children":187},{},[188],{"type":40,"tag":114,"props":189,"children":190},{},[191,193,203],{"type":45,"value":192},"the run-with-verbosity workflow in\n",{"type":40,"tag":60,"props":194,"children":196},{"href":195},"TASKS.md#run",[197],{"type":40,"tag":65,"props":198,"children":200},{"className":199},[],[201],{"type":45,"value":202},"TASKS.md ## run",{"type":45,"value":204},".",{"type":40,"tag":114,"props":206,"children":207},{},[208,213,214,219,221,231],{"type":40,"tag":52,"props":209,"children":210},{},[211],{"type":45,"value":212},"\"How do I capture state for a forum question or bug report?\"",{"type":45,"value":167},{"type":40,"tag":125,"props":215,"children":216},{},[217],{"type":45,"value":218},"\"I'd like to file a forum question with reproducible\ncontext — what should I include?\"",{"type":45,"value":220}," Answered by the\ncapture-a-reproducible-state workflow in\n",{"type":40,"tag":60,"props":222,"children":224},{"href":223},"TASKS.md#test",[225],{"type":40,"tag":65,"props":226,"children":228},{"className":227},[],[229],{"type":45,"value":230},"TASKS.md ## test",{"type":45,"value":204},{"type":40,"tag":114,"props":233,"children":234},{},[235,240,241,253,255,265],{"type":40,"tag":52,"props":236,"children":237},{},[238],{"type":45,"value":239},"\"What does this DOCA tool's output mean \u002F which tool answers\nthis?\"",{"type":45,"value":123},{"type":40,"tag":125,"props":242,"children":243},{},[244,245,251],{"type":45,"value":129},{"type":40,"tag":65,"props":246,"children":248},{"className":247},[],[249],{"type":45,"value":250},"doca_caps",{"type":45,"value":252}," returned nothing for RDMA\n— does that mean unsupported?\"",{"type":45,"value":254}," Answered by the\ntool-vs-capability decision tree in\n",{"type":40,"tag":60,"props":256,"children":258},{"href":257},"CAPABILITIES.md#capabilities-and-modes",[259],{"type":40,"tag":65,"props":260,"children":262},{"className":261},[],[263],{"type":45,"value":264},"CAPABILITIES.md ## Capabilities and modes",{"type":40,"tag":110,"props":266,"children":267},{},[268],{"type":40,"tag":114,"props":269,"children":270},{},[271,273,279],{"type":45,"value":272},"cross-link to ",{"type":40,"tag":60,"props":274,"children":276},{"href":275},"..\u002Ftools\u002Fdoca-caps\u002FSKILL.md",[277],{"type":45,"value":278},"doca-caps",{"type":45,"value":204},{"type":40,"tag":114,"props":281,"children":282},{},[283,288,289,301,303,313],{"type":40,"tag":52,"props":284,"children":285},{},[286],{"type":45,"value":287},"\"I'm debugging inside the NGC container — what's\nobservable?\"",{"type":45,"value":123},{"type":40,"tag":125,"props":290,"children":291},{},[292,293,299],{"type":45,"value":129},{"type":40,"tag":65,"props":294,"children":296},{"className":295},[],[297],{"type":45,"value":298},"hugepages",{"type":45,"value":300}," is empty inside the\ncontainer; is that a real problem or a container thing?\"",{"type":45,"value":302},"\nAnswered by the container-specific debug constraints in\n",{"type":40,"tag":60,"props":304,"children":306},{"href":305},"CAPABILITIES.md#safety-policy",[307],{"type":40,"tag":65,"props":308,"children":310},{"className":309},[],[311],{"type":45,"value":312},"CAPABILITIES.md ## Safety policy",{"type":40,"tag":110,"props":314,"children":315},{},[316],{"type":40,"tag":114,"props":317,"children":318},{},[319,327],{"type":40,"tag":60,"props":320,"children":321},{"href":87},[322],{"type":40,"tag":65,"props":323,"children":325},{"className":324},[],[326],{"type":45,"value":94},{"type":45,"value":328}," layer 5 (Runtime).",{"type":40,"tag":114,"props":330,"children":331},{},[332,337,338,343,345,353,355,361],{"type":40,"tag":52,"props":333,"children":334},{},[335],{"type":45,"value":336},"\"Where do I ask for help with this?\"",{"type":45,"value":123},{"type":40,"tag":125,"props":339,"children":340},{},[341],{"type":45,"value":342},"\"Where\nis the customer-facing DOCA forum and what should the post\ncontain?\"",{"type":45,"value":344}," Answered by the Developer-Forum routing rule in\n",{"type":40,"tag":60,"props":346,"children":347},{"href":87},[348],{"type":40,"tag":65,"props":349,"children":351},{"className":350},[],[352],{"type":45,"value":94},{"type":45,"value":354}," plus\n",{"type":40,"tag":60,"props":356,"children":358},{"href":357},"..\u002Fdoca-public-knowledge-map\u002FSKILL.md",[359],{"type":45,"value":360},"doca-public-knowledge-map",{"type":45,"value":204},{"type":40,"tag":48,"props":363,"children":364},{},[365,367,375,377,385],{"type":45,"value":366},"If the symptom is purely env-class, route to\n",{"type":40,"tag":60,"props":368,"children":369},{"href":62},[370],{"type":40,"tag":65,"props":371,"children":373},{"className":372},[],[374],{"type":45,"value":70},{"type":45,"value":376},"; if purely\nprogram-class, route to\n",{"type":40,"tag":60,"props":378,"children":379},{"href":75},[380],{"type":40,"tag":65,"props":381,"children":383},{"className":382},[],[384],{"type":45,"value":82},{"type":45,"value":386},".\nThis skill is the cross-cutting layer both call into.",{"type":40,"tag":98,"props":388,"children":390},{"id":389},"when-to-load-this-skill",[391],{"type":45,"value":392},"When to load this skill",{"type":40,"tag":48,"props":394,"children":395},{},[396,398,404,406,412],{"type":45,"value":397},"Load this skill when the user is debugging anything DOCA-related — a build that won't compile, a link step that can't resolve a ",{"type":40,"tag":65,"props":399,"children":401},{"className":400},[],[402],{"type":45,"value":403},"doca_*",{"type":45,"value":405}," symbol, a runtime call that returns ",{"type":40,"tag":65,"props":407,"children":409},{"className":408},[],[410],{"type":45,"value":411},"DOCA_ERROR_*",{"type":45,"value":413},", a packet that does not appear on the wire, a service that won't start, or a tool that returns no useful output. Concretely:",{"type":40,"tag":110,"props":415,"children":416},{},[417,422,427,432,445],{"type":40,"tag":114,"props":418,"children":419},{},[420],{"type":45,"value":421},"The user reports a symptom and needs to find the layer that caused it (install \u002F version \u002F build \u002F link \u002F runtime \u002F program).",{"type":40,"tag":114,"props":423,"children":424},{},[425],{"type":45,"value":426},"The user asks \"how do I get more logs?\" or \"how do I turn up the verbosity?\" for any DOCA library or tool.",{"type":40,"tag":114,"props":428,"children":429},{},[430],{"type":45,"value":431},"The user wants to capture state for a forum question or an internal bug report (the bundle does not own the internal-bug-report channel; it routes to the public DOCA Developer Forum).",{"type":40,"tag":114,"props":433,"children":434},{},[435,437,443],{"type":45,"value":436},"The user is reading a stack trace, a ",{"type":40,"tag":65,"props":438,"children":440},{"className":439},[],[441],{"type":45,"value":442},"valgrind",{"type":45,"value":444}," output, or a core dump from a DOCA program and wants to know where to look first.",{"type":40,"tag":114,"props":446,"children":447},{},[448,450,455],{"type":45,"value":449},"The user is debugging ",{"type":40,"tag":125,"props":451,"children":452},{},[453],{"type":45,"value":454},"inside",{"type":45,"value":456}," the NGC DOCA container and needs to know what is and is not observable from inside it.",{"type":40,"tag":48,"props":458,"children":459},{},[460,462,467],{"type":45,"value":461},"Do ",{"type":40,"tag":52,"props":463,"children":464},{},[465],{"type":45,"value":466},"not",{"type":45,"value":468}," load this skill for:",{"type":40,"tag":110,"props":470,"children":471},{},[472,509,545],{"type":40,"tag":114,"props":473,"children":474},{},[475,488,490,495,497,507],{"type":40,"tag":125,"props":476,"children":477},{},[478,480,486],{"type":45,"value":479},"\"What is ",{"type":40,"tag":65,"props":481,"children":483},{"className":482},[],[484],{"type":45,"value":485},"DOCA_ERROR_BAD_STATE",{"type":45,"value":487},"?\", \"what error codes does DOCA return?\"",{"type":45,"value":489}," — that is the cross-library ",{"type":40,"tag":125,"props":491,"children":492},{},[493],{"type":45,"value":494},"error taxonomy",{"type":45,"value":496},", owned by ",{"type":40,"tag":60,"props":498,"children":500},{"href":499},"..\u002Fdoca-programming-guide\u002FCAPABILITIES.md#error-taxonomy",[501],{"type":40,"tag":65,"props":502,"children":504},{"className":503},[],[505],{"type":45,"value":506},"doca-programming-guide CAPABILITIES.md ## Error taxonomy",{"type":45,"value":508},". This skill consumes that taxonomy; it does not redefine it.",{"type":40,"tag":114,"props":510,"children":511},{},[512,533,535,543],{"type":40,"tag":125,"props":513,"children":514},{},[515,517,523,525,531],{"type":45,"value":516},"\"My ",{"type":40,"tag":65,"props":518,"children":520},{"className":519},[],[521],{"type":45,"value":522},"pkg-config",{"type":45,"value":524}," cannot find ",{"type":40,"tag":65,"props":526,"children":528},{"className":527},[],[529],{"type":45,"value":530},"doca-flow",{"type":45,"value":532},"\", \"hugepages are not mounted\", \"my representor isn't visible\"",{"type":45,"value":534}," — those are env-class symptoms, owned by ",{"type":40,"tag":60,"props":536,"children":537},{"href":62},[538],{"type":40,"tag":65,"props":539,"children":541},{"className":540},[],[542],{"type":45,"value":70},{"type":45,"value":544},". This skill is the canonical pointer that env-class debug ladder redirects to once the symptom escalates beyond install \u002F version \u002F build prerequisites.",{"type":40,"tag":114,"props":546,"children":547},{},[548,553,555,565],{"type":40,"tag":125,"props":549,"children":550},{},[551],{"type":45,"value":552},"Library-specific debugging",{"type":45,"value":554}," (Flow pipe trace, RDMA queue-pair state, Comch channel statistics) — those live in the matching library skill (e.g. ",{"type":40,"tag":60,"props":556,"children":558},{"href":557},"..\u002Flibs\u002Fdoca-flow\u002FTASKS.md#debug",[559],{"type":40,"tag":65,"props":560,"children":562},{"className":561},[],[563],{"type":45,"value":564},"doca-flow ## debug",{"type":45,"value":566},"). This skill provides the cross-cutting debug ladder; library skills layer their library-specific debug surface on top of it.",{"type":40,"tag":98,"props":568,"children":570},{"id":569},"what-this-skill-provides",[571],{"type":45,"value":572},"What this skill provides",{"type":40,"tag":48,"props":574,"children":575},{},[576,578,583],{"type":45,"value":577},"This is a ",{"type":40,"tag":52,"props":579,"children":580},{},[581],{"type":45,"value":582},"thin loader",{"type":45,"value":584},". The body keeps only the orientation needed to pick the right next file. The substantive debug material lives in two companion files:",{"type":40,"tag":110,"props":586,"children":587},{},[588,644],{"type":40,"tag":114,"props":589,"children":590},{},[591,596,598,603,605,610,612,618,620,626,628,634,636,642],{"type":40,"tag":60,"props":592,"children":594},{"href":593},"CAPABILITIES.md",[595],{"type":45,"value":593},{"type":45,"value":597}," — what ",{"type":40,"tag":125,"props":599,"children":600},{},[601],{"type":45,"value":602},"kinds of debug surface",{"type":45,"value":604}," DOCA exposes: the layered debug model (install \u002F version \u002F build \u002F link \u002F runtime \u002F program), the read-only-first stance, version-availability of debug tools (e.g. ",{"type":40,"tag":65,"props":606,"children":608},{"className":607},[],[609],{"type":45,"value":250},{"type":45,"value":611}," since DOCA 2.6.0), the cross-library error taxonomy (cross-link only — owned by ",{"type":40,"tag":65,"props":613,"children":615},{"className":614},[],[616],{"type":45,"value":617},"doca-programming-guide",{"type":45,"value":619},"), the observability primitives DOCA emits (",{"type":40,"tag":65,"props":621,"children":623},{"className":622},[],[624],{"type":45,"value":625},"stderr",{"type":45,"value":627}," logs, ",{"type":40,"tag":65,"props":629,"children":631},{"className":630},[],[632],{"type":45,"value":633},"--sdk-log-level",{"type":45,"value":635},", the ",{"type":40,"tag":65,"props":637,"children":639},{"className":638},[],[640],{"type":45,"value":641},"doca-\u003Clib>-trace",{"type":45,"value":643}," build flavor, library counters), and the safety constraints on debug actions (read-only first, don't mutate install tree mid-investigation).",{"type":40,"tag":114,"props":645,"children":646},{},[647,652,654,660,662,668,670,676,678,683,685,690,692,698,700,706,707,713,715,724],{"type":40,"tag":60,"props":648,"children":650},{"href":649},"TASKS.md",[651],{"type":45,"value":649},{"type":45,"value":653}," — the actual debug workflows: ",{"type":40,"tag":65,"props":655,"children":657},{"className":656},[],[658],{"type":45,"value":659},"## configure",{"type":45,"value":661}," (set up env for high-verbosity debug), ",{"type":40,"tag":65,"props":663,"children":665},{"className":664},[],[666],{"type":45,"value":667},"## test",{"type":45,"value":669}," (capture a reproducible state), ",{"type":40,"tag":65,"props":671,"children":673},{"className":672},[],[674],{"type":45,"value":675},"## debug",{"type":45,"value":677}," (the canonical layered ladder, the universal entry point that every library ",{"type":40,"tag":65,"props":679,"children":681},{"className":680},[],[682],{"type":45,"value":675},{"type":45,"value":684}," redirects to), and the ",{"type":40,"tag":125,"props":686,"children":687},{},[688],{"type":45,"value":689},"Where to ask for help",{"type":45,"value":691}," routing (NVIDIA DOCA Developer Forum). Three other anchors (",{"type":40,"tag":65,"props":693,"children":695},{"className":694},[],[696],{"type":45,"value":697},"build",{"type":45,"value":699},", ",{"type":40,"tag":65,"props":701,"children":703},{"className":702},[],[704],{"type":45,"value":705},"modify",{"type":45,"value":699},{"type":40,"tag":65,"props":708,"children":710},{"className":709},[],[711],{"type":45,"value":712},"run",{"type":45,"value":714},") exist for lint compliance and route to ",{"type":40,"tag":60,"props":716,"children":718},{"href":717},"..\u002Fdoca-programming-guide\u002FSKILL.md",[719],{"type":40,"tag":65,"props":720,"children":722},{"className":721},[],[723],{"type":45,"value":617},{"type":45,"value":725},", which owns those verbs after the env \u002F program split.",{"type":40,"tag":98,"props":727,"children":729},{"id":728},"loading-order",[730],{"type":45,"value":731},"Loading order",{"type":40,"tag":733,"props":734,"children":735},"ol",{},[736,756,769,828],{"type":40,"tag":114,"props":737,"children":738},{},[739,741,747,749,754],{"type":45,"value":740},"Read this ",{"type":40,"tag":65,"props":742,"children":744},{"className":743},[],[745],{"type":45,"value":746},"SKILL.md",{"type":45,"value":748}," first to confirm the user's symptom is ",{"type":40,"tag":125,"props":750,"children":751},{},[752],{"type":45,"value":753},"cross-cutting",{"type":45,"value":755}," debug (not env-class only, not program-class only, not library-internal only).",{"type":40,"tag":114,"props":757,"children":758},{},[759],{"type":40,"tag":52,"props":760,"children":761},{},[762,764,768],{"type":45,"value":763},"For the layered debug model, the read-only stance, the version-availability of debug tools, and the observability surface DOCA emits, see ",{"type":40,"tag":60,"props":765,"children":766},{"href":593},[767],{"type":45,"value":593},{"type":45,"value":204},{"type":40,"tag":114,"props":770,"children":771},{},[772,789,791,796,797,802,804,809,811,816,818,826],{"type":40,"tag":52,"props":773,"children":774},{},[775,777,782,784,788],{"type":45,"value":776},"For the canonical layered debug ladder and the ",{"type":40,"tag":125,"props":778,"children":779},{},[780],{"type":45,"value":781},"capture-and-report",{"type":45,"value":783}," workflow, see ",{"type":40,"tag":60,"props":785,"children":786},{"href":649},[787],{"type":45,"value":649},{"type":45,"value":204},{"type":45,"value":790}," The ",{"type":40,"tag":65,"props":792,"children":794},{"className":793},[],[795],{"type":45,"value":697},{"type":45,"value":699},{"type":40,"tag":65,"props":798,"children":800},{"className":799},[],[801],{"type":45,"value":705},{"type":45,"value":803},", and ",{"type":40,"tag":65,"props":805,"children":807},{"className":806},[],[808],{"type":45,"value":712},{"type":45,"value":810}," anchors in ",{"type":40,"tag":65,"props":812,"children":814},{"className":813},[],[815],{"type":45,"value":649},{"type":45,"value":817}," are stubs that route to ",{"type":40,"tag":60,"props":819,"children":820},{"href":717},[821],{"type":40,"tag":65,"props":822,"children":824},{"className":823},[],[825],{"type":45,"value":617},{"type":45,"value":827},"; their substance lives there.",{"type":40,"tag":114,"props":829,"children":830},{},[831,833,841,843,851,853,858,860,868],{"type":45,"value":832},"If the user's symptom turns out to be env-class (install \u002F build prerequisites), hand off to ",{"type":40,"tag":60,"props":834,"children":835},{"href":62},[836],{"type":40,"tag":65,"props":837,"children":839},{"className":838},[],[840],{"type":45,"value":70},{"type":45,"value":842},". If program-class, hand off to ",{"type":40,"tag":60,"props":844,"children":845},{"href":75},[846],{"type":40,"tag":65,"props":847,"children":849},{"className":848},[],[850],{"type":45,"value":82},{"type":45,"value":852},". If library-internal, hand off to the matching library skill's ",{"type":40,"tag":65,"props":854,"children":856},{"className":855},[],[857],{"type":45,"value":675},{"type":45,"value":859}," (e.g. ",{"type":40,"tag":60,"props":861,"children":862},{"href":557},[863],{"type":40,"tag":65,"props":864,"children":866},{"className":865},[],[867],{"type":45,"value":564},{"type":45,"value":869},").",{"type":40,"tag":48,"props":871,"children":872},{},[873,875,883,885,890,892,897],{"type":45,"value":874},"The two companion files cross-link to each other and to ",{"type":40,"tag":60,"props":876,"children":877},{"href":357},[878],{"type":40,"tag":65,"props":879,"children":881},{"className":880},[],[882],{"type":45,"value":360},{"type":45,"value":884}," whenever the right answer is ",{"type":40,"tag":125,"props":886,"children":887},{},[888],{"type":45,"value":889},"\"look it up in the public docs or the installed package layout\"",{"type":45,"value":891}," rather than ",{"type":40,"tag":125,"props":893,"children":894},{},[895],{"type":45,"value":896},"\"debug-specific guidance\"",{"type":45,"value":204},{"type":40,"tag":98,"props":899,"children":901},{"id":900},"related-skills",[902],{"type":45,"value":903},"Related skills",{"type":40,"tag":110,"props":905,"children":906},{},[907,939,968,1017],{"type":40,"tag":114,"props":908,"children":909},{},[910,918,920,925,926,931,932,937],{"type":40,"tag":60,"props":911,"children":912},{"href":357},[913],{"type":40,"tag":65,"props":914,"children":916},{"className":915},[],[917],{"type":45,"value":360},{"type":45,"value":919}," — public DOCA documentation routing and the on-disk layout of an installed DOCA package. This skill defers all ",{"type":40,"tag":125,"props":921,"children":922},{},[923],{"type":45,"value":924},"\"where is X documented\"",{"type":45,"value":699},{"type":40,"tag":125,"props":927,"children":928},{},[929],{"type":45,"value":930},"\"where on disk is Y\"",{"type":45,"value":803},{"type":40,"tag":125,"props":933,"children":934},{},[935],{"type":45,"value":936},"\"how do I check the installed version\"",{"type":45,"value":938}," questions to the knowledge-map.",{"type":40,"tag":114,"props":940,"children":941},{},[942,952,954,959,961,966],{"type":40,"tag":60,"props":943,"children":945},{"href":944},"..\u002Fdoca-setup\u002FSKILL.md",[946],{"type":40,"tag":65,"props":947,"children":949},{"className":948},[],[950],{"type":45,"value":951},"doca-setup",{"type":45,"value":953}," — env-class debug (install \u002F build prerequisites): ",{"type":40,"tag":65,"props":955,"children":957},{"className":956},[],[958],{"type":45,"value":522},{"type":45,"value":960}," failures, missing hugepages, representors not visible, header-vs-runtime version mismatches. ",{"type":40,"tag":65,"props":962,"children":964},{"className":963},[],[965],{"type":45,"value":70},{"type":45,"value":967}," is the env-class layered ladder; this skill is the cross-cutting debug ladder both env and program ladders escalate to.",{"type":40,"tag":114,"props":969,"children":970},{},[971,979,981,986,988,994,996,1001,1003,1009,1010,1015],{"type":40,"tag":60,"props":972,"children":973},{"href":717},[974],{"type":40,"tag":65,"props":975,"children":977},{"className":976},[],[978],{"type":45,"value":617},{"type":45,"value":980}," — program-class debug (lifecycle order, ",{"type":40,"tag":65,"props":982,"children":984},{"className":983},[],[985],{"type":45,"value":411},{"type":45,"value":987}," interpretation, ",{"type":40,"tag":65,"props":989,"children":991},{"className":990},[],[992],{"type":45,"value":993},"doca_error_get_descr()",{"type":45,"value":995}," use, the validate-before-commit rule). ",{"type":40,"tag":65,"props":997,"children":999},{"className":998},[],[1000],{"type":45,"value":82},{"type":45,"value":1002}," is the program-class layered ladder; this skill picks up where it leaves off when the symptom involves cross-library tooling (",{"type":40,"tag":65,"props":1004,"children":1006},{"className":1005},[],[1007],{"type":45,"value":1008},"gdb",{"type":45,"value":699},{"type":40,"tag":65,"props":1011,"children":1013},{"className":1012},[],[1014],{"type":45,"value":442},{"type":45,"value":1016},", container introspection, core dumps).",{"type":40,"tag":114,"props":1018,"children":1019},{},[1020,1022,1031,1032,1042,1043,1051,1053,1058],{"type":45,"value":1021},"Library skills (e.g. ",{"type":40,"tag":60,"props":1023,"children":1025},{"href":1024},"..\u002Flibs\u002Fdoca-flow\u002FSKILL.md",[1026],{"type":40,"tag":65,"props":1027,"children":1029},{"className":1028},[],[1030],{"type":45,"value":530},{"type":45,"value":699},{"type":40,"tag":60,"props":1033,"children":1035},{"href":1034},"..\u002Fservices\u002Fdoca-dms\u002FSKILL.md",[1036],{"type":40,"tag":65,"props":1037,"children":1039},{"className":1038},[],[1040],{"type":45,"value":1041},"doca-dms",{"type":45,"value":699},{"type":40,"tag":60,"props":1044,"children":1045},{"href":275},[1046],{"type":40,"tag":65,"props":1047,"children":1049},{"className":1048},[],[1050],{"type":45,"value":278},{"type":45,"value":1052},") — library-specific debug overlays. Each library's ",{"type":40,"tag":65,"props":1054,"children":1056},{"className":1055},[],[1057],{"type":45,"value":675},{"type":45,"value":1059}," builds on the cross-cutting ladder defined here, then adds its own counters, traces, and inspector tools.",{"items":1061,"total":1218},[1062,1080,1098,1109,1121,1133,1146,1160,1173,1184,1198,1207],{"slug":1063,"name":1063,"fn":1064,"description":1065,"org":1066,"tags":1067,"stars":1077,"repoUrl":1078,"updatedAt":1079},"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},[1068,1071,1074],{"name":1069,"slug":1070,"type":13},"Documentation","documentation",{"name":1072,"slug":1073,"type":13},"MCP","mcp",{"name":1075,"slug":1076,"type":13},"Search","search",21777,"https:\u002F\u002Fgithub.com\u002FNVIDIA\u002FNemoClaw","2026-07-20T06:00:01.461044",{"slug":1081,"name":1081,"fn":1082,"description":1083,"org":1084,"tags":1085,"stars":1095,"repoUrl":1096,"updatedAt":1097},"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},[1086,1089,1092],{"name":1087,"slug":1088,"type":13},"Containers","containers",{"name":1090,"slug":1091,"type":13},"Deployment","deployment",{"name":1093,"slug":1094,"type":13},"Python","python",17049,"https:\u002F\u002Fgithub.com\u002FNVIDIA\u002FMegatron-LM","2026-07-27T06:06:11.249662",{"slug":1099,"name":1099,"fn":1100,"description":1101,"org":1102,"tags":1103,"stars":1095,"repoUrl":1096,"updatedAt":1108},"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},[1104,1107],{"name":1105,"slug":1106,"type":13},"CI\u002FCD","ci-cd",{"name":1090,"slug":1091,"type":13},"2026-07-14T05:25:59.97109",{"slug":1110,"name":1110,"fn":1111,"description":1112,"org":1113,"tags":1114,"stars":1095,"repoUrl":1096,"updatedAt":1120},"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},[1115,1116,1117],{"name":1105,"slug":1106,"type":13},{"name":1090,"slug":1091,"type":13},{"name":1118,"slug":1119,"type":13},"GitHub","github","2026-07-27T06:06:12.278222",{"slug":1122,"name":1122,"fn":1123,"description":1124,"org":1125,"tags":1126,"stars":1095,"repoUrl":1096,"updatedAt":1132},"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},[1127,1128,1129],{"name":18,"slug":19,"type":13},{"name":1118,"slug":1119,"type":13},{"name":1130,"slug":1131,"type":13},"Triage","triage","2026-07-14T05:25:57.442089",{"slug":1134,"name":1134,"fn":1135,"description":1136,"org":1137,"tags":1138,"stars":1095,"repoUrl":1096,"updatedAt":1145},"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},[1139,1142],{"name":1140,"slug":1141,"type":13},"Best Practices","best-practices",{"name":1143,"slug":1144,"type":13},"Code Analysis","code-analysis","2026-07-14T05:25:56.18433",{"slug":1147,"name":1147,"fn":1148,"description":1149,"org":1150,"tags":1151,"stars":1095,"repoUrl":1096,"updatedAt":1159},"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},[1152,1155,1158],{"name":1153,"slug":1154,"type":13},"Machine Learning","machine-learning",{"name":1156,"slug":1157,"type":13},"Migration","migration",{"name":9,"slug":8,"type":13},"2026-07-17T06:07:11.777011",{"slug":1161,"name":1161,"fn":1162,"description":1163,"org":1164,"tags":1165,"stars":1095,"repoUrl":1096,"updatedAt":1172},"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},[1166,1169],{"name":1167,"slug":1168,"type":13},"QA","qa",{"name":1170,"slug":1171,"type":13},"Testing","testing","2026-07-14T05:25:53.673039",{"slug":1174,"name":1174,"fn":1175,"description":1176,"org":1177,"tags":1178,"stars":1095,"repoUrl":1096,"updatedAt":1183},"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},[1179,1180],{"name":1090,"slug":1091,"type":13},{"name":1181,"slug":1182,"type":13},"Infrastructure","infrastructure","2026-07-14T05:25:49.362534",{"slug":1185,"name":1185,"fn":1186,"description":1187,"org":1188,"tags":1189,"stars":1095,"repoUrl":1096,"updatedAt":1197},"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},[1190,1193,1194],{"name":1191,"slug":1192,"type":13},"Code Review","code-review",{"name":1118,"slug":1119,"type":13},{"name":1195,"slug":1196,"type":13},"Pull Requests","pull-requests","2026-07-14T05:26:01.226578",{"slug":1199,"name":1199,"fn":1200,"description":1201,"org":1202,"tags":1203,"stars":1095,"repoUrl":1096,"updatedAt":1206},"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},[1204,1205],{"name":1167,"slug":1168,"type":13},{"name":1170,"slug":1171,"type":13},"2026-07-14T05:25:54.928983",{"slug":1208,"name":1208,"fn":1209,"description":1210,"org":1211,"tags":1212,"stars":1095,"repoUrl":1096,"updatedAt":1217},"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},[1213,1216],{"name":1214,"slug":1215,"type":13},"Automation","automation",{"name":1105,"slug":1106,"type":13},"2026-07-30T05:29:03.275638",496,{"items":1220,"total":1316},[1221,1238,1248,1262,1272,1287,1302],{"slug":1222,"name":1222,"fn":1223,"description":1224,"org":1225,"tags":1226,"stars":20,"repoUrl":21,"updatedAt":1237},"accelerated-computing-cudf","accelerate data processing with cuDF","Official NVIDIA-authored guidance for NVIDIA cuDF GPU DataFrames, pandas acceleration, dask-cuDF, ETL, joins, groupby, CSV\u002FParquet I\u002FO, nullable semantics, and multi-GPU DataFrame workloads.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1227,1230,1233,1234],{"name":1228,"slug":1229,"type":13},"Data Analysis","data-analysis",{"name":1231,"slug":1232,"type":13},"Data Engineering","data-engineering",{"name":9,"slug":8,"type":13},{"name":1235,"slug":1236,"type":13},"Performance","performance","2026-07-14T05:28:43.176466",{"slug":1239,"name":1239,"fn":1240,"description":1241,"org":1242,"tags":1243,"stars":20,"repoUrl":21,"updatedAt":1247},"aiq-deploy","deploy and manage NVIDIA AI-Q infrastructure","Use when asked to install, deploy, run, validate, troubleshoot, or stop NVIDIA AI-Q Blueprint infrastructure.\n",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1244,1245,1246],{"name":1090,"slug":1091,"type":13},{"name":1181,"slug":1182,"type":13},{"name":9,"slug":8,"type":13},"2026-07-14T05:29:06.667109",{"slug":1249,"name":1249,"fn":1250,"description":1251,"org":1252,"tags":1253,"stars":20,"repoUrl":21,"updatedAt":1261},"aiq-research","conduct deep research with AI-Q","Use when asked to run deep research or AI-Q research through a reachable NVIDIA AI-Q Blueprint backend.\n",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1254,1257,1258],{"name":1255,"slug":1256,"type":13},"Agents","agents",{"name":9,"slug":8,"type":13},{"name":1259,"slug":1260,"type":13},"Research","research","2026-07-14T05:28:06.816956",{"slug":1263,"name":1263,"fn":1264,"description":1265,"org":1266,"tags":1267,"stars":20,"repoUrl":21,"updatedAt":1271},"amc-run-sample-calibration","run AMC sample dataset calibration","Run end-to-end calibration on the shipped sample dataset (sdg_08_2_sample_data_010926.zip) against a running AMC microservice. Use when user says 'test sample dataset', 'run sample calibration', 'verify AMC install', or 'launch and test'.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1268,1269,1270],{"name":1228,"slug":1229,"type":13},{"name":9,"slug":8,"type":13},{"name":1170,"slug":1171,"type":13},"2026-07-17T05:29:03.913266",{"slug":1273,"name":1273,"fn":1274,"description":1275,"org":1276,"tags":1277,"stars":20,"repoUrl":21,"updatedAt":1286},"amc-run-video-calibration","calibrate video datasets with AutoMagicCalib","Calibrate a new dataset from pre-recorded video files via the AutoMagicCalib REST API. Use when user has local MP4s and says 'calibrate my videos', 'run AMC on these videos', or similar. For RTSP\u002Flive streams, use amc-run-rtsp-calibration instead.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1278,1279,1282,1283],{"name":1214,"slug":1215,"type":13},{"name":1280,"slug":1281,"type":13},"Imaging","imaging",{"name":9,"slug":8,"type":13},{"name":1284,"slug":1285,"type":13},"Video","video","2026-07-17T05:28:53.905004",{"slug":1288,"name":1288,"fn":1289,"description":1290,"org":1291,"tags":1292,"stars":20,"repoUrl":21,"updatedAt":1301},"amc-setup-calibration-stack","deploy AutoMagicCalib microservice with Docker","Launch AutoMagicCalib microservice and web UI from NGC release images via Docker Compose. Use when user says 'deploy auto calibration', 'launch auto calibration', 'launch AMC', 'start MS+UI', or 'set up auto-magic-calib'. Requires NGC API key.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1293,1294,1297,1298],{"name":1090,"slug":1091,"type":13},{"name":1295,"slug":1296,"type":13},"Docker","docker",{"name":9,"slug":8,"type":13},{"name":1299,"slug":1300,"type":13},"Operations","operations","2026-07-17T05:28:56.913999",{"slug":1303,"name":1303,"fn":1304,"description":1305,"org":1306,"tags":1307,"stars":20,"repoUrl":21,"updatedAt":1315},"cudaq-guide","develop quantum applications with CUDA-Q","CUDA-Q onboarding guide for installation, test programs, GPU simulation, QPU hardware, and quantum applications.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},[1308,1309,1312],{"name":9,"slug":8,"type":13},{"name":1310,"slug":1311,"type":13},"Quantum Computing","quantum-computing",{"name":1313,"slug":1314,"type":13},"Simulation","simulation","2026-07-14T05:26:58.898253",305]