[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill-nvidia-earth2studio-discover":3,"mdc-xyemxg-key":31,"related-repo-nvidia-earth2studio-discover":942,"related-org-nvidia-earth2studio-discover":1046},{"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},"earth2studio-discover","discover Earth2Studio models and data sources","Find Earth2Studio models, data sources, and examples for a weather\u002Fclimate use case. Do NOT use for writing inference code, downloading data, or installation.\n",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},"nvidia","NVIDIA","https:\u002F\u002Fpexgzepcugksgbtrxkhf.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Forg-logos\u002Fnvidia.png",[12,16,17],{"name":13,"slug":14,"type":15},"Research","research","tag",{"name":9,"slug":8,"type":15},{"name":18,"slug":19,"type":15},"Search","search",2473,"https:\u002F\u002Fgithub.com\u002FNVIDIA\u002Fskills","2026-07-14T05:27:16.785177","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\u002Fearth2studio-discover","---\nname: earth2studio-discover\nversion: 0.16.0\nlicense: Apache-2.0\nmetadata:\n  author: NVIDIA Earth-2 Team\n  tags:\n    - earth2studio\n    - earth2\n    - python\n    - discovery\n    - models\n    - data-sources\ndescription: >\n  Find Earth2Studio models, data sources, and examples for a weather\u002Fclimate use\n  case. Do NOT use for writing inference code, downloading data, or installation.\n---\n\n# Earth2Studio Discoverability Skill\n\n## Purpose\n\nHelp users identify the right Earth2Studio models, data sources, and examples for\ntheir weather\u002Fclimate task. Use when: comparing models by GPU\u002FVRAM requirements,\nchoosing forecast class (nowcast, medium-range, seasonal), finding compatible\ndata sources via lexicons, or locating gallery examples for downscaling,\nensemble generation, or data assimilation.\n\n## Prerequisites\n\n- Internet access to fetch live documentation pages from nvidia.github.io\n- Familiarity with Earth2Studio badge system (Class, Region, VRAM, Release)\n\nYou are helping a user find the right Earth2Studio components for their use case. Your job is to understand what they want to do, then point them at the models, data sources, and examples that fit — verified against live documentation.\n\n## Core principle: discover from live docs, don't memorize\n\nEarth2Studio adds models, data sources, and examples every release. Model classes get new badges, new data sources appear, examples get reorganized. Any static list in this skill will rot.\n\n**Rules:**\n1. Always fetch the relevant live doc pages before recommending components.\n2. Use badge metadata (Region, Class, VRAM, Release) from the docs to filter candidates.\n3. Verify data-source ↔ model compatibility using the lexicon system (see Step 4).\n4. Cite doc URLs so the user can explore further.\n\n## Live doc references\n\nFetch these pages as needed (not all at once — only what the user's question requires):\n\n| Category | URL |\n|----------|-----|\n| Prognostic models | https:\u002F\u002Fnvidia.github.io\u002Fearth2studio\u002Fmodules\u002Fmodels_px.html |\n| Diagnostic models | https:\u002F\u002Fnvidia.github.io\u002Fearth2studio\u002Fmodules\u002Fmodels_dx.html |\n| Data assimilation | https:\u002F\u002Fnvidia.github.io\u002Fearth2studio\u002Fmodules\u002Fmodels_da.html |\n| Data sources (analysis) | https:\u002F\u002Fnvidia.github.io\u002Fearth2studio\u002Fmodules\u002Fdatasources_analysis.html |\n| Data sources (forecast) | https:\u002F\u002Fnvidia.github.io\u002Fearth2studio\u002Fmodules\u002Fdatasources_forecast.html |\n| Data sources (dataframe) | https:\u002F\u002Fnvidia.github.io\u002Fearth2studio\u002Fmodules\u002Fdatasources_dataframe.html |\n| Examples gallery | https:\u002F\u002Fnvidia.github.io\u002Fearth2studio\u002Fexamples\u002Findex.html |\n| Lexicon source | https:\u002F\u002Fgithub.com\u002FNVIDIA\u002Fearth2studio\u002Ftree\u002Fmain\u002Fearth2studio\u002Flexicon |\n\n## Interaction protocol\n\n### Step 1. Understand the user's problem\n\nExtract from what the user has said (ask follow-ups if needed, cap at 3 questions):\n\n- **Task type** — medium-range forecasting, nowcasting, downscaling\u002Fsuper-resolution, seasonal\u002Fsubseasonal, data assimilation, climate projection, ensemble generation, derived diagnostics\n- **Region** — global, North America, Europe, Asia, specific country\u002Farea\n- **Temporal scale** — hours ahead (nowcast), days ahead (medium-range), weeks\u002Fmonths (seasonal), climate\n- **Variables of interest** — temperature, precipitation, wind, pressure, radiation, specific levels, etc.\n- **Hardware constraints** — GPU type, available VRAM (40GB, 48GB, 80GB, 96GB)\n- **Deterministic vs. ensemble** — single forecast or probabilistic\n\nGood follow-up phrasing: *\"Are you looking for a single best-estimate forecast or an ensemble with uncertainty?\"* — not *\"what's your use case?\"*\n\n### Step 2. Fetch relevant model docs\n\nBased on the user's task type, fetch the appropriate model page(s):\n\n- Forecasting → prognostic models (px)\n- Post-processing \u002F downscaling \u002F derived variables → diagnostic models (dx)\n- Observation integration → data assimilation (da)\n- Often a workflow chains px → dx, so check both\n\nFrom the doc pages, extract for each candidate model:\n- **Class badge** — NWC, DS, MR, S2S, DA, CM\n- **Region badge** — Global, NA, EU, AS, etc.\n- **Rec VRAM badge** — minimum GPU memory\n- **Release year** — newer models generally supersede older ones in the same class\n\nFilter to models matching the user's task type, region, and hardware. Present a short-list (not the full catalog) with badge metadata.\n\n### Step 3. Fetch relevant data source docs\n\nBased on the user's data needs, fetch the appropriate data source page:\n\n- Historical reanalysis → analysis data sources\n- Real-time or operational → forecast data sources\n- Observations \u002F station data → dataframe data sources\n\nNote which data sources cover the user's region and variables.\n\n### Step 4. Verify compatibility via lexicon\n\nThis is the key technical step. Earth2Studio models declare their required input variables via `input_coords()`. Data sources expose available variables through their lexicon VOCAB. If a data source's lexicon VOCAB keys contain all variables in a model's `input_coords` (the \"variable\" dimension), they are compatible.\n\nTo verify:\n1. Check the model's doc page or source for its `input_coords` — specifically the variable list\n2. Check the data source's lexicon file at `earth2studio\u002Flexicon\u002F\u003Csource>.py` for its VOCAB keys\n3. Confirm the data source VOCAB covers all variables the model needs\n\nIf checking source code directly (e.g. user has a local clone), the lexicon files are at:\n```\nearth2studio\u002Flexicon\u002Fgfs.py\nearth2studio\u002Flexicon\u002Fhrrr.py\nearth2studio\u002Flexicon\u002Fcds.py\nearth2studio\u002Flexicon\u002Farco.py\nearth2studio\u002Flexicon\u002Fwb2.py\n... (one per data source)\n```\n\nEach defines a `VOCAB: dict[str, str | tuple]` mapping Earth2Studio variable names to source-specific identifiers.\n\nSurface compatibility results clearly: *\"GraphCastOperational needs [list of variables] — GFS and ERA5 (via ARCO\u002FCDS) both provide these, but HRRR does not cover pressure levels above X.\"*\n\n### Step 5. Suggest examples\n\nFetch the examples gallery and identify examples that demonstrate the user's workflow pattern. Examples are organized by category:\n\n- `01_getting_started` — basic deterministic, diagnostic, ensemble pipelines\n- `02_medium_range` — ensemble extension, perturbation, cyclone tracking\n- `03_downscaling` — CorrDiff, CBottle, ensemble downscaling\n- `04_nowcasting` — StormCast, StormScope\n- `05_data_assimilation` — StormCast SDA, HealDA\n- `06_seasonal` — DLESyM, statistical methods\n- `07_misc` — distributed inference, IO, custom data, generation\n- `08_extend` — building custom models, diagnostics, data sources\n\nPoint the user at the most relevant 1–3 examples as starting points. Explain what each demonstrates and how it relates to their problem.\n\n### Step 6. Return recommendations\n\nOutput structure (omit empty sections):\n\n```\n## Your use case\n[1-2 sentence restatement of what the user wants to do]\n\n## Recommended models\n| Model | Class | Region | VRAM | Why |\n|-------|-------|--------|------|-----|\n[Short-list with rationale per row]\n\n## Compatible data sources\n| Data Source | Coverage | Compatible with |\n|-------------|----------|-----------------|\n[Verified via lexicon]\n\n## Relevant examples\n- [Example name](link) — what it demonstrates\n\n## Next steps\n[What to install, what to read next]\n```\n\nKeep recommendations to 2–4 models maximum. If multiple options exist, explain the tradeoff (accuracy vs. speed, deterministic vs. ensemble, VRAM, etc.) rather than listing everything.\n\n## Limitations\n\n- Recommendations are only as current as the live docs; unreleased models are not discoverable.\n- Badge metadata may be incomplete for newly added models.\n- Lexicon compatibility checks require source code access for full accuracy; doc-only checks are approximate.\n\n## Troubleshooting\n\n| Error | Cause | Solution |\n|-------|-------|----------|\n| Model page returns 404 | URL changed after a release | Check https:\u002F\u002Fnvidia.github.io\u002Fearth2studio\u002F for updated navigation |\n| Lexicon file not found | Data source is new or renamed | Search `earth2studio\u002Flexicon\u002F` directory for current filenames |\n| Badge missing from model | Model docs not yet updated | Fall back to the model's source code `__init__` or README for specs |\n\n## Ownership and out-of-scope\n\n**Owns:** component discovery, model\u002Fdata-source compatibility checking, badge-based filtering, example recommendation, hardware-fit assessment.\n\n**Does not own:** installation (use earth2studio-install skill), writing inference code, model training, custom model development, runtime debugging, PhysicsNeMo model discovery.\n",{"data":32,"body":43},{"name":4,"version":33,"license":23,"metadata":34,"description":6},"0.16.0",{"author":35,"tags":36},"NVIDIA Earth-2 Team",[37,38,39,40,41,42],"earth2studio","earth2","python","discovery","models","data-sources",{"type":44,"children":45},"root",[46,55,62,68,74,89,94,100,105,114,138,144,149,315,321,328,333,396,414,420,425,448,453,496,501,507,512,530,535,541,563,568,601,606,618,631,649,655,660,751,756,762,767,776,781,787,805,811,916,922,932],{"type":47,"tag":48,"props":49,"children":51},"element","h1",{"id":50},"earth2studio-discoverability-skill",[52],{"type":53,"value":54},"text","Earth2Studio Discoverability Skill",{"type":47,"tag":56,"props":57,"children":59},"h2",{"id":58},"purpose",[60],{"type":53,"value":61},"Purpose",{"type":47,"tag":63,"props":64,"children":65},"p",{},[66],{"type":53,"value":67},"Help users identify the right Earth2Studio models, data sources, and examples for\ntheir weather\u002Fclimate task. Use when: comparing models by GPU\u002FVRAM requirements,\nchoosing forecast class (nowcast, medium-range, seasonal), finding compatible\ndata sources via lexicons, or locating gallery examples for downscaling,\nensemble generation, or data assimilation.",{"type":47,"tag":56,"props":69,"children":71},{"id":70},"prerequisites",[72],{"type":53,"value":73},"Prerequisites",{"type":47,"tag":75,"props":76,"children":77},"ul",{},[78,84],{"type":47,"tag":79,"props":80,"children":81},"li",{},[82],{"type":53,"value":83},"Internet access to fetch live documentation pages from nvidia.github.io",{"type":47,"tag":79,"props":85,"children":86},{},[87],{"type":53,"value":88},"Familiarity with Earth2Studio badge system (Class, Region, VRAM, Release)",{"type":47,"tag":63,"props":90,"children":91},{},[92],{"type":53,"value":93},"You are helping a user find the right Earth2Studio components for their use case. Your job is to understand what they want to do, then point them at the models, data sources, and examples that fit — verified against live documentation.",{"type":47,"tag":56,"props":95,"children":97},{"id":96},"core-principle-discover-from-live-docs-dont-memorize",[98],{"type":53,"value":99},"Core principle: discover from live docs, don't memorize",{"type":47,"tag":63,"props":101,"children":102},{},[103],{"type":53,"value":104},"Earth2Studio adds models, data sources, and examples every release. Model classes get new badges, new data sources appear, examples get reorganized. Any static list in this skill will rot.",{"type":47,"tag":63,"props":106,"children":107},{},[108],{"type":47,"tag":109,"props":110,"children":111},"strong",{},[112],{"type":53,"value":113},"Rules:",{"type":47,"tag":115,"props":116,"children":117},"ol",{},[118,123,128,133],{"type":47,"tag":79,"props":119,"children":120},{},[121],{"type":53,"value":122},"Always fetch the relevant live doc pages before recommending components.",{"type":47,"tag":79,"props":124,"children":125},{},[126],{"type":53,"value":127},"Use badge metadata (Region, Class, VRAM, Release) from the docs to filter candidates.",{"type":47,"tag":79,"props":129,"children":130},{},[131],{"type":53,"value":132},"Verify data-source ↔ model compatibility using the lexicon system (see Step 4).",{"type":47,"tag":79,"props":134,"children":135},{},[136],{"type":53,"value":137},"Cite doc URLs so the user can explore further.",{"type":47,"tag":56,"props":139,"children":141},{"id":140},"live-doc-references",[142],{"type":53,"value":143},"Live doc references",{"type":47,"tag":63,"props":145,"children":146},{},[147],{"type":53,"value":148},"Fetch these pages as needed (not all at once — only what the user's question requires):",{"type":47,"tag":150,"props":151,"children":152},"table",{},[153,172],{"type":47,"tag":154,"props":155,"children":156},"thead",{},[157],{"type":47,"tag":158,"props":159,"children":160},"tr",{},[161,167],{"type":47,"tag":162,"props":163,"children":164},"th",{},[165],{"type":53,"value":166},"Category",{"type":47,"tag":162,"props":168,"children":169},{},[170],{"type":53,"value":171},"URL",{"type":47,"tag":173,"props":174,"children":175},"tbody",{},[176,196,213,230,247,264,281,298],{"type":47,"tag":158,"props":177,"children":178},{},[179,185],{"type":47,"tag":180,"props":181,"children":182},"td",{},[183],{"type":53,"value":184},"Prognostic models",{"type":47,"tag":180,"props":186,"children":187},{},[188],{"type":47,"tag":189,"props":190,"children":194},"a",{"href":191,"rel":192},"https:\u002F\u002Fnvidia.github.io\u002Fearth2studio\u002Fmodules\u002Fmodels_px.html",[193],"nofollow",[195],{"type":53,"value":191},{"type":47,"tag":158,"props":197,"children":198},{},[199,204],{"type":47,"tag":180,"props":200,"children":201},{},[202],{"type":53,"value":203},"Diagnostic models",{"type":47,"tag":180,"props":205,"children":206},{},[207],{"type":47,"tag":189,"props":208,"children":211},{"href":209,"rel":210},"https:\u002F\u002Fnvidia.github.io\u002Fearth2studio\u002Fmodules\u002Fmodels_dx.html",[193],[212],{"type":53,"value":209},{"type":47,"tag":158,"props":214,"children":215},{},[216,221],{"type":47,"tag":180,"props":217,"children":218},{},[219],{"type":53,"value":220},"Data assimilation",{"type":47,"tag":180,"props":222,"children":223},{},[224],{"type":47,"tag":189,"props":225,"children":228},{"href":226,"rel":227},"https:\u002F\u002Fnvidia.github.io\u002Fearth2studio\u002Fmodules\u002Fmodels_da.html",[193],[229],{"type":53,"value":226},{"type":47,"tag":158,"props":231,"children":232},{},[233,238],{"type":47,"tag":180,"props":234,"children":235},{},[236],{"type":53,"value":237},"Data sources (analysis)",{"type":47,"tag":180,"props":239,"children":240},{},[241],{"type":47,"tag":189,"props":242,"children":245},{"href":243,"rel":244},"https:\u002F\u002Fnvidia.github.io\u002Fearth2studio\u002Fmodules\u002Fdatasources_analysis.html",[193],[246],{"type":53,"value":243},{"type":47,"tag":158,"props":248,"children":249},{},[250,255],{"type":47,"tag":180,"props":251,"children":252},{},[253],{"type":53,"value":254},"Data sources (forecast)",{"type":47,"tag":180,"props":256,"children":257},{},[258],{"type":47,"tag":189,"props":259,"children":262},{"href":260,"rel":261},"https:\u002F\u002Fnvidia.github.io\u002Fearth2studio\u002Fmodules\u002Fdatasources_forecast.html",[193],[263],{"type":53,"value":260},{"type":47,"tag":158,"props":265,"children":266},{},[267,272],{"type":47,"tag":180,"props":268,"children":269},{},[270],{"type":53,"value":271},"Data sources (dataframe)",{"type":47,"tag":180,"props":273,"children":274},{},[275],{"type":47,"tag":189,"props":276,"children":279},{"href":277,"rel":278},"https:\u002F\u002Fnvidia.github.io\u002Fearth2studio\u002Fmodules\u002Fdatasources_dataframe.html",[193],[280],{"type":53,"value":277},{"type":47,"tag":158,"props":282,"children":283},{},[284,289],{"type":47,"tag":180,"props":285,"children":286},{},[287],{"type":53,"value":288},"Examples gallery",{"type":47,"tag":180,"props":290,"children":291},{},[292],{"type":47,"tag":189,"props":293,"children":296},{"href":294,"rel":295},"https:\u002F\u002Fnvidia.github.io\u002Fearth2studio\u002Fexamples\u002Findex.html",[193],[297],{"type":53,"value":294},{"type":47,"tag":158,"props":299,"children":300},{},[301,306],{"type":47,"tag":180,"props":302,"children":303},{},[304],{"type":53,"value":305},"Lexicon source",{"type":47,"tag":180,"props":307,"children":308},{},[309],{"type":47,"tag":189,"props":310,"children":313},{"href":311,"rel":312},"https:\u002F\u002Fgithub.com\u002FNVIDIA\u002Fearth2studio\u002Ftree\u002Fmain\u002Fearth2studio\u002Flexicon",[193],[314],{"type":53,"value":311},{"type":47,"tag":56,"props":316,"children":318},{"id":317},"interaction-protocol",[319],{"type":53,"value":320},"Interaction protocol",{"type":47,"tag":322,"props":323,"children":325},"h3",{"id":324},"step-1-understand-the-users-problem",[326],{"type":53,"value":327},"Step 1. Understand the user's problem",{"type":47,"tag":63,"props":329,"children":330},{},[331],{"type":53,"value":332},"Extract from what the user has said (ask follow-ups if needed, cap at 3 questions):",{"type":47,"tag":75,"props":334,"children":335},{},[336,346,356,366,376,386],{"type":47,"tag":79,"props":337,"children":338},{},[339,344],{"type":47,"tag":109,"props":340,"children":341},{},[342],{"type":53,"value":343},"Task type",{"type":53,"value":345}," — medium-range forecasting, nowcasting, downscaling\u002Fsuper-resolution, seasonal\u002Fsubseasonal, data assimilation, climate projection, ensemble generation, derived diagnostics",{"type":47,"tag":79,"props":347,"children":348},{},[349,354],{"type":47,"tag":109,"props":350,"children":351},{},[352],{"type":53,"value":353},"Region",{"type":53,"value":355}," — global, North America, Europe, Asia, specific country\u002Farea",{"type":47,"tag":79,"props":357,"children":358},{},[359,364],{"type":47,"tag":109,"props":360,"children":361},{},[362],{"type":53,"value":363},"Temporal scale",{"type":53,"value":365}," — hours ahead (nowcast), days ahead (medium-range), weeks\u002Fmonths (seasonal), climate",{"type":47,"tag":79,"props":367,"children":368},{},[369,374],{"type":47,"tag":109,"props":370,"children":371},{},[372],{"type":53,"value":373},"Variables of interest",{"type":53,"value":375}," — temperature, precipitation, wind, pressure, radiation, specific levels, etc.",{"type":47,"tag":79,"props":377,"children":378},{},[379,384],{"type":47,"tag":109,"props":380,"children":381},{},[382],{"type":53,"value":383},"Hardware constraints",{"type":53,"value":385}," — GPU type, available VRAM (40GB, 48GB, 80GB, 96GB)",{"type":47,"tag":79,"props":387,"children":388},{},[389,394],{"type":47,"tag":109,"props":390,"children":391},{},[392],{"type":53,"value":393},"Deterministic vs. ensemble",{"type":53,"value":395}," — single forecast or probabilistic",{"type":47,"tag":63,"props":397,"children":398},{},[399,401,407,409],{"type":53,"value":400},"Good follow-up phrasing: ",{"type":47,"tag":402,"props":403,"children":404},"em",{},[405],{"type":53,"value":406},"\"Are you looking for a single best-estimate forecast or an ensemble with uncertainty?\"",{"type":53,"value":408}," — not ",{"type":47,"tag":402,"props":410,"children":411},{},[412],{"type":53,"value":413},"\"what's your use case?\"",{"type":47,"tag":322,"props":415,"children":417},{"id":416},"step-2-fetch-relevant-model-docs",[418],{"type":53,"value":419},"Step 2. Fetch relevant model docs",{"type":47,"tag":63,"props":421,"children":422},{},[423],{"type":53,"value":424},"Based on the user's task type, fetch the appropriate model page(s):",{"type":47,"tag":75,"props":426,"children":427},{},[428,433,438,443],{"type":47,"tag":79,"props":429,"children":430},{},[431],{"type":53,"value":432},"Forecasting → prognostic models (px)",{"type":47,"tag":79,"props":434,"children":435},{},[436],{"type":53,"value":437},"Post-processing \u002F downscaling \u002F derived variables → diagnostic models (dx)",{"type":47,"tag":79,"props":439,"children":440},{},[441],{"type":53,"value":442},"Observation integration → data assimilation (da)",{"type":47,"tag":79,"props":444,"children":445},{},[446],{"type":53,"value":447},"Often a workflow chains px → dx, so check both",{"type":47,"tag":63,"props":449,"children":450},{},[451],{"type":53,"value":452},"From the doc pages, extract for each candidate model:",{"type":47,"tag":75,"props":454,"children":455},{},[456,466,476,486],{"type":47,"tag":79,"props":457,"children":458},{},[459,464],{"type":47,"tag":109,"props":460,"children":461},{},[462],{"type":53,"value":463},"Class badge",{"type":53,"value":465}," — NWC, DS, MR, S2S, DA, CM",{"type":47,"tag":79,"props":467,"children":468},{},[469,474],{"type":47,"tag":109,"props":470,"children":471},{},[472],{"type":53,"value":473},"Region badge",{"type":53,"value":475}," — Global, NA, EU, AS, etc.",{"type":47,"tag":79,"props":477,"children":478},{},[479,484],{"type":47,"tag":109,"props":480,"children":481},{},[482],{"type":53,"value":483},"Rec VRAM badge",{"type":53,"value":485}," — minimum GPU memory",{"type":47,"tag":79,"props":487,"children":488},{},[489,494],{"type":47,"tag":109,"props":490,"children":491},{},[492],{"type":53,"value":493},"Release year",{"type":53,"value":495}," — newer models generally supersede older ones in the same class",{"type":47,"tag":63,"props":497,"children":498},{},[499],{"type":53,"value":500},"Filter to models matching the user's task type, region, and hardware. Present a short-list (not the full catalog) with badge metadata.",{"type":47,"tag":322,"props":502,"children":504},{"id":503},"step-3-fetch-relevant-data-source-docs",[505],{"type":53,"value":506},"Step 3. Fetch relevant data source docs",{"type":47,"tag":63,"props":508,"children":509},{},[510],{"type":53,"value":511},"Based on the user's data needs, fetch the appropriate data source page:",{"type":47,"tag":75,"props":513,"children":514},{},[515,520,525],{"type":47,"tag":79,"props":516,"children":517},{},[518],{"type":53,"value":519},"Historical reanalysis → analysis data sources",{"type":47,"tag":79,"props":521,"children":522},{},[523],{"type":53,"value":524},"Real-time or operational → forecast data sources",{"type":47,"tag":79,"props":526,"children":527},{},[528],{"type":53,"value":529},"Observations \u002F station data → dataframe data sources",{"type":47,"tag":63,"props":531,"children":532},{},[533],{"type":53,"value":534},"Note which data sources cover the user's region and variables.",{"type":47,"tag":322,"props":536,"children":538},{"id":537},"step-4-verify-compatibility-via-lexicon",[539],{"type":53,"value":540},"Step 4. Verify compatibility via lexicon",{"type":47,"tag":63,"props":542,"children":543},{},[544,546,553,555,561],{"type":53,"value":545},"This is the key technical step. Earth2Studio models declare their required input variables via ",{"type":47,"tag":547,"props":548,"children":550},"code",{"className":549},[],[551],{"type":53,"value":552},"input_coords()",{"type":53,"value":554},". Data sources expose available variables through their lexicon VOCAB. If a data source's lexicon VOCAB keys contain all variables in a model's ",{"type":47,"tag":547,"props":556,"children":558},{"className":557},[],[559],{"type":53,"value":560},"input_coords",{"type":53,"value":562}," (the \"variable\" dimension), they are compatible.",{"type":47,"tag":63,"props":564,"children":565},{},[566],{"type":53,"value":567},"To verify:",{"type":47,"tag":115,"props":569,"children":570},{},[571,583,596],{"type":47,"tag":79,"props":572,"children":573},{},[574,576,581],{"type":53,"value":575},"Check the model's doc page or source for its ",{"type":47,"tag":547,"props":577,"children":579},{"className":578},[],[580],{"type":53,"value":560},{"type":53,"value":582}," — specifically the variable list",{"type":47,"tag":79,"props":584,"children":585},{},[586,588,594],{"type":53,"value":587},"Check the data source's lexicon file at ",{"type":47,"tag":547,"props":589,"children":591},{"className":590},[],[592],{"type":53,"value":593},"earth2studio\u002Flexicon\u002F\u003Csource>.py",{"type":53,"value":595}," for its VOCAB keys",{"type":47,"tag":79,"props":597,"children":598},{},[599],{"type":53,"value":600},"Confirm the data source VOCAB covers all variables the model needs",{"type":47,"tag":63,"props":602,"children":603},{},[604],{"type":53,"value":605},"If checking source code directly (e.g. user has a local clone), the lexicon files are at:",{"type":47,"tag":607,"props":608,"children":612},"pre",{"className":609,"code":611,"language":53},[610],"language-text","earth2studio\u002Flexicon\u002Fgfs.py\nearth2studio\u002Flexicon\u002Fhrrr.py\nearth2studio\u002Flexicon\u002Fcds.py\nearth2studio\u002Flexicon\u002Farco.py\nearth2studio\u002Flexicon\u002Fwb2.py\n... (one per data source)\n",[613],{"type":47,"tag":547,"props":614,"children":616},{"__ignoreMap":615},"",[617],{"type":53,"value":611},{"type":47,"tag":63,"props":619,"children":620},{},[621,623,629],{"type":53,"value":622},"Each defines a ",{"type":47,"tag":547,"props":624,"children":626},{"className":625},[],[627],{"type":53,"value":628},"VOCAB: dict[str, str | tuple]",{"type":53,"value":630}," mapping Earth2Studio variable names to source-specific identifiers.",{"type":47,"tag":63,"props":632,"children":633},{},[634,636],{"type":53,"value":635},"Surface compatibility results clearly: ",{"type":47,"tag":402,"props":637,"children":638},{},[639,641,647],{"type":53,"value":640},"\"GraphCastOperational needs ",{"type":47,"tag":642,"props":643,"children":644},"span",{},[645],{"type":53,"value":646},"list of variables",{"type":53,"value":648}," — GFS and ERA5 (via ARCO\u002FCDS) both provide these, but HRRR does not cover pressure levels above X.\"",{"type":47,"tag":322,"props":650,"children":652},{"id":651},"step-5-suggest-examples",[653],{"type":53,"value":654},"Step 5. Suggest examples",{"type":47,"tag":63,"props":656,"children":657},{},[658],{"type":53,"value":659},"Fetch the examples gallery and identify examples that demonstrate the user's workflow pattern. Examples are organized by category:",{"type":47,"tag":75,"props":661,"children":662},{},[663,674,685,696,707,718,729,740],{"type":47,"tag":79,"props":664,"children":665},{},[666,672],{"type":47,"tag":547,"props":667,"children":669},{"className":668},[],[670],{"type":53,"value":671},"01_getting_started",{"type":53,"value":673}," — basic deterministic, diagnostic, ensemble pipelines",{"type":47,"tag":79,"props":675,"children":676},{},[677,683],{"type":47,"tag":547,"props":678,"children":680},{"className":679},[],[681],{"type":53,"value":682},"02_medium_range",{"type":53,"value":684}," — ensemble extension, perturbation, cyclone tracking",{"type":47,"tag":79,"props":686,"children":687},{},[688,694],{"type":47,"tag":547,"props":689,"children":691},{"className":690},[],[692],{"type":53,"value":693},"03_downscaling",{"type":53,"value":695}," — CorrDiff, CBottle, ensemble downscaling",{"type":47,"tag":79,"props":697,"children":698},{},[699,705],{"type":47,"tag":547,"props":700,"children":702},{"className":701},[],[703],{"type":53,"value":704},"04_nowcasting",{"type":53,"value":706}," — StormCast, StormScope",{"type":47,"tag":79,"props":708,"children":709},{},[710,716],{"type":47,"tag":547,"props":711,"children":713},{"className":712},[],[714],{"type":53,"value":715},"05_data_assimilation",{"type":53,"value":717}," — StormCast SDA, HealDA",{"type":47,"tag":79,"props":719,"children":720},{},[721,727],{"type":47,"tag":547,"props":722,"children":724},{"className":723},[],[725],{"type":53,"value":726},"06_seasonal",{"type":53,"value":728}," — DLESyM, statistical methods",{"type":47,"tag":79,"props":730,"children":731},{},[732,738],{"type":47,"tag":547,"props":733,"children":735},{"className":734},[],[736],{"type":53,"value":737},"07_misc",{"type":53,"value":739}," — distributed inference, IO, custom data, generation",{"type":47,"tag":79,"props":741,"children":742},{},[743,749],{"type":47,"tag":547,"props":744,"children":746},{"className":745},[],[747],{"type":53,"value":748},"08_extend",{"type":53,"value":750}," — building custom models, diagnostics, data sources",{"type":47,"tag":63,"props":752,"children":753},{},[754],{"type":53,"value":755},"Point the user at the most relevant 1–3 examples as starting points. Explain what each demonstrates and how it relates to their problem.",{"type":47,"tag":322,"props":757,"children":759},{"id":758},"step-6-return-recommendations",[760],{"type":53,"value":761},"Step 6. Return recommendations",{"type":47,"tag":63,"props":763,"children":764},{},[765],{"type":53,"value":766},"Output structure (omit empty sections):",{"type":47,"tag":607,"props":768,"children":771},{"className":769,"code":770,"language":53},[610],"## Your use case\n[1-2 sentence restatement of what the user wants to do]\n\n## Recommended models\n| Model | Class | Region | VRAM | Why |\n|-------|-------|--------|------|-----|\n[Short-list with rationale per row]\n\n## Compatible data sources\n| Data Source | Coverage | Compatible with |\n|-------------|----------|-----------------|\n[Verified via lexicon]\n\n## Relevant examples\n- [Example name](link) — what it demonstrates\n\n## Next steps\n[What to install, what to read next]\n",[772],{"type":47,"tag":547,"props":773,"children":774},{"__ignoreMap":615},[775],{"type":53,"value":770},{"type":47,"tag":63,"props":777,"children":778},{},[779],{"type":53,"value":780},"Keep recommendations to 2–4 models maximum. If multiple options exist, explain the tradeoff (accuracy vs. speed, deterministic vs. ensemble, VRAM, etc.) rather than listing everything.",{"type":47,"tag":56,"props":782,"children":784},{"id":783},"limitations",[785],{"type":53,"value":786},"Limitations",{"type":47,"tag":75,"props":788,"children":789},{},[790,795,800],{"type":47,"tag":79,"props":791,"children":792},{},[793],{"type":53,"value":794},"Recommendations are only as current as the live docs; unreleased models are not discoverable.",{"type":47,"tag":79,"props":796,"children":797},{},[798],{"type":53,"value":799},"Badge metadata may be incomplete for newly added models.",{"type":47,"tag":79,"props":801,"children":802},{},[803],{"type":53,"value":804},"Lexicon compatibility checks require source code access for full accuracy; doc-only checks are approximate.",{"type":47,"tag":56,"props":806,"children":808},{"id":807},"troubleshooting",[809],{"type":53,"value":810},"Troubleshooting",{"type":47,"tag":150,"props":812,"children":813},{},[814,835],{"type":47,"tag":154,"props":815,"children":816},{},[817],{"type":47,"tag":158,"props":818,"children":819},{},[820,825,830],{"type":47,"tag":162,"props":821,"children":822},{},[823],{"type":53,"value":824},"Error",{"type":47,"tag":162,"props":826,"children":827},{},[828],{"type":53,"value":829},"Cause",{"type":47,"tag":162,"props":831,"children":832},{},[833],{"type":53,"value":834},"Solution",{"type":47,"tag":173,"props":836,"children":837},{},[838,864,890],{"type":47,"tag":158,"props":839,"children":840},{},[841,846,851],{"type":47,"tag":180,"props":842,"children":843},{},[844],{"type":53,"value":845},"Model page returns 404",{"type":47,"tag":180,"props":847,"children":848},{},[849],{"type":53,"value":850},"URL changed after a release",{"type":47,"tag":180,"props":852,"children":853},{},[854,856,862],{"type":53,"value":855},"Check ",{"type":47,"tag":189,"props":857,"children":860},{"href":858,"rel":859},"https:\u002F\u002Fnvidia.github.io\u002Fearth2studio\u002F",[193],[861],{"type":53,"value":858},{"type":53,"value":863}," for updated navigation",{"type":47,"tag":158,"props":865,"children":866},{},[867,872,877],{"type":47,"tag":180,"props":868,"children":869},{},[870],{"type":53,"value":871},"Lexicon file not found",{"type":47,"tag":180,"props":873,"children":874},{},[875],{"type":53,"value":876},"Data source is new or renamed",{"type":47,"tag":180,"props":878,"children":879},{},[880,882,888],{"type":53,"value":881},"Search ",{"type":47,"tag":547,"props":883,"children":885},{"className":884},[],[886],{"type":53,"value":887},"earth2studio\u002Flexicon\u002F",{"type":53,"value":889}," directory for current filenames",{"type":47,"tag":158,"props":891,"children":892},{},[893,898,903],{"type":47,"tag":180,"props":894,"children":895},{},[896],{"type":53,"value":897},"Badge missing from model",{"type":47,"tag":180,"props":899,"children":900},{},[901],{"type":53,"value":902},"Model docs not yet updated",{"type":47,"tag":180,"props":904,"children":905},{},[906,908,914],{"type":53,"value":907},"Fall back to the model's source code ",{"type":47,"tag":547,"props":909,"children":911},{"className":910},[],[912],{"type":53,"value":913},"__init__",{"type":53,"value":915}," or README for specs",{"type":47,"tag":56,"props":917,"children":919},{"id":918},"ownership-and-out-of-scope",[920],{"type":53,"value":921},"Ownership and out-of-scope",{"type":47,"tag":63,"props":923,"children":924},{},[925,930],{"type":47,"tag":109,"props":926,"children":927},{},[928],{"type":53,"value":929},"Owns:",{"type":53,"value":931}," component discovery, model\u002Fdata-source compatibility checking, badge-based filtering, example recommendation, hardware-fit assessment.",{"type":47,"tag":63,"props":933,"children":934},{},[935,940],{"type":47,"tag":109,"props":936,"children":937},{},[938],{"type":53,"value":939},"Does not own:",{"type":53,"value":941}," installation (use earth2studio-install skill), writing inference code, model training, custom model development, runtime debugging, PhysicsNeMo model discovery.",{"items":943,"total":1045},[944,961,975,987,999,1016,1031],{"slug":945,"name":945,"fn":946,"description":947,"org":948,"tags":949,"stars":20,"repoUrl":21,"updatedAt":960},"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},[950,953,956,957],{"name":951,"slug":952,"type":15},"Data Analysis","data-analysis",{"name":954,"slug":955,"type":15},"Data Engineering","data-engineering",{"name":9,"slug":8,"type":15},{"name":958,"slug":959,"type":15},"Performance","performance","2026-07-14T05:28:43.176466",{"slug":962,"name":962,"fn":963,"description":964,"org":965,"tags":966,"stars":20,"repoUrl":21,"updatedAt":974},"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},[967,970,973],{"name":968,"slug":969,"type":15},"Deployment","deployment",{"name":971,"slug":972,"type":15},"Infrastructure","infrastructure",{"name":9,"slug":8,"type":15},"2026-07-14T05:29:06.667109",{"slug":976,"name":976,"fn":977,"description":978,"org":979,"tags":980,"stars":20,"repoUrl":21,"updatedAt":986},"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},[981,984,985],{"name":982,"slug":983,"type":15},"Agents","agents",{"name":9,"slug":8,"type":15},{"name":13,"slug":14,"type":15},"2026-07-14T05:28:06.816956",{"slug":988,"name":988,"fn":989,"description":990,"org":991,"tags":992,"stars":20,"repoUrl":21,"updatedAt":998},"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},[993,994,995],{"name":951,"slug":952,"type":15},{"name":9,"slug":8,"type":15},{"name":996,"slug":997,"type":15},"Testing","testing","2026-07-17T05:29:03.913266",{"slug":1000,"name":1000,"fn":1001,"description":1002,"org":1003,"tags":1004,"stars":20,"repoUrl":21,"updatedAt":1015},"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},[1005,1008,1011,1012],{"name":1006,"slug":1007,"type":15},"Automation","automation",{"name":1009,"slug":1010,"type":15},"Imaging","imaging",{"name":9,"slug":8,"type":15},{"name":1013,"slug":1014,"type":15},"Video","video","2026-07-17T05:28:53.905004",{"slug":1017,"name":1017,"fn":1018,"description":1019,"org":1020,"tags":1021,"stars":20,"repoUrl":21,"updatedAt":1030},"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},[1022,1023,1026,1027],{"name":968,"slug":969,"type":15},{"name":1024,"slug":1025,"type":15},"Docker","docker",{"name":9,"slug":8,"type":15},{"name":1028,"slug":1029,"type":15},"Operations","operations","2026-07-17T05:28:56.913999",{"slug":1032,"name":1032,"fn":1033,"description":1034,"org":1035,"tags":1036,"stars":20,"repoUrl":21,"updatedAt":1044},"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},[1037,1038,1041],{"name":9,"slug":8,"type":15},{"name":1039,"slug":1040,"type":15},"Quantum Computing","quantum-computing",{"name":1042,"slug":1043,"type":15},"Simulation","simulation","2026-07-14T05:26:58.898253",305,{"items":1047,"total":1195},[1048,1064,1079,1090,1102,1116,1129,1143,1154,1163,1177,1186],{"slug":1049,"name":1049,"fn":1050,"description":1051,"org":1052,"tags":1053,"stars":1061,"repoUrl":1062,"updatedAt":1063},"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},[1054,1057,1060],{"name":1055,"slug":1056,"type":15},"Documentation","documentation",{"name":1058,"slug":1059,"type":15},"MCP","mcp",{"name":18,"slug":19,"type":15},21777,"https:\u002F\u002Fgithub.com\u002FNVIDIA\u002FNemoClaw","2026-07-20T06:00:01.461044",{"slug":1065,"name":1065,"fn":1066,"description":1067,"org":1068,"tags":1069,"stars":1076,"repoUrl":1077,"updatedAt":1078},"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},[1070,1073,1074],{"name":1071,"slug":1072,"type":15},"Containers","containers",{"name":968,"slug":969,"type":15},{"name":1075,"slug":39,"type":15},"Python",17049,"https:\u002F\u002Fgithub.com\u002FNVIDIA\u002FMegatron-LM","2026-07-27T06:06:11.249662",{"slug":1080,"name":1080,"fn":1081,"description":1082,"org":1083,"tags":1084,"stars":1076,"repoUrl":1077,"updatedAt":1089},"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},[1085,1088],{"name":1086,"slug":1087,"type":15},"CI\u002FCD","ci-cd",{"name":968,"slug":969,"type":15},"2026-07-14T05:25:59.97109",{"slug":1091,"name":1091,"fn":1092,"description":1093,"org":1094,"tags":1095,"stars":1076,"repoUrl":1077,"updatedAt":1101},"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},[1096,1097,1098],{"name":1086,"slug":1087,"type":15},{"name":968,"slug":969,"type":15},{"name":1099,"slug":1100,"type":15},"GitHub","github","2026-07-27T06:06:12.278222",{"slug":1103,"name":1103,"fn":1104,"description":1105,"org":1106,"tags":1107,"stars":1076,"repoUrl":1077,"updatedAt":1115},"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},[1108,1111,1112],{"name":1109,"slug":1110,"type":15},"Debugging","debugging",{"name":1099,"slug":1100,"type":15},{"name":1113,"slug":1114,"type":15},"Triage","triage","2026-07-14T05:25:57.442089",{"slug":1117,"name":1117,"fn":1118,"description":1119,"org":1120,"tags":1121,"stars":1076,"repoUrl":1077,"updatedAt":1128},"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},[1122,1125],{"name":1123,"slug":1124,"type":15},"Best Practices","best-practices",{"name":1126,"slug":1127,"type":15},"Code Analysis","code-analysis","2026-07-14T05:25:56.18433",{"slug":1130,"name":1130,"fn":1131,"description":1132,"org":1133,"tags":1134,"stars":1076,"repoUrl":1077,"updatedAt":1142},"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},[1135,1138,1141],{"name":1136,"slug":1137,"type":15},"Machine Learning","machine-learning",{"name":1139,"slug":1140,"type":15},"Migration","migration",{"name":9,"slug":8,"type":15},"2026-07-17T06:07:11.777011",{"slug":1144,"name":1144,"fn":1145,"description":1146,"org":1147,"tags":1148,"stars":1076,"repoUrl":1077,"updatedAt":1153},"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},[1149,1152],{"name":1150,"slug":1151,"type":15},"QA","qa",{"name":996,"slug":997,"type":15},"2026-07-14T05:25:53.673039",{"slug":1155,"name":1155,"fn":1156,"description":1157,"org":1158,"tags":1159,"stars":1076,"repoUrl":1077,"updatedAt":1162},"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},[1160,1161],{"name":968,"slug":969,"type":15},{"name":971,"slug":972,"type":15},"2026-07-14T05:25:49.362534",{"slug":1164,"name":1164,"fn":1165,"description":1166,"org":1167,"tags":1168,"stars":1076,"repoUrl":1077,"updatedAt":1176},"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},[1169,1172,1173],{"name":1170,"slug":1171,"type":15},"Code Review","code-review",{"name":1099,"slug":1100,"type":15},{"name":1174,"slug":1175,"type":15},"Pull Requests","pull-requests","2026-07-14T05:26:01.226578",{"slug":1178,"name":1178,"fn":1179,"description":1180,"org":1181,"tags":1182,"stars":1076,"repoUrl":1077,"updatedAt":1185},"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},[1183,1184],{"name":1150,"slug":1151,"type":15},{"name":996,"slug":997,"type":15},"2026-07-14T05:25:54.928983",{"slug":1187,"name":1187,"fn":1188,"description":1189,"org":1190,"tags":1191,"stars":1076,"repoUrl":1077,"updatedAt":1194},"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},[1192,1193],{"name":1006,"slug":1007,"type":15},{"name":1086,"slug":1087,"type":15},"2026-07-30T05:29:03.275638",496]