[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill-nvidia-i4h-workflow-create":3,"mdc-xzvwfd-key":31,"related-org-nvidia-i4h-workflow-create":1217,"related-repo-nvidia-i4h-workflow-create":1377},{"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},"i4h-workflow-create","create new agentic environments","Create a new agentic env by forking an existing env. Use for new env\u002Ftask scaffolding, not scene edits or baking.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},"nvidia","NVIDIA","https:\u002F\u002Fpexgzepcugksgbtrxkhf.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Forg-logos\u002Fnvidia.png",[12,16,17],{"name":13,"slug":14,"type":15},"Robotics","robotics","tag",{"name":9,"slug":8,"type":15},{"name":18,"slug":19,"type":15},"Engineering","engineering",2473,"https:\u002F\u002Fgithub.com\u002FNVIDIA\u002Fskills","2026-07-23T06:06:13.18648","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\u002Fi4h-workflow-create","---\nname: i4h-workflow-create\nversion: \"0.7.0\"\ndescription: Create a new agentic env by forking an existing env. Use for new env\u002Ftask scaffolding, not scene edits or baking.\nlicense: Apache-2.0\nmetadata:\n  author: \"Isaac for Healthcare Team \u003Cisaac-for-healthcare-support@nvidia.com>\"\n  tags:\n    - isaac-for-healthcare\n    - i4h\n    - agentic-workflow\n    - environment\n    - scaffolding\n---\n\n# i4h Workflow - Create Env\n\n## Purpose\n\nCreate the first runnable version of a new `workflows\u002Fagentic` environment by forking the closest existing env. Keep this skill focused on env scaffolding: YAML, assets, task, env class, runtime, and validation. Do not use this skill to polish an existing scene, add optional props\u002Fcameras, or bake bridge edits; use [[i4h-workflow-scene-edit]] for that.\n\n## Base Code\n\nResolve and work from the i4h-workflows root:\n\n```bash\nROOT=\"${I4H_WORKFLOWS:-$(git rev-parse --show-toplevel 2>\u002Fdev\u002Fnull)}\"\nif [ ! -d \"$ROOT\u002Fworkflows\u002Fagentic\" ]; then\n  ROOT=\"${I4H_WORKFLOWS:-$HOME\u002Fi4h-workflows}\"\n  [ -d \"$ROOT\u002Fworkflows\u002Fagentic\" ] || git clone https:\u002F\u002Fgithub.com\u002Fisaac-for-healthcare\u002Fi4h-workflows \"$ROOT\"\nfi\nexport I4H_WORKFLOWS=\"$ROOT\"; cd \"$ROOT\"\n```\n\n## What to Load\n\nBefore editing, load only the references that match the request:\n\n- Always load `skills\u002Fi4h-workflow\u002Freferences\u002Frepo-map.md`.\n- Always load `skills\u002Fi4h-workflow-create\u002Freferences\u002Fcreate-contract.md`.\n- Load `skills\u002Fi4h-workflow-create\u002Freferences\u002Fenv-authoring-patterns.md` to choose the source env, robot owner, policy stack, and YAML pattern.\n- Load `skills\u002Fi4h-workflow-create\u002Freferences\u002Fhybrid-layout-rules.md` only for hybrid scene+robot envs, G1 footprint\u002Ftable-height work, catalog USD scale, or support-surface layout.\n- Load a recipe file only when the prompt matches it exactly enough to avoid re-deciding components.\n\nRecipe routing:\n\n| Prompt shape | Reference |\n|---|---|\n| Surgical tool sorting using G1 based on `scissor_pick_and_place` | `skills\u002Fi4h-workflow-create\u002Freferences\u002Fg1-surgical-tool-sort.md` |\n\n## Create Contract\n\nCreate the env shell first. A normal create task should produce exactly these surfaces unless the chosen existing pattern requires an explicit exception:\n\n| Surface | Path |\n|---|---|\n| Env YAML | `workflows\u002Fagentic\u002Fconfig\u002Fenvironments\u002F\u003Cenv>.yaml` |\n| Assets | `workflows\u002Fagentic\u002Farena\u002Farena\u002Fassets\u002F\u003Cenv>.py` |\n| Task | `workflows\u002Fagentic\u002Farena\u002Farena\u002Ftasks\u002F\u003Cenv>.py` |\n| Env class | `workflows\u002Fagentic\u002Farena\u002Farena\u002Fenvironments\u002F\u003Cenv>_environment.py` |\n| Runtime | `workflows\u002Fagentic\u002Farena\u002Farena\u002Fruntimes\u002F\u003Cenv>.py` |\n\nKeep all paths repo-root relative and keep the `workflows\u002Fagentic\u002F` prefix. Do not create a new policy package, README, docs, or shared-module edits unless the user explicitly asks and the policy stack truly needs it.\n\n## Workflow\n\n1. Resolve components: env id, source env, scene\u002Fassets source, robot owner, policy stack, model\u002Fcheckpoint, cameras, objects\u002Fdestinations, success rule.\n2. Inspect the selected source env YAML, env class, assets, task, runtime, and policy stack files before writing code.\n3. Fork the closest working pattern. Preserve inline-scene vs registry-asset style; do not invent a new architecture.\n4. Keep the first version minimal and runnable. Optional visual polish, extra props, new cameras, and layout changes happen through scene-edit after the env exists.\n5. Run validation and fix source until the env builds and the rendered scene is visually sane.\n\nFor an exact recipe match, do not re-derive the architecture after loading the recipe. Read only the source files the recipe names, skip large source runtime files when the recipe says the runtime is a re-export, then write the five contract files immediately. Do not inspect third-party framework internals unless a validation error requires it.\n\nPlan shape:\n\n```text\nEnv id:\nSource env \u002F recipe:\nScene\u002Fassets source:\nRobot owner:\nPolicy stack + model\u002Fcheckpoint:\nObjects\u002Fdestinations:\nSuccess rule:\nFiles to create:\nValidation:\n```\n\n## Hard Rules\n\n- Env YAML is the source of truth for robot, policy, cameras, task text, dataset mapping, and train defaults.\n- Fork from the nearest existing implementation. For hybrid envs, robot integration comes from the robot owner and scene construction comes from the scene source.\n- Sorting tasks need at least two object types, at least two destinations, and a success rule that fails swapped placements.\n- Static\u002Fkinematic destination props may not expose `.data.root_pos_w`; success checks must fall back to `entity.get_world_poses()`.\n- Functions referenced by `func=` inside config classes must be defined above those classes.\n- G1 WBC envs need footprint clearance and matched ground\u002Fbase-height values; load `hybrid-layout-rules.md` before choosing those numbers.\n- Additional fixed cameras are normally scene-edit\u002Fbake work. If a create prompt explicitly requires a policy\u002Fdataset camera, load the scene-edit camera and bake references and wire every surface in one pass.\n\n## Validation\n\nRun these static checks:\n\n```bash\npython -m py_compile \u003Cchanged-python-files>\nworkflows\u002Fagentic\u002Fpolicy\u002Frun.sh --list-envs\nworkflows\u002Fagentic\u002Farena\u002Frun.sh --env \u003Cenv> --dry-run\nworkflows\u002Fagentic\u002Fpolicy\u002Frun.sh --env \u003Cenv> --dry-run\n```\n\nThen run the real build\u002Fvisual gate from `skills\u002Fi4h-workflow-create\u002Freferences\u002Fcreate-validation.md`. `--dry-run` is shallow: it does not prove the task\u002Fassets instantiate or the scene is visually usable. Do not report a new env as ready until the bridge reaches ready, key objects are valid, a viewport capture has been inspected, and the bridge is stopped with:\n\n```bash\nworkflows\u002Fagentic\u002Farena\u002Fstop.sh --env \u003Cenv>\n```\n\nIf Isaac Sim cannot launch on the host, report that as a blocker and include the static check results; do not present static-only as success.\n\n## Hand Off to Scene Edit\n\nAfter the env shell passes create validation, use [[i4h-workflow-scene-edit]] for:\n\n- Adding, moving, resizing, or replacing objects\u002Fassets.\n- Adding fixed room\u002Foverhead\u002Fwrist cameras.\n- Live bridge edits.\n- Baking live changes into source.\n- Running `local-agent\u002Fvalidate-bake.sh`.\n\nThe scene-edit workflow owns object snippets, camera snippets, bridge endpoint details, and bake checklists so this create workflow stays small.\n\n## Prerequisites\n\n- Workflow setup has completed via [[i4h-workflow-setup]]; `.venv`, third-party checkouts, and Isaac Sim launch support are present.\n- The source env and target env id are known, or the prompt provides enough information to choose them from existing patterns.\n- Bridge validation runs on a GPU host that can launch Isaac Sim.\n\n## Limitations\n\n- Create one env per invocation.\n- Fork existing patterns; do not invent a new policy stack or shared framework.\n- Use scene-edit for post-create polish, optional props, camera baking, and source persistence of live edits.\n- `--dry-run` is not a substitute for the bridge build and visual validation gate.\n\n## Troubleshooting\n\n- **Missing setup**: if `.venv`, third-party checkouts, or `run.sh` entrypoints are missing, run [[i4h-workflow-setup]] first.\n- **Env not listed**: check that the YAML exists at `workflows\u002Fagentic\u002Fconfig\u002Fenvironments\u002F\u003Cenv>.yaml` and all files use the full `workflows\u002Fagentic\u002F` repo-root prefix.\n- **Build fails after static checks pass**: inspect the first Isaac\u002Farena stack trace; common causes are bad cfg kwargs, a helper defined below a config class, or an observation pointing at a missing sensor.\n- **Scene looks wrong**: fix source or use scene-edit to live-adjust and bake; do not report success from static checks alone.\n\n## Final Response\n\nReport the env id, source choices, files created, validation commands and results, capture paths, and any blocker. Do not commit changes unless the user explicitly asks.\n",{"data":32,"body":42},{"name":4,"version":33,"description":6,"license":23,"metadata":34},"0.7.0",{"author":35,"tags":36},"Isaac for Healthcare Team \u003Cisaac-for-healthcare-support@nvidia.com>",[37,38,39,40,41],"isaac-for-healthcare","i4h","agentic-workflow","environment","scaffolding",{"type":43,"children":44},"root",[45,54,61,76,82,87,420,426,431,490,495,546,552,557,664,677,683,712,717,722,732,738,807,813,818,943,964,999,1004,1010,1015,1050,1055,1061,1087,1093,1121,1127,1200,1206,1211],{"type":46,"tag":47,"props":48,"children":50},"element","h1",{"id":49},"i4h-workflow-create-env",[51],{"type":52,"value":53},"text","i4h Workflow - Create Env",{"type":46,"tag":55,"props":56,"children":58},"h2",{"id":57},"purpose",[59],{"type":52,"value":60},"Purpose",{"type":46,"tag":62,"props":63,"children":64},"p",{},[65,67,74],{"type":52,"value":66},"Create the first runnable version of a new ",{"type":46,"tag":68,"props":69,"children":71},"code",{"className":70},[],[72],{"type":52,"value":73},"workflows\u002Fagentic",{"type":52,"value":75}," environment by forking the closest existing env. Keep this skill focused on env scaffolding: YAML, assets, task, env class, runtime, and validation. Do not use this skill to polish an existing scene, add optional props\u002Fcameras, or bake bridge edits; use [[i4h-workflow-scene-edit]] for that.",{"type":46,"tag":55,"props":77,"children":79},{"id":78},"base-code",[80],{"type":52,"value":81},"Base Code",{"type":46,"tag":62,"props":83,"children":84},{},[85],{"type":52,"value":86},"Resolve and work from the i4h-workflows root:",{"type":46,"tag":88,"props":89,"children":94},"pre",{"className":90,"code":91,"language":92,"meta":93,"style":93},"language-bash shiki shiki-themes material-theme-lighter material-theme material-theme-palenight","ROOT=\"${I4H_WORKFLOWS:-$(git rev-parse --show-toplevel 2>\u002Fdev\u002Fnull)}\"\nif [ ! -d \"$ROOT\u002Fworkflows\u002Fagentic\" ]; then\n  ROOT=\"${I4H_WORKFLOWS:-$HOME\u002Fi4h-workflows}\"\n  [ -d \"$ROOT\u002Fworkflows\u002Fagentic\" ] || git clone https:\u002F\u002Fgithub.com\u002Fisaac-for-healthcare\u002Fi4h-workflows \"$ROOT\"\nfi\nexport I4H_WORKFLOWS=\"$ROOT\"; cd \"$ROOT\"\n","bash","",[95],{"type":46,"tag":68,"props":96,"children":97},{"__ignoreMap":93},[98,191,246,289,357,366],{"type":46,"tag":99,"props":100,"children":103},"span",{"class":101,"line":102},"line",1,[104,110,116,121,126,131,137,142,147,152,157,162,167,172,176,181,186],{"type":46,"tag":99,"props":105,"children":107},{"style":106},"--shiki-light:#90A4AE;--shiki-default:#EEFFFF;--shiki-dark:#BABED8",[108],{"type":52,"value":109},"ROOT",{"type":46,"tag":99,"props":111,"children":113},{"style":112},"--shiki-light:#39ADB5;--shiki-default:#89DDFF;--shiki-dark:#89DDFF",[114],{"type":52,"value":115},"=",{"type":46,"tag":99,"props":117,"children":118},{"style":112},[119],{"type":52,"value":120},"\"${",{"type":46,"tag":99,"props":122,"children":123},{"style":106},[124],{"type":52,"value":125},"I4H_WORKFLOWS",{"type":46,"tag":99,"props":127,"children":128},{"style":112},[129],{"type":52,"value":130},":-",{"type":46,"tag":99,"props":132,"children":134},{"style":133},"--shiki-light:#91B859;--shiki-default:#C3E88D;--shiki-dark:#C3E88D",[135],{"type":52,"value":136},"$(",{"type":46,"tag":99,"props":138,"children":139},{"style":106},[140],{"type":52,"value":141},"git",{"type":46,"tag":99,"props":143,"children":144},{"style":106},[145],{"type":52,"value":146}," rev-parse",{"type":46,"tag":99,"props":148,"children":149},{"style":106},[150],{"type":52,"value":151}," --show-toplevel",{"type":46,"tag":99,"props":153,"children":154},{"style":106},[155],{"type":52,"value":156}," 2",{"type":46,"tag":99,"props":158,"children":159},{"style":133},[160],{"type":52,"value":161},">",{"type":46,"tag":99,"props":163,"children":164},{"style":112},[165],{"type":52,"value":166},"\u002F",{"type":46,"tag":99,"props":168,"children":169},{"style":106},[170],{"type":52,"value":171},"dev",{"type":46,"tag":99,"props":173,"children":174},{"style":112},[175],{"type":52,"value":166},{"type":46,"tag":99,"props":177,"children":178},{"style":106},[179],{"type":52,"value":180},"null",{"type":46,"tag":99,"props":182,"children":183},{"style":133},[184],{"type":52,"value":185},")",{"type":46,"tag":99,"props":187,"children":188},{"style":112},[189],{"type":52,"value":190},"}\"\n",{"type":46,"tag":99,"props":192,"children":194},{"class":101,"line":193},2,[195,201,206,211,216,221,226,231,236,241],{"type":46,"tag":99,"props":196,"children":198},{"style":197},"--shiki-light:#39ADB5;--shiki-light-font-style:italic;--shiki-default:#89DDFF;--shiki-default-font-style:italic;--shiki-dark:#89DDFF;--shiki-dark-font-style:italic",[199],{"type":52,"value":200},"if",{"type":46,"tag":99,"props":202,"children":203},{"style":112},[204],{"type":52,"value":205}," [",{"type":46,"tag":99,"props":207,"children":208},{"style":112},[209],{"type":52,"value":210}," !",{"type":46,"tag":99,"props":212,"children":213},{"style":112},[214],{"type":52,"value":215}," -d",{"type":46,"tag":99,"props":217,"children":218},{"style":112},[219],{"type":52,"value":220}," \"",{"type":46,"tag":99,"props":222,"children":223},{"style":106},[224],{"type":52,"value":225},"$ROOT",{"type":46,"tag":99,"props":227,"children":228},{"style":133},[229],{"type":52,"value":230},"\u002Fworkflows\u002Fagentic",{"type":46,"tag":99,"props":232,"children":233},{"style":112},[234],{"type":52,"value":235},"\"",{"type":46,"tag":99,"props":237,"children":238},{"style":112},[239],{"type":52,"value":240}," ];",{"type":46,"tag":99,"props":242,"children":243},{"style":197},[244],{"type":52,"value":245}," then\n",{"type":46,"tag":99,"props":247,"children":249},{"class":101,"line":248},3,[250,255,259,263,267,271,276,280,285],{"type":46,"tag":99,"props":251,"children":252},{"style":106},[253],{"type":52,"value":254},"  ROOT",{"type":46,"tag":99,"props":256,"children":257},{"style":112},[258],{"type":52,"value":115},{"type":46,"tag":99,"props":260,"children":261},{"style":112},[262],{"type":52,"value":120},{"type":46,"tag":99,"props":264,"children":265},{"style":106},[266],{"type":52,"value":125},{"type":46,"tag":99,"props":268,"children":269},{"style":112},[270],{"type":52,"value":130},{"type":46,"tag":99,"props":272,"children":273},{"style":106},[274],{"type":52,"value":275},"$HOME",{"type":46,"tag":99,"props":277,"children":278},{"style":112},[279],{"type":52,"value":166},{"type":46,"tag":99,"props":281,"children":282},{"style":106},[283],{"type":52,"value":284},"i4h-workflows",{"type":46,"tag":99,"props":286,"children":287},{"style":112},[288],{"type":52,"value":190},{"type":46,"tag":99,"props":290,"children":292},{"class":101,"line":291},4,[293,298,302,306,310,314,318,323,328,334,339,344,348,352],{"type":46,"tag":99,"props":294,"children":295},{"style":112},[296],{"type":52,"value":297},"  [",{"type":46,"tag":99,"props":299,"children":300},{"style":112},[301],{"type":52,"value":215},{"type":46,"tag":99,"props":303,"children":304},{"style":112},[305],{"type":52,"value":220},{"type":46,"tag":99,"props":307,"children":308},{"style":106},[309],{"type":52,"value":225},{"type":46,"tag":99,"props":311,"children":312},{"style":133},[313],{"type":52,"value":230},{"type":46,"tag":99,"props":315,"children":316},{"style":112},[317],{"type":52,"value":235},{"type":46,"tag":99,"props":319,"children":320},{"style":112},[321],{"type":52,"value":322}," ]",{"type":46,"tag":99,"props":324,"children":325},{"style":112},[326],{"type":52,"value":327}," ||",{"type":46,"tag":99,"props":329,"children":331},{"style":330},"--shiki-light:#E2931D;--shiki-default:#FFCB6B;--shiki-dark:#FFCB6B",[332],{"type":52,"value":333}," git",{"type":46,"tag":99,"props":335,"children":336},{"style":133},[337],{"type":52,"value":338}," clone",{"type":46,"tag":99,"props":340,"children":341},{"style":133},[342],{"type":52,"value":343}," https:\u002F\u002Fgithub.com\u002Fisaac-for-healthcare\u002Fi4h-workflows",{"type":46,"tag":99,"props":345,"children":346},{"style":112},[347],{"type":52,"value":220},{"type":46,"tag":99,"props":349,"children":350},{"style":106},[351],{"type":52,"value":225},{"type":46,"tag":99,"props":353,"children":354},{"style":112},[355],{"type":52,"value":356},"\"\n",{"type":46,"tag":99,"props":358,"children":360},{"class":101,"line":359},5,[361],{"type":46,"tag":99,"props":362,"children":363},{"style":197},[364],{"type":52,"value":365},"fi\n",{"type":46,"tag":99,"props":367,"children":369},{"class":101,"line":368},6,[370,376,381,385,389,393,397,402,408,412,416],{"type":46,"tag":99,"props":371,"children":373},{"style":372},"--shiki-light:#9C3EDA;--shiki-default:#C792EA;--shiki-dark:#C792EA",[374],{"type":52,"value":375},"export",{"type":46,"tag":99,"props":377,"children":378},{"style":106},[379],{"type":52,"value":380}," I4H_WORKFLOWS",{"type":46,"tag":99,"props":382,"children":383},{"style":112},[384],{"type":52,"value":115},{"type":46,"tag":99,"props":386,"children":387},{"style":112},[388],{"type":52,"value":235},{"type":46,"tag":99,"props":390,"children":391},{"style":106},[392],{"type":52,"value":225},{"type":46,"tag":99,"props":394,"children":395},{"style":112},[396],{"type":52,"value":235},{"type":46,"tag":99,"props":398,"children":399},{"style":112},[400],{"type":52,"value":401},";",{"type":46,"tag":99,"props":403,"children":405},{"style":404},"--shiki-light:#6182B8;--shiki-default:#82AAFF;--shiki-dark:#82AAFF",[406],{"type":52,"value":407}," cd",{"type":46,"tag":99,"props":409,"children":410},{"style":112},[411],{"type":52,"value":220},{"type":46,"tag":99,"props":413,"children":414},{"style":106},[415],{"type":52,"value":225},{"type":46,"tag":99,"props":417,"children":418},{"style":112},[419],{"type":52,"value":356},{"type":46,"tag":55,"props":421,"children":423},{"id":422},"what-to-load",[424],{"type":52,"value":425},"What to Load",{"type":46,"tag":62,"props":427,"children":428},{},[429],{"type":52,"value":430},"Before editing, load only the references that match the request:",{"type":46,"tag":432,"props":433,"children":434},"ul",{},[435,449,460,473,485],{"type":46,"tag":436,"props":437,"children":438},"li",{},[439,441,447],{"type":52,"value":440},"Always load ",{"type":46,"tag":68,"props":442,"children":444},{"className":443},[],[445],{"type":52,"value":446},"skills\u002Fi4h-workflow\u002Freferences\u002Frepo-map.md",{"type":52,"value":448},".",{"type":46,"tag":436,"props":450,"children":451},{},[452,453,459],{"type":52,"value":440},{"type":46,"tag":68,"props":454,"children":456},{"className":455},[],[457],{"type":52,"value":458},"skills\u002Fi4h-workflow-create\u002Freferences\u002Fcreate-contract.md",{"type":52,"value":448},{"type":46,"tag":436,"props":461,"children":462},{},[463,465,471],{"type":52,"value":464},"Load ",{"type":46,"tag":68,"props":466,"children":468},{"className":467},[],[469],{"type":52,"value":470},"skills\u002Fi4h-workflow-create\u002Freferences\u002Fenv-authoring-patterns.md",{"type":52,"value":472}," to choose the source env, robot owner, policy stack, and YAML pattern.",{"type":46,"tag":436,"props":474,"children":475},{},[476,477,483],{"type":52,"value":464},{"type":46,"tag":68,"props":478,"children":480},{"className":479},[],[481],{"type":52,"value":482},"skills\u002Fi4h-workflow-create\u002Freferences\u002Fhybrid-layout-rules.md",{"type":52,"value":484}," only for hybrid scene+robot envs, G1 footprint\u002Ftable-height work, catalog USD scale, or support-surface layout.",{"type":46,"tag":436,"props":486,"children":487},{},[488],{"type":52,"value":489},"Load a recipe file only when the prompt matches it exactly enough to avoid re-deciding components.",{"type":46,"tag":62,"props":491,"children":492},{},[493],{"type":52,"value":494},"Recipe routing:",{"type":46,"tag":496,"props":497,"children":498},"table",{},[499,518],{"type":46,"tag":500,"props":501,"children":502},"thead",{},[503],{"type":46,"tag":504,"props":505,"children":506},"tr",{},[507,513],{"type":46,"tag":508,"props":509,"children":510},"th",{},[511],{"type":52,"value":512},"Prompt shape",{"type":46,"tag":508,"props":514,"children":515},{},[516],{"type":52,"value":517},"Reference",{"type":46,"tag":519,"props":520,"children":521},"tbody",{},[522],{"type":46,"tag":504,"props":523,"children":524},{},[525,537],{"type":46,"tag":526,"props":527,"children":528},"td",{},[529,531],{"type":52,"value":530},"Surgical tool sorting using G1 based on ",{"type":46,"tag":68,"props":532,"children":534},{"className":533},[],[535],{"type":52,"value":536},"scissor_pick_and_place",{"type":46,"tag":526,"props":538,"children":539},{},[540],{"type":46,"tag":68,"props":541,"children":543},{"className":542},[],[544],{"type":52,"value":545},"skills\u002Fi4h-workflow-create\u002Freferences\u002Fg1-surgical-tool-sort.md",{"type":46,"tag":55,"props":547,"children":549},{"id":548},"create-contract",[550],{"type":52,"value":551},"Create Contract",{"type":46,"tag":62,"props":553,"children":554},{},[555],{"type":52,"value":556},"Create the env shell first. A normal create task should produce exactly these surfaces unless the chosen existing pattern requires an explicit exception:",{"type":46,"tag":496,"props":558,"children":559},{},[560,576],{"type":46,"tag":500,"props":561,"children":562},{},[563],{"type":46,"tag":504,"props":564,"children":565},{},[566,571],{"type":46,"tag":508,"props":567,"children":568},{},[569],{"type":52,"value":570},"Surface",{"type":46,"tag":508,"props":572,"children":573},{},[574],{"type":52,"value":575},"Path",{"type":46,"tag":519,"props":577,"children":578},{},[579,596,613,630,647],{"type":46,"tag":504,"props":580,"children":581},{},[582,587],{"type":46,"tag":526,"props":583,"children":584},{},[585],{"type":52,"value":586},"Env YAML",{"type":46,"tag":526,"props":588,"children":589},{},[590],{"type":46,"tag":68,"props":591,"children":593},{"className":592},[],[594],{"type":52,"value":595},"workflows\u002Fagentic\u002Fconfig\u002Fenvironments\u002F\u003Cenv>.yaml",{"type":46,"tag":504,"props":597,"children":598},{},[599,604],{"type":46,"tag":526,"props":600,"children":601},{},[602],{"type":52,"value":603},"Assets",{"type":46,"tag":526,"props":605,"children":606},{},[607],{"type":46,"tag":68,"props":608,"children":610},{"className":609},[],[611],{"type":52,"value":612},"workflows\u002Fagentic\u002Farena\u002Farena\u002Fassets\u002F\u003Cenv>.py",{"type":46,"tag":504,"props":614,"children":615},{},[616,621],{"type":46,"tag":526,"props":617,"children":618},{},[619],{"type":52,"value":620},"Task",{"type":46,"tag":526,"props":622,"children":623},{},[624],{"type":46,"tag":68,"props":625,"children":627},{"className":626},[],[628],{"type":52,"value":629},"workflows\u002Fagentic\u002Farena\u002Farena\u002Ftasks\u002F\u003Cenv>.py",{"type":46,"tag":504,"props":631,"children":632},{},[633,638],{"type":46,"tag":526,"props":634,"children":635},{},[636],{"type":52,"value":637},"Env class",{"type":46,"tag":526,"props":639,"children":640},{},[641],{"type":46,"tag":68,"props":642,"children":644},{"className":643},[],[645],{"type":52,"value":646},"workflows\u002Fagentic\u002Farena\u002Farena\u002Fenvironments\u002F\u003Cenv>_environment.py",{"type":46,"tag":504,"props":648,"children":649},{},[650,655],{"type":46,"tag":526,"props":651,"children":652},{},[653],{"type":52,"value":654},"Runtime",{"type":46,"tag":526,"props":656,"children":657},{},[658],{"type":46,"tag":68,"props":659,"children":661},{"className":660},[],[662],{"type":52,"value":663},"workflows\u002Fagentic\u002Farena\u002Farena\u002Fruntimes\u002F\u003Cenv>.py",{"type":46,"tag":62,"props":665,"children":666},{},[667,669,675],{"type":52,"value":668},"Keep all paths repo-root relative and keep the ",{"type":46,"tag":68,"props":670,"children":672},{"className":671},[],[673],{"type":52,"value":674},"workflows\u002Fagentic\u002F",{"type":52,"value":676}," prefix. Do not create a new policy package, README, docs, or shared-module edits unless the user explicitly asks and the policy stack truly needs it.",{"type":46,"tag":55,"props":678,"children":680},{"id":679},"workflow",[681],{"type":52,"value":682},"Workflow",{"type":46,"tag":684,"props":685,"children":686},"ol",{},[687,692,697,702,707],{"type":46,"tag":436,"props":688,"children":689},{},[690],{"type":52,"value":691},"Resolve components: env id, source env, scene\u002Fassets source, robot owner, policy stack, model\u002Fcheckpoint, cameras, objects\u002Fdestinations, success rule.",{"type":46,"tag":436,"props":693,"children":694},{},[695],{"type":52,"value":696},"Inspect the selected source env YAML, env class, assets, task, runtime, and policy stack files before writing code.",{"type":46,"tag":436,"props":698,"children":699},{},[700],{"type":52,"value":701},"Fork the closest working pattern. Preserve inline-scene vs registry-asset style; do not invent a new architecture.",{"type":46,"tag":436,"props":703,"children":704},{},[705],{"type":52,"value":706},"Keep the first version minimal and runnable. Optional visual polish, extra props, new cameras, and layout changes happen through scene-edit after the env exists.",{"type":46,"tag":436,"props":708,"children":709},{},[710],{"type":52,"value":711},"Run validation and fix source until the env builds and the rendered scene is visually sane.",{"type":46,"tag":62,"props":713,"children":714},{},[715],{"type":52,"value":716},"For an exact recipe match, do not re-derive the architecture after loading the recipe. Read only the source files the recipe names, skip large source runtime files when the recipe says the runtime is a re-export, then write the five contract files immediately. Do not inspect third-party framework internals unless a validation error requires it.",{"type":46,"tag":62,"props":718,"children":719},{},[720],{"type":52,"value":721},"Plan shape:",{"type":46,"tag":88,"props":723,"children":727},{"className":724,"code":726,"language":52,"meta":93},[725],"language-text","Env id:\nSource env \u002F recipe:\nScene\u002Fassets source:\nRobot owner:\nPolicy stack + model\u002Fcheckpoint:\nObjects\u002Fdestinations:\nSuccess rule:\nFiles to create:\nValidation:\n",[728],{"type":46,"tag":68,"props":729,"children":730},{"__ignoreMap":93},[731],{"type":52,"value":726},{"type":46,"tag":55,"props":733,"children":735},{"id":734},"hard-rules",[736],{"type":52,"value":737},"Hard Rules",{"type":46,"tag":432,"props":739,"children":740},{},[741,746,751,756,776,789,802],{"type":46,"tag":436,"props":742,"children":743},{},[744],{"type":52,"value":745},"Env YAML is the source of truth for robot, policy, cameras, task text, dataset mapping, and train defaults.",{"type":46,"tag":436,"props":747,"children":748},{},[749],{"type":52,"value":750},"Fork from the nearest existing implementation. For hybrid envs, robot integration comes from the robot owner and scene construction comes from the scene source.",{"type":46,"tag":436,"props":752,"children":753},{},[754],{"type":52,"value":755},"Sorting tasks need at least two object types, at least two destinations, and a success rule that fails swapped placements.",{"type":46,"tag":436,"props":757,"children":758},{},[759,761,767,769,775],{"type":52,"value":760},"Static\u002Fkinematic destination props may not expose ",{"type":46,"tag":68,"props":762,"children":764},{"className":763},[],[765],{"type":52,"value":766},".data.root_pos_w",{"type":52,"value":768},"; success checks must fall back to ",{"type":46,"tag":68,"props":770,"children":772},{"className":771},[],[773],{"type":52,"value":774},"entity.get_world_poses()",{"type":52,"value":448},{"type":46,"tag":436,"props":777,"children":778},{},[779,781,787],{"type":52,"value":780},"Functions referenced by ",{"type":46,"tag":68,"props":782,"children":784},{"className":783},[],[785],{"type":52,"value":786},"func=",{"type":52,"value":788}," inside config classes must be defined above those classes.",{"type":46,"tag":436,"props":790,"children":791},{},[792,794,800],{"type":52,"value":793},"G1 WBC envs need footprint clearance and matched ground\u002Fbase-height values; load ",{"type":46,"tag":68,"props":795,"children":797},{"className":796},[],[798],{"type":52,"value":799},"hybrid-layout-rules.md",{"type":52,"value":801}," before choosing those numbers.",{"type":46,"tag":436,"props":803,"children":804},{},[805],{"type":52,"value":806},"Additional fixed cameras are normally scene-edit\u002Fbake work. If a create prompt explicitly requires a policy\u002Fdataset camera, load the scene-edit camera and bake references and wire every surface in one pass.",{"type":46,"tag":55,"props":808,"children":810},{"id":809},"validation",[811],{"type":52,"value":812},"Validation",{"type":46,"tag":62,"props":814,"children":815},{},[816],{"type":52,"value":817},"Run these static checks:",{"type":46,"tag":88,"props":819,"children":821},{"className":90,"code":820,"language":92,"meta":93,"style":93},"python -m py_compile \u003Cchanged-python-files>\nworkflows\u002Fagentic\u002Fpolicy\u002Frun.sh --list-envs\nworkflows\u002Fagentic\u002Farena\u002Frun.sh --env \u003Cenv> --dry-run\nworkflows\u002Fagentic\u002Fpolicy\u002Frun.sh --env \u003Cenv> --dry-run\n",[822],{"type":46,"tag":68,"props":823,"children":824},{"__ignoreMap":93},[825,863,876,912],{"type":46,"tag":99,"props":826,"children":827},{"class":101,"line":102},[828,833,838,843,848,853,858],{"type":46,"tag":99,"props":829,"children":830},{"style":330},[831],{"type":52,"value":832},"python",{"type":46,"tag":99,"props":834,"children":835},{"style":133},[836],{"type":52,"value":837}," -m",{"type":46,"tag":99,"props":839,"children":840},{"style":133},[841],{"type":52,"value":842}," py_compile",{"type":46,"tag":99,"props":844,"children":845},{"style":112},[846],{"type":52,"value":847}," \u003C",{"type":46,"tag":99,"props":849,"children":850},{"style":133},[851],{"type":52,"value":852},"changed-python-file",{"type":46,"tag":99,"props":854,"children":855},{"style":106},[856],{"type":52,"value":857},"s",{"type":46,"tag":99,"props":859,"children":860},{"style":112},[861],{"type":52,"value":862},">\n",{"type":46,"tag":99,"props":864,"children":865},{"class":101,"line":193},[866,871],{"type":46,"tag":99,"props":867,"children":868},{"style":330},[869],{"type":52,"value":870},"workflows\u002Fagentic\u002Fpolicy\u002Frun.sh",{"type":46,"tag":99,"props":872,"children":873},{"style":133},[874],{"type":52,"value":875}," --list-envs\n",{"type":46,"tag":99,"props":877,"children":878},{"class":101,"line":248},[879,884,889,893,898,903,907],{"type":46,"tag":99,"props":880,"children":881},{"style":330},[882],{"type":52,"value":883},"workflows\u002Fagentic\u002Farena\u002Frun.sh",{"type":46,"tag":99,"props":885,"children":886},{"style":133},[887],{"type":52,"value":888}," --env",{"type":46,"tag":99,"props":890,"children":891},{"style":112},[892],{"type":52,"value":847},{"type":46,"tag":99,"props":894,"children":895},{"style":133},[896],{"type":52,"value":897},"en",{"type":46,"tag":99,"props":899,"children":900},{"style":106},[901],{"type":52,"value":902},"v",{"type":46,"tag":99,"props":904,"children":905},{"style":112},[906],{"type":52,"value":161},{"type":46,"tag":99,"props":908,"children":909},{"style":133},[910],{"type":52,"value":911}," --dry-run\n",{"type":46,"tag":99,"props":913,"children":914},{"class":101,"line":291},[915,919,923,927,931,935,939],{"type":46,"tag":99,"props":916,"children":917},{"style":330},[918],{"type":52,"value":870},{"type":46,"tag":99,"props":920,"children":921},{"style":133},[922],{"type":52,"value":888},{"type":46,"tag":99,"props":924,"children":925},{"style":112},[926],{"type":52,"value":847},{"type":46,"tag":99,"props":928,"children":929},{"style":133},[930],{"type":52,"value":897},{"type":46,"tag":99,"props":932,"children":933},{"style":106},[934],{"type":52,"value":902},{"type":46,"tag":99,"props":936,"children":937},{"style":112},[938],{"type":52,"value":161},{"type":46,"tag":99,"props":940,"children":941},{"style":133},[942],{"type":52,"value":911},{"type":46,"tag":62,"props":944,"children":945},{},[946,948,954,956,962],{"type":52,"value":947},"Then run the real build\u002Fvisual gate from ",{"type":46,"tag":68,"props":949,"children":951},{"className":950},[],[952],{"type":52,"value":953},"skills\u002Fi4h-workflow-create\u002Freferences\u002Fcreate-validation.md",{"type":52,"value":955},". ",{"type":46,"tag":68,"props":957,"children":959},{"className":958},[],[960],{"type":52,"value":961},"--dry-run",{"type":52,"value":963}," is shallow: it does not prove the task\u002Fassets instantiate or the scene is visually usable. Do not report a new env as ready until the bridge reaches ready, key objects are valid, a viewport capture has been inspected, and the bridge is stopped with:",{"type":46,"tag":88,"props":965,"children":967},{"className":90,"code":966,"language":92,"meta":93,"style":93},"workflows\u002Fagentic\u002Farena\u002Fstop.sh --env \u003Cenv>\n",[968],{"type":46,"tag":68,"props":969,"children":970},{"__ignoreMap":93},[971],{"type":46,"tag":99,"props":972,"children":973},{"class":101,"line":102},[974,979,983,987,991,995],{"type":46,"tag":99,"props":975,"children":976},{"style":330},[977],{"type":52,"value":978},"workflows\u002Fagentic\u002Farena\u002Fstop.sh",{"type":46,"tag":99,"props":980,"children":981},{"style":133},[982],{"type":52,"value":888},{"type":46,"tag":99,"props":984,"children":985},{"style":112},[986],{"type":52,"value":847},{"type":46,"tag":99,"props":988,"children":989},{"style":133},[990],{"type":52,"value":897},{"type":46,"tag":99,"props":992,"children":993},{"style":106},[994],{"type":52,"value":902},{"type":46,"tag":99,"props":996,"children":997},{"style":112},[998],{"type":52,"value":862},{"type":46,"tag":62,"props":1000,"children":1001},{},[1002],{"type":52,"value":1003},"If Isaac Sim cannot launch on the host, report that as a blocker and include the static check results; do not present static-only as success.",{"type":46,"tag":55,"props":1005,"children":1007},{"id":1006},"hand-off-to-scene-edit",[1008],{"type":52,"value":1009},"Hand Off to Scene Edit",{"type":46,"tag":62,"props":1011,"children":1012},{},[1013],{"type":52,"value":1014},"After the env shell passes create validation, use [[i4h-workflow-scene-edit]] for:",{"type":46,"tag":432,"props":1016,"children":1017},{},[1018,1023,1028,1033,1038],{"type":46,"tag":436,"props":1019,"children":1020},{},[1021],{"type":52,"value":1022},"Adding, moving, resizing, or replacing objects\u002Fassets.",{"type":46,"tag":436,"props":1024,"children":1025},{},[1026],{"type":52,"value":1027},"Adding fixed room\u002Foverhead\u002Fwrist cameras.",{"type":46,"tag":436,"props":1029,"children":1030},{},[1031],{"type":52,"value":1032},"Live bridge edits.",{"type":46,"tag":436,"props":1034,"children":1035},{},[1036],{"type":52,"value":1037},"Baking live changes into source.",{"type":46,"tag":436,"props":1039,"children":1040},{},[1041,1043,1049],{"type":52,"value":1042},"Running ",{"type":46,"tag":68,"props":1044,"children":1046},{"className":1045},[],[1047],{"type":52,"value":1048},"local-agent\u002Fvalidate-bake.sh",{"type":52,"value":448},{"type":46,"tag":62,"props":1051,"children":1052},{},[1053],{"type":52,"value":1054},"The scene-edit workflow owns object snippets, camera snippets, bridge endpoint details, and bake checklists so this create workflow stays small.",{"type":46,"tag":55,"props":1056,"children":1058},{"id":1057},"prerequisites",[1059],{"type":52,"value":1060},"Prerequisites",{"type":46,"tag":432,"props":1062,"children":1063},{},[1064,1077,1082],{"type":46,"tag":436,"props":1065,"children":1066},{},[1067,1069,1075],{"type":52,"value":1068},"Workflow setup has completed via [[i4h-workflow-setup]]; ",{"type":46,"tag":68,"props":1070,"children":1072},{"className":1071},[],[1073],{"type":52,"value":1074},".venv",{"type":52,"value":1076},", third-party checkouts, and Isaac Sim launch support are present.",{"type":46,"tag":436,"props":1078,"children":1079},{},[1080],{"type":52,"value":1081},"The source env and target env id are known, or the prompt provides enough information to choose them from existing patterns.",{"type":46,"tag":436,"props":1083,"children":1084},{},[1085],{"type":52,"value":1086},"Bridge validation runs on a GPU host that can launch Isaac Sim.",{"type":46,"tag":55,"props":1088,"children":1090},{"id":1089},"limitations",[1091],{"type":52,"value":1092},"Limitations",{"type":46,"tag":432,"props":1094,"children":1095},{},[1096,1101,1106,1111],{"type":46,"tag":436,"props":1097,"children":1098},{},[1099],{"type":52,"value":1100},"Create one env per invocation.",{"type":46,"tag":436,"props":1102,"children":1103},{},[1104],{"type":52,"value":1105},"Fork existing patterns; do not invent a new policy stack or shared framework.",{"type":46,"tag":436,"props":1107,"children":1108},{},[1109],{"type":52,"value":1110},"Use scene-edit for post-create polish, optional props, camera baking, and source persistence of live edits.",{"type":46,"tag":436,"props":1112,"children":1113},{},[1114,1119],{"type":46,"tag":68,"props":1115,"children":1117},{"className":1116},[],[1118],{"type":52,"value":961},{"type":52,"value":1120}," is not a substitute for the bridge build and visual validation gate.",{"type":46,"tag":55,"props":1122,"children":1124},{"id":1123},"troubleshooting",[1125],{"type":52,"value":1126},"Troubleshooting",{"type":46,"tag":432,"props":1128,"children":1129},{},[1130,1156,1180,1190],{"type":46,"tag":436,"props":1131,"children":1132},{},[1133,1139,1141,1146,1148,1154],{"type":46,"tag":1134,"props":1135,"children":1136},"strong",{},[1137],{"type":52,"value":1138},"Missing setup",{"type":52,"value":1140},": if ",{"type":46,"tag":68,"props":1142,"children":1144},{"className":1143},[],[1145],{"type":52,"value":1074},{"type":52,"value":1147},", third-party checkouts, or ",{"type":46,"tag":68,"props":1149,"children":1151},{"className":1150},[],[1152],{"type":52,"value":1153},"run.sh",{"type":52,"value":1155}," entrypoints are missing, run [[i4h-workflow-setup]] first.",{"type":46,"tag":436,"props":1157,"children":1158},{},[1159,1164,1166,1171,1173,1178],{"type":46,"tag":1134,"props":1160,"children":1161},{},[1162],{"type":52,"value":1163},"Env not listed",{"type":52,"value":1165},": check that the YAML exists at ",{"type":46,"tag":68,"props":1167,"children":1169},{"className":1168},[],[1170],{"type":52,"value":595},{"type":52,"value":1172}," and all files use the full ",{"type":46,"tag":68,"props":1174,"children":1176},{"className":1175},[],[1177],{"type":52,"value":674},{"type":52,"value":1179}," repo-root prefix.",{"type":46,"tag":436,"props":1181,"children":1182},{},[1183,1188],{"type":46,"tag":1134,"props":1184,"children":1185},{},[1186],{"type":52,"value":1187},"Build fails after static checks pass",{"type":52,"value":1189},": inspect the first Isaac\u002Farena stack trace; common causes are bad cfg kwargs, a helper defined below a config class, or an observation pointing at a missing sensor.",{"type":46,"tag":436,"props":1191,"children":1192},{},[1193,1198],{"type":46,"tag":1134,"props":1194,"children":1195},{},[1196],{"type":52,"value":1197},"Scene looks wrong",{"type":52,"value":1199},": fix source or use scene-edit to live-adjust and bake; do not report success from static checks alone.",{"type":46,"tag":55,"props":1201,"children":1203},{"id":1202},"final-response",[1204],{"type":52,"value":1205},"Final Response",{"type":46,"tag":62,"props":1207,"children":1208},{},[1209],{"type":52,"value":1210},"Report the env id, source choices, files created, validation commands and results, capture paths, and any blocker. Do not commit changes unless the user explicitly asks.",{"type":46,"tag":1212,"props":1213,"children":1214},"style",{},[1215],{"type":52,"value":1216},"html .light .shiki span {color: var(--shiki-light);background: var(--shiki-light-bg);font-style: var(--shiki-light-font-style);font-weight: var(--shiki-light-font-weight);text-decoration: var(--shiki-light-text-decoration);}html.light .shiki span {color: var(--shiki-light);background: var(--shiki-light-bg);font-style: var(--shiki-light-font-style);font-weight: var(--shiki-light-font-weight);text-decoration: var(--shiki-light-text-decoration);}html .default .shiki span {color: var(--shiki-default);background: var(--shiki-default-bg);font-style: var(--shiki-default-font-style);font-weight: var(--shiki-default-font-weight);text-decoration: var(--shiki-default-text-decoration);}html .shiki span {color: var(--shiki-default);background: var(--shiki-default-bg);font-style: var(--shiki-default-font-style);font-weight: var(--shiki-default-font-weight);text-decoration: var(--shiki-default-text-decoration);}html .dark .shiki span {color: var(--shiki-dark);background: var(--shiki-dark-bg);font-style: var(--shiki-dark-font-style);font-weight: var(--shiki-dark-font-weight);text-decoration: var(--shiki-dark-text-decoration);}html.dark .shiki span {color: var(--shiki-dark);background: var(--shiki-dark-bg);font-style: var(--shiki-dark-font-style);font-weight: var(--shiki-dark-font-weight);text-decoration: var(--shiki-dark-text-decoration);}",{"items":1218,"total":1376},[1219,1237,1254,1265,1277,1291,1304,1318,1331,1342,1356,1365],{"slug":1220,"name":1220,"fn":1221,"description":1222,"org":1223,"tags":1224,"stars":1234,"repoUrl":1235,"updatedAt":1236},"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},[1225,1228,1231],{"name":1226,"slug":1227,"type":15},"Documentation","documentation",{"name":1229,"slug":1230,"type":15},"MCP","mcp",{"name":1232,"slug":1233,"type":15},"Search","search",21777,"https:\u002F\u002Fgithub.com\u002FNVIDIA\u002FNemoClaw","2026-07-20T06:00:01.461044",{"slug":1238,"name":1238,"fn":1239,"description":1240,"org":1241,"tags":1242,"stars":1251,"repoUrl":1252,"updatedAt":1253},"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},[1243,1246,1249],{"name":1244,"slug":1245,"type":15},"Containers","containers",{"name":1247,"slug":1248,"type":15},"Deployment","deployment",{"name":1250,"slug":832,"type":15},"Python",17049,"https:\u002F\u002Fgithub.com\u002FNVIDIA\u002FMegatron-LM","2026-07-27T06:06:11.249662",{"slug":1255,"name":1255,"fn":1256,"description":1257,"org":1258,"tags":1259,"stars":1251,"repoUrl":1252,"updatedAt":1264},"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},[1260,1263],{"name":1261,"slug":1262,"type":15},"CI\u002FCD","ci-cd",{"name":1247,"slug":1248,"type":15},"2026-07-14T05:25:59.97109",{"slug":1266,"name":1266,"fn":1267,"description":1268,"org":1269,"tags":1270,"stars":1251,"repoUrl":1252,"updatedAt":1276},"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},[1271,1272,1273],{"name":1261,"slug":1262,"type":15},{"name":1247,"slug":1248,"type":15},{"name":1274,"slug":1275,"type":15},"GitHub","github","2026-07-27T06:06:12.278222",{"slug":1278,"name":1278,"fn":1279,"description":1280,"org":1281,"tags":1282,"stars":1251,"repoUrl":1252,"updatedAt":1290},"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},[1283,1286,1287],{"name":1284,"slug":1285,"type":15},"Debugging","debugging",{"name":1274,"slug":1275,"type":15},{"name":1288,"slug":1289,"type":15},"Triage","triage","2026-07-14T05:25:57.442089",{"slug":1292,"name":1292,"fn":1293,"description":1294,"org":1295,"tags":1296,"stars":1251,"repoUrl":1252,"updatedAt":1303},"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},[1297,1300],{"name":1298,"slug":1299,"type":15},"Best Practices","best-practices",{"name":1301,"slug":1302,"type":15},"Code Analysis","code-analysis","2026-07-14T05:25:56.18433",{"slug":1305,"name":1305,"fn":1306,"description":1307,"org":1308,"tags":1309,"stars":1251,"repoUrl":1252,"updatedAt":1317},"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},[1310,1313,1316],{"name":1311,"slug":1312,"type":15},"Machine Learning","machine-learning",{"name":1314,"slug":1315,"type":15},"Migration","migration",{"name":9,"slug":8,"type":15},"2026-07-17T06:07:11.777011",{"slug":1319,"name":1319,"fn":1320,"description":1321,"org":1322,"tags":1323,"stars":1251,"repoUrl":1252,"updatedAt":1330},"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},[1324,1327],{"name":1325,"slug":1326,"type":15},"QA","qa",{"name":1328,"slug":1329,"type":15},"Testing","testing","2026-07-14T05:25:53.673039",{"slug":1332,"name":1332,"fn":1333,"description":1334,"org":1335,"tags":1336,"stars":1251,"repoUrl":1252,"updatedAt":1341},"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},[1337,1338],{"name":1247,"slug":1248,"type":15},{"name":1339,"slug":1340,"type":15},"Infrastructure","infrastructure","2026-07-14T05:25:49.362534",{"slug":1343,"name":1343,"fn":1344,"description":1345,"org":1346,"tags":1347,"stars":1251,"repoUrl":1252,"updatedAt":1355},"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},[1348,1351,1352],{"name":1349,"slug":1350,"type":15},"Code Review","code-review",{"name":1274,"slug":1275,"type":15},{"name":1353,"slug":1354,"type":15},"Pull Requests","pull-requests","2026-07-14T05:26:01.226578",{"slug":1357,"name":1357,"fn":1358,"description":1359,"org":1360,"tags":1361,"stars":1251,"repoUrl":1252,"updatedAt":1364},"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},[1362,1363],{"name":1325,"slug":1326,"type":15},{"name":1328,"slug":1329,"type":15},"2026-07-14T05:25:54.928983",{"slug":1366,"name":1366,"fn":1367,"description":1368,"org":1369,"tags":1370,"stars":1251,"repoUrl":1252,"updatedAt":1375},"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},[1371,1374],{"name":1372,"slug":1373,"type":15},"Automation","automation",{"name":1261,"slug":1262,"type":15},"2026-07-30T05:29:03.275638",496,{"items":1378,"total":1474},[1379,1396,1406,1420,1430,1445,1460],{"slug":1380,"name":1380,"fn":1381,"description":1382,"org":1383,"tags":1384,"stars":20,"repoUrl":21,"updatedAt":1395},"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},[1385,1388,1391,1392],{"name":1386,"slug":1387,"type":15},"Data Analysis","data-analysis",{"name":1389,"slug":1390,"type":15},"Data Engineering","data-engineering",{"name":9,"slug":8,"type":15},{"name":1393,"slug":1394,"type":15},"Performance","performance","2026-07-14T05:28:43.176466",{"slug":1397,"name":1397,"fn":1398,"description":1399,"org":1400,"tags":1401,"stars":20,"repoUrl":21,"updatedAt":1405},"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},[1402,1403,1404],{"name":1247,"slug":1248,"type":15},{"name":1339,"slug":1340,"type":15},{"name":9,"slug":8,"type":15},"2026-07-14T05:29:06.667109",{"slug":1407,"name":1407,"fn":1408,"description":1409,"org":1410,"tags":1411,"stars":20,"repoUrl":21,"updatedAt":1419},"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},[1412,1415,1416],{"name":1413,"slug":1414,"type":15},"Agents","agents",{"name":9,"slug":8,"type":15},{"name":1417,"slug":1418,"type":15},"Research","research","2026-07-14T05:28:06.816956",{"slug":1421,"name":1421,"fn":1422,"description":1423,"org":1424,"tags":1425,"stars":20,"repoUrl":21,"updatedAt":1429},"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},[1426,1427,1428],{"name":1386,"slug":1387,"type":15},{"name":9,"slug":8,"type":15},{"name":1328,"slug":1329,"type":15},"2026-07-17T05:29:03.913266",{"slug":1431,"name":1431,"fn":1432,"description":1433,"org":1434,"tags":1435,"stars":20,"repoUrl":21,"updatedAt":1444},"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},[1436,1437,1440,1441],{"name":1372,"slug":1373,"type":15},{"name":1438,"slug":1439,"type":15},"Imaging","imaging",{"name":9,"slug":8,"type":15},{"name":1442,"slug":1443,"type":15},"Video","video","2026-07-17T05:28:53.905004",{"slug":1446,"name":1446,"fn":1447,"description":1448,"org":1449,"tags":1450,"stars":20,"repoUrl":21,"updatedAt":1459},"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},[1451,1452,1455,1456],{"name":1247,"slug":1248,"type":15},{"name":1453,"slug":1454,"type":15},"Docker","docker",{"name":9,"slug":8,"type":15},{"name":1457,"slug":1458,"type":15},"Operations","operations","2026-07-17T05:28:56.913999",{"slug":1461,"name":1461,"fn":1462,"description":1463,"org":1464,"tags":1465,"stars":20,"repoUrl":21,"updatedAt":1473},"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},[1466,1467,1470],{"name":9,"slug":8,"type":15},{"name":1468,"slug":1469,"type":15},"Quantum Computing","quantum-computing",{"name":1471,"slug":1472,"type":15},"Simulation","simulation","2026-07-14T05:26:58.898253",305]