[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill-nvidia-jetson-customize-camera":3,"mdc-gohmcs-key":34,"related-org-nvidia-jetson-customize-camera":1256,"related-repo-nvidia-jetson-customize-camera":1417},{"slug":4,"name":4,"fn":5,"description":6,"org":7,"tags":11,"stars":23,"repoUrl":24,"updatedAt":25,"license":26,"forks":27,"topics":28,"repo":29,"sourceUrl":32,"mdContent":33},"jetson-customize-camera","customize camera sensor drivers for Jetson","Enable MIPI\u002FGMSL camera sensors on a Jetson Thor or Orin custom carrier by rendering a kernel-DT overlay from the in-tree sensor DTSI. Do NOT use for UPHY lane allocation or ODMDATA edits.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},"nvidia","NVIDIA","https:\u002F\u002Fpexgzepcugksgbtrxkhf.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Forg-logos\u002Fnvidia.png",[12,16,19,20],{"name":13,"slug":14,"type":15},"Firmware","firmware","tag",{"name":17,"slug":18,"type":15},"Hardware","hardware",{"name":9,"slug":8,"type":15},{"name":21,"slug":22,"type":15},"Engineering","engineering",2473,"https:\u002F\u002Fgithub.com\u002FNVIDIA\u002Fskills","2026-07-14T05:28:03.092885","Apache-2.0",281,[],{"repoUrl":24,"stars":23,"forks":27,"topics":30,"description":31},[],"AI agent skills published by NVIDIA","https:\u002F\u002Fgithub.com\u002FNVIDIA\u002Fskills\u002Ftree\u002FHEAD\u002Fskills\u002Fjetson-customize-camera","---\nname: jetson-customize-camera\ndescription: >-\n  Enable MIPI\u002FGMSL camera sensors on a Jetson Thor or Orin custom\n  carrier by rendering a kernel-DT overlay from the in-tree sensor\n  DTSI. Do NOT use for UPHY lane allocation or ODMDATA edits.\nversion: 0.0.1\nlicense: \"Apache-2.0\"\nmetadata:\n  data-classification: public\n  author: \"Jetson Team\"\n  tags:\n    - bsp\n    - phase-2\n    - io\n    - camera\n    - csi\n  domain: meta\n---\n\n# Customize camera (CSI \u002F MIPI \u002F GMSL sensor bring-up)\n\n## Overview\n\nTegra264 (Thor) and Tegra234 (Orin) expose a single `tegra-capture-vi`\ncontroller fronted by NVCSI and a fixed set of CSI ports. Camera\nbring-up is:\n\n1. **Sensor selection** — picked from the set NVIDIA ships in-tree\n   `.dtsi` references for on the active platform.\n2. **Carrier + module support check** — verified against the Camera\n   Development Guide, Adaptation Guide §Camera, carrier schematic,\n   Module TRM, and carrier pinmap.\n3. **Wiring** — derived from the in-tree\n   `tegra\u003Csoc>-camera-\u003Csensor>*.dtsi` when one exists (**the DTSI IS\n   the wiring source of truth**); captured per-sensor from the user\n   when the sensor is custom.\n4. **Kernel-DT overlay** — cpp-expand the in-tree DTSI, extract its\n   `fragment@N` body, append into the composite custom overlay\n   `.dts` for the active target (per\n   [`..\u002F..\u002Freferences\u002Fbsp-customization-kernel-dtb.md`](..\u002F..\u002Freferences\u002Fbsp-customization-kernel-dtb.md)),\n   verify the composite with `fdtoverlay`.\n   `\u002Fjetson-build-source` compiles the composite and owns the\n   carrier conf's `OVERLAY_DTB_FILE+=` registration.\n\n**Agentic, not table-driven** — sensor list is built at runtime by\nglobbing in-tree per-sensor dtbos. No `_THOR_CAMERAS` dict, no\n`questions.json`, no Python renderer in the question path.\n\n**No ODMDATA edit** — cameras don't consume UPHY lanes (CSI is a\nseparate PHY pool). The skill emits only a kernel-DT overlay; the\nODMDATA line in the carrier conf is untouched by this skill.\n\nThe output is **one commit**:\n- Camera `fragment@N` block (plus `jetson-header-name` on the\n  composite root if not already present) appended to the composite\n  custom overlay `.dts` per\n  [`..\u002F..\u002Freferences\u002Fbsp-customization-kernel-dtb.md`](..\u002F..\u002Freferences\u002Fbsp-customization-kernel-dtb.md)\n  → committed to the `bsp_sources\u002F` hardware repo.\n  `\u002Fjetson-build-source` compiles the composite to `.dtbo` and\n  owns its Makefile + flash-conf registration.\n\n## When to invoke\n\n- The user says \"enable camera\", \"configure CSI\", \"wire a Hawk \u002F\n  Owl \u002F IMX sensor\", \"MIPI camera\", \"GMSL camera\", or asks to bring\n  up `tegra-capture-vi` \u002F NVCSI on a custom carrier.\n- Flash boots but `v4l2-ctl --list-devices` shows no\n  `tegra-capture-vi` channels, OR sensor enumeration on a fresh\n  daughter-card needs to be confirmed.\n- A sensor was previously enabled and the user wants to add another\n  (multi-sensor bring-up).\n\n**Prerequisites:**\n\n- Active profile with `reference_devkit:` + `custom_carrier:` blocks.\n- `\u003Csource.root_path>\u002FLinux_for_Tegra\u002F.git` exists\n  (`\u002Fjetson-init-source`).\n- `\u002Fjetson-derive-carrier` has run — the carrier flash-conf fork is\n  in the overlay tracker.\n- `\u003Csource.root_path>\u002Fbsp_sources\u002Fhardware\u002Fnvidia\u002F\u003Cchip-dir>\u002Fnv-public\u002Foverlay\u002F`\n  exists and contains the in-tree per-sensor `.dtsi` files (sourced\n  by `\u002Fjetson-init-source`'s Branch A archive extract).\n- `\u003Csource.root_path>\u002Fbsp_sources\u002Fkernel\u002Fkernel-noble\u002Finclude\u002Fdt-bindings\u002F`\n  contains the macro headers cpp needs (`source_sync.sh` may need to\n  run if Branch B was used — see Step 5a.i below).\n- Source-of-truth docs registered or supplied at prompt:\n  Camera Development Guide (in `bsp_developer_guide` mirror or\n  separate path), Adaptation Guide §Camera, carrier schematic, SoC\n  TRM, Module Design Guide.\n- `dtc`, `cpp`, `fdtoverlay` on PATH.\n\n## Procedure\n\nDetailed step-by-step procedure (Steps 1–7, with all tables, code\nblocks, and gates) lives in\n[`references\u002Fprocedure.md`](references\u002Fprocedure.md). Summary:\n\n1. **Step 1** — Resolve active target + open source-of-truth docs.\n2. **Step 2** — Enumerate supported sensors by globbing in-tree\n   per-platform camera dtbos; classify as DPHY-direct \u002F GMSL \u002F\n   custom. Never invent sensors.\n3. **Step 3 \u002F 3a** — Cross-check carrier + module support against\n   DTSI, Camera Development Guide, Adaptation Guide §Camera, SoC TRM,\n   Module Design Guide, schematic, and carrier pinmap. Render the\n   wiring table FIRST, then issue the confirm-or-customize gate.\n4. **Step 4** (custom path only) — Batched per-sensor wiring\n   questions auto-filled from the carrier pinmap.\n5. **Step 5** — Append exactly ONE `\u002F* custom-bsp: camera:\u003Csensor> *\u002F`\n   fragment to the composite custom overlay `.dts` (see\n   [`..\u002F..\u002Freferences\u002Fbsp-customization-kernel-dtb.md`](..\u002F..\u002Freferences\u002Fbsp-customization-kernel-dtb.md)).\n   Clone path cpp-expands the in-tree DTSI; custom path splices Step-4\n   answers + mode tables in-place. Idempotently set\n   `jetson-header-name` on the composite root. Verify with\n   `dtc` + `fdtoverlay` (pre-compile single-fragment gate;\n   post-compile deep-tree uniqueness gate). Commit via the\n   workflow's commit-message preview gate.\n6. **Step 6** — Verify ancillary CAM pin SFIOs (`cam_i2c_*`,\n   `extperiph\u003Cm>_clk`, reset\u002FPWDN\u002FPWR_EN GPIOs) via\n   `pin_verifier.py`; route mismatches to `\u002Fjetson-customize-pinmux`.\n7. **Step 7** — Atomic-write run-state JSON sidecar at\n   `\u003Cworkspace>\u002Ftarget-platform\u002F\u003Cprofile-stem>.jetson-customize-camera.json`\n   and emit the headline, then drive the downstream next-step chain via\n   sequential `AskUserQuestion` prompts per `references\u002Fprocedure.md`\n   Step 7. **The chain is a documented workflow gate, not a clarifying\n   question — auto-mode does NOT exempt it.** Never substitute a\n   printed \"Next step: …\" line for the prompts.\n\n## Gotchas\n\n- **Dual-fragment trap.** Contribute exactly ONE camera-tagged\n  `fragment@N` to the composite. A second one carrying status\n  overrides triggers dtc deep-merge → duplicate sibling subtrees\n  (e.g. two `tca9546@70`) → runtime first-match drops the dtsi-\n  supplied deep tree → camera silently doesn't enumerate. Gate on\n  this skill's marker only (Step 5c).\n- **Composite root `compatible` is owned globally, not by this\n  skill.** Don't widen from any in-tree per-sensor dtbo's\n  `compatible` (devkit-SKU-gated). Fix the composite root if needed.\n- **`jetson-header-name` from any in-tree per-sensor dtbo.** Fixed,\n  carrier-agnostic; read once, paste onto the metadata root.\n- **DO NOT also append the in-tree per-sensor dtbo to\n  `OVERLAY_DTB_FILE`.** Registering both your rendered overlay AND\n  the in-tree `tegra\u003Csoc>-p3971-camera-\u003Csensor>-overlay.dtbo`\n  produces a phantom subdev bind that bricks camera enumeration.\n- **Stub overlay is a known footgun.** Committing\n  `tegra-capture-vi { status=\"okay\"; num-channels=\u003CN>; }` with no\n  ports \u002F sensor \u002F nvcsi body bricks the camera (`all channel init\n  failed`). Splice the FULL sensor body via cpp + dtc.\n- **Sensor mode tables must be spliced, never hand-authored.**\n  `mode\u003CN>`, `sensor_modes`, `pixel_phase` — copy verbatim from the\n  closest in-tree DTSI.\n- **`camera_common_regulator_get (null) ERR: -EINVAL`** = missing\n  `avdd-reg` \u002F `iovdd-reg` \u002F `dvdd-reg` strings — splice the FULL\n  sensor body; always-on rails fall back to dummy regulator.\n- **External `&label` refs must exist in base DTB's `__symbols__`.**\n  Use `target-path = \"\u002Ftegra-capture-vi\"` when the label is absent;\n  `fdtoverlay` exits non-zero with `FDT_ERR_NOTFOUND` otherwise.\n- **`cpp` failure on `dt-bindings\u002Fgpio\u002Fgpio.h: No such file`** =\n  L4T source tree isn't staged. Re-run `\u002Fjetson-init-source` (Branch\n  B's `source_sync.sh` fetches the headers). Never fabricate the\n  macro expansion.\n- **No ODMDATA edit, no flash-conf edit.** Camera doesn't consume\n  UPHY lanes. The carrier conf's `ODMDATA=\"...\"` is untouched.\n  `OVERLAY_DTB_FILE+=` is owned by `\u002Fjetson-build-source` Step\n  5.0a — this skill never touches the carrier flash conf.\n- **Don't touch the upstream BSP at `\u003Cbsp_image.root_path>`.** All\n  edits land in `\u003Csource.root_path>\u002FLinux_for_Tegra\u002F` (overlay\n  tracker) and `\u003Csource.root_path>\u002Fbsp_sources\u002F` (overlay `.dts`)\n  under the pristine + customization commit pattern.\n\n## References\n\n- [`references\u002Fprocedure.md`](references\u002Fprocedure.md) — full\n  step-by-step Steps 1–7 procedure (extracted from this SKILL.md).\n- [`references\u002Fcsi-dt-bindings.md`](references\u002Fcsi-dt-bindings.md) —\n  CSI \u002F nvcsi \u002F vi DT binding reference notes.\n- [`references\u002Foverlay-template.md`](references\u002Foverlay-template.md) —\n  guidance on the metadata-root + clone-body overlay shape.\n- [`references\u002Fcamera-overlay-templates\u002F`](references\u002Fcamera-overlay-templates\u002F)\n  — starter `.dts.tmpl` templates: `dphy-direct.dts.tmpl`,\n  `gmsl-serdes.dts.tmpl`.\n- [`..\u002F..\u002Fscripts\u002Fpin_verifier.py`](..\u002F..\u002Fscripts\u002Fpin_verifier.py)\n  — shared HSIO pin verifier (Step 6).\n- [`..\u002F..\u002Freferences\u002Fplatform_template.yaml`](..\u002F..\u002Freferences\u002Fplatform_template.yaml)\n  — `documents:` block consumed by Step 1.\n- [`..\u002F..\u002Fcontext\u002Fbsp-customization-workflow.md`](..\u002F..\u002Fcontext\u002Fbsp-customization-workflow.md#workflow-invariants)\n  — overlay edit protocol.\n- [`..\u002Fjetson-customize-pinmux\u002FSKILL.md`](..\u002Fjetson-customize-pinmux\u002FSKILL.md) —\n  sibling skill auto-invoked by Step 6 to fix HSIO pin SFIO\n  mismatches (CAM I²C, MCLK, reset GPIOs).\n- [`..\u002Fjetson-derive-carrier\u002FSKILL.md`](..\u002Fjetson-derive-carrier\u002FSKILL.md)\n  — must run first; produces the carrier base overlay (the\n  `*-dynamic.dtbo`) this skill's composite stacks after.\n- [`..\u002Fjetson-init-source\u002FSKILL.md`](..\u002Fjetson-init-source\u002FSKILL.md) —\n  produces the overlay tracker + `bsp_sources` repo (with the\n  `hardware\u002Fnvidia\u002F\u003Cchip-dir>\u002F` per-sensor DTSI tree) this skill\n  reads and commits into.\n",{"data":35,"body":47},{"name":4,"description":6,"version":36,"license":26,"metadata":37},"0.0.1",{"data-classification":38,"author":39,"tags":40,"domain":46},"public","Jetson Team",[41,42,43,44,45],"bsp","phase-2","io","camera","csi","meta",{"type":48,"children":49},"root",[50,59,66,81,202,228,238,250,314,320,360,368,504,510,526,706,712,1053,1059],{"type":51,"tag":52,"props":53,"children":55},"element","h1",{"id":54},"customize-camera-csi-mipi-gmsl-sensor-bring-up",[56],{"type":57,"value":58},"text","Customize camera (CSI \u002F MIPI \u002F GMSL sensor bring-up)",{"type":51,"tag":60,"props":61,"children":63},"h2",{"id":62},"overview",[64],{"type":57,"value":65},"Overview",{"type":51,"tag":67,"props":68,"children":69},"p",{},[70,72,79],{"type":57,"value":71},"Tegra264 (Thor) and Tegra234 (Orin) expose a single ",{"type":51,"tag":73,"props":74,"children":76},"code",{"className":75},[],[77],{"type":57,"value":78},"tegra-capture-vi",{"type":57,"value":80},"\ncontroller fronted by NVCSI and a fixed set of CSI ports. Camera\nbring-up is:",{"type":51,"tag":82,"props":83,"children":84},"ol",{},[85,105,115,140],{"type":51,"tag":86,"props":87,"children":88},"li",{},[89,95,97,103],{"type":51,"tag":90,"props":91,"children":92},"strong",{},[93],{"type":57,"value":94},"Sensor selection",{"type":57,"value":96}," — picked from the set NVIDIA ships in-tree\n",{"type":51,"tag":73,"props":98,"children":100},{"className":99},[],[101],{"type":57,"value":102},".dtsi",{"type":57,"value":104}," references for on the active platform.",{"type":51,"tag":86,"props":106,"children":107},{},[108,113],{"type":51,"tag":90,"props":109,"children":110},{},[111],{"type":57,"value":112},"Carrier + module support check",{"type":57,"value":114}," — verified against the Camera\nDevelopment Guide, Adaptation Guide §Camera, carrier schematic,\nModule TRM, and carrier pinmap.",{"type":51,"tag":86,"props":116,"children":117},{},[118,123,125,131,133,138],{"type":51,"tag":90,"props":119,"children":120},{},[121],{"type":57,"value":122},"Wiring",{"type":57,"value":124}," — derived from the in-tree\n",{"type":51,"tag":73,"props":126,"children":128},{"className":127},[],[129],{"type":57,"value":130},"tegra\u003Csoc>-camera-\u003Csensor>*.dtsi",{"type":57,"value":132}," when one exists (",{"type":51,"tag":90,"props":134,"children":135},{},[136],{"type":57,"value":137},"the DTSI IS\nthe wiring source of truth",{"type":57,"value":139},"); captured per-sensor from the user\nwhen the sensor is custom.",{"type":51,"tag":86,"props":141,"children":142},{},[143,148,150,156,158,164,166,176,178,184,186,192,194,200],{"type":51,"tag":90,"props":144,"children":145},{},[146],{"type":57,"value":147},"Kernel-DT overlay",{"type":57,"value":149}," — cpp-expand the in-tree DTSI, extract its\n",{"type":51,"tag":73,"props":151,"children":153},{"className":152},[],[154],{"type":57,"value":155},"fragment@N",{"type":57,"value":157}," body, append into the composite custom overlay\n",{"type":51,"tag":73,"props":159,"children":161},{"className":160},[],[162],{"type":57,"value":163},".dts",{"type":57,"value":165}," for the active target (per\n",{"type":51,"tag":167,"props":168,"children":170},"a",{"href":169},"..\u002F..\u002Freferences\u002Fbsp-customization-kernel-dtb.md",[171],{"type":51,"tag":73,"props":172,"children":174},{"className":173},[],[175],{"type":57,"value":169},{"type":57,"value":177},"),\nverify the composite with ",{"type":51,"tag":73,"props":179,"children":181},{"className":180},[],[182],{"type":57,"value":183},"fdtoverlay",{"type":57,"value":185},".\n",{"type":51,"tag":73,"props":187,"children":189},{"className":188},[],[190],{"type":57,"value":191},"\u002Fjetson-build-source",{"type":57,"value":193}," compiles the composite and owns the\ncarrier conf's ",{"type":51,"tag":73,"props":195,"children":197},{"className":196},[],[198],{"type":57,"value":199},"OVERLAY_DTB_FILE+=",{"type":57,"value":201}," registration.",{"type":51,"tag":67,"props":203,"children":204},{},[205,210,212,218,220,226],{"type":51,"tag":90,"props":206,"children":207},{},[208],{"type":57,"value":209},"Agentic, not table-driven",{"type":57,"value":211}," — sensor list is built at runtime by\nglobbing in-tree per-sensor dtbos. No ",{"type":51,"tag":73,"props":213,"children":215},{"className":214},[],[216],{"type":57,"value":217},"_THOR_CAMERAS",{"type":57,"value":219}," dict, no\n",{"type":51,"tag":73,"props":221,"children":223},{"className":222},[],[224],{"type":57,"value":225},"questions.json",{"type":57,"value":227},", no Python renderer in the question path.",{"type":51,"tag":67,"props":229,"children":230},{},[231,236],{"type":51,"tag":90,"props":232,"children":233},{},[234],{"type":57,"value":235},"No ODMDATA edit",{"type":57,"value":237}," — cameras don't consume UPHY lanes (CSI is a\nseparate PHY pool). The skill emits only a kernel-DT overlay; the\nODMDATA line in the carrier conf is untouched by this skill.",{"type":51,"tag":67,"props":239,"children":240},{},[241,243,248],{"type":57,"value":242},"The output is ",{"type":51,"tag":90,"props":244,"children":245},{},[246],{"type":57,"value":247},"one commit",{"type":57,"value":249},":",{"type":51,"tag":251,"props":252,"children":253},"ul",{},[254],{"type":51,"tag":86,"props":255,"children":256},{},[257,259,264,266,272,274,279,281,289,291,297,299,304,306,312],{"type":57,"value":258},"Camera ",{"type":51,"tag":73,"props":260,"children":262},{"className":261},[],[263],{"type":57,"value":155},{"type":57,"value":265}," block (plus ",{"type":51,"tag":73,"props":267,"children":269},{"className":268},[],[270],{"type":57,"value":271},"jetson-header-name",{"type":57,"value":273}," on the\ncomposite root if not already present) appended to the composite\ncustom overlay ",{"type":51,"tag":73,"props":275,"children":277},{"className":276},[],[278],{"type":57,"value":163},{"type":57,"value":280}," per\n",{"type":51,"tag":167,"props":282,"children":283},{"href":169},[284],{"type":51,"tag":73,"props":285,"children":287},{"className":286},[],[288],{"type":57,"value":169},{"type":57,"value":290},"\n→ committed to the ",{"type":51,"tag":73,"props":292,"children":294},{"className":293},[],[295],{"type":57,"value":296},"bsp_sources\u002F",{"type":57,"value":298}," hardware repo.\n",{"type":51,"tag":73,"props":300,"children":302},{"className":301},[],[303],{"type":57,"value":191},{"type":57,"value":305}," compiles the composite to ",{"type":51,"tag":73,"props":307,"children":309},{"className":308},[],[310],{"type":57,"value":311},".dtbo",{"type":57,"value":313}," and\nowns its Makefile + flash-conf registration.",{"type":51,"tag":60,"props":315,"children":317},{"id":316},"when-to-invoke",[318],{"type":57,"value":319},"When to invoke",{"type":51,"tag":251,"props":321,"children":322},{},[323,335,355],{"type":51,"tag":86,"props":324,"children":325},{},[326,328,333],{"type":57,"value":327},"The user says \"enable camera\", \"configure CSI\", \"wire a Hawk \u002F\nOwl \u002F IMX sensor\", \"MIPI camera\", \"GMSL camera\", or asks to bring\nup ",{"type":51,"tag":73,"props":329,"children":331},{"className":330},[],[332],{"type":57,"value":78},{"type":57,"value":334}," \u002F NVCSI on a custom carrier.",{"type":51,"tag":86,"props":336,"children":337},{},[338,340,346,348,353],{"type":57,"value":339},"Flash boots but ",{"type":51,"tag":73,"props":341,"children":343},{"className":342},[],[344],{"type":57,"value":345},"v4l2-ctl --list-devices",{"type":57,"value":347}," shows no\n",{"type":51,"tag":73,"props":349,"children":351},{"className":350},[],[352],{"type":57,"value":78},{"type":57,"value":354}," channels, OR sensor enumeration on a fresh\ndaughter-card needs to be confirmed.",{"type":51,"tag":86,"props":356,"children":357},{},[358],{"type":57,"value":359},"A sensor was previously enabled and the user wants to add another\n(multi-sensor bring-up).",{"type":51,"tag":67,"props":361,"children":362},{},[363],{"type":51,"tag":90,"props":364,"children":365},{},[366],{"type":57,"value":367},"Prerequisites:",{"type":51,"tag":251,"props":369,"children":370},{},[371,392,411,422,447,466,479],{"type":51,"tag":86,"props":372,"children":373},{},[374,376,382,384,390],{"type":57,"value":375},"Active profile with ",{"type":51,"tag":73,"props":377,"children":379},{"className":378},[],[380],{"type":57,"value":381},"reference_devkit:",{"type":57,"value":383}," + ",{"type":51,"tag":73,"props":385,"children":387},{"className":386},[],[388],{"type":57,"value":389},"custom_carrier:",{"type":57,"value":391}," blocks.",{"type":51,"tag":86,"props":393,"children":394},{},[395,401,403,409],{"type":51,"tag":73,"props":396,"children":398},{"className":397},[],[399],{"type":57,"value":400},"\u003Csource.root_path>\u002FLinux_for_Tegra\u002F.git",{"type":57,"value":402}," exists\n(",{"type":51,"tag":73,"props":404,"children":406},{"className":405},[],[407],{"type":57,"value":408},"\u002Fjetson-init-source",{"type":57,"value":410},").",{"type":51,"tag":86,"props":412,"children":413},{},[414,420],{"type":51,"tag":73,"props":415,"children":417},{"className":416},[],[418],{"type":57,"value":419},"\u002Fjetson-derive-carrier",{"type":57,"value":421}," has run — the carrier flash-conf fork is\nin the overlay tracker.",{"type":51,"tag":86,"props":423,"children":424},{},[425,431,433,438,440,445],{"type":51,"tag":73,"props":426,"children":428},{"className":427},[],[429],{"type":57,"value":430},"\u003Csource.root_path>\u002Fbsp_sources\u002Fhardware\u002Fnvidia\u002F\u003Cchip-dir>\u002Fnv-public\u002Foverlay\u002F",{"type":57,"value":432},"\nexists and contains the in-tree per-sensor ",{"type":51,"tag":73,"props":434,"children":436},{"className":435},[],[437],{"type":57,"value":102},{"type":57,"value":439}," files (sourced\nby ",{"type":51,"tag":73,"props":441,"children":443},{"className":442},[],[444],{"type":57,"value":408},{"type":57,"value":446},"'s Branch A archive extract).",{"type":51,"tag":86,"props":448,"children":449},{},[450,456,458,464],{"type":51,"tag":73,"props":451,"children":453},{"className":452},[],[454],{"type":57,"value":455},"\u003Csource.root_path>\u002Fbsp_sources\u002Fkernel\u002Fkernel-noble\u002Finclude\u002Fdt-bindings\u002F",{"type":57,"value":457},"\ncontains the macro headers cpp needs (",{"type":51,"tag":73,"props":459,"children":461},{"className":460},[],[462],{"type":57,"value":463},"source_sync.sh",{"type":57,"value":465}," may need to\nrun if Branch B was used — see Step 5a.i below).",{"type":51,"tag":86,"props":467,"children":468},{},[469,471,477],{"type":57,"value":470},"Source-of-truth docs registered or supplied at prompt:\nCamera Development Guide (in ",{"type":51,"tag":73,"props":472,"children":474},{"className":473},[],[475],{"type":57,"value":476},"bsp_developer_guide",{"type":57,"value":478}," mirror or\nseparate path), Adaptation Guide §Camera, carrier schematic, SoC\nTRM, Module Design Guide.",{"type":51,"tag":86,"props":480,"children":481},{},[482,488,490,496,497,502],{"type":51,"tag":73,"props":483,"children":485},{"className":484},[],[486],{"type":57,"value":487},"dtc",{"type":57,"value":489},", ",{"type":51,"tag":73,"props":491,"children":493},{"className":492},[],[494],{"type":57,"value":495},"cpp",{"type":57,"value":489},{"type":51,"tag":73,"props":498,"children":500},{"className":499},[],[501],{"type":57,"value":183},{"type":57,"value":503}," on PATH.",{"type":51,"tag":60,"props":505,"children":507},{"id":506},"procedure",[508],{"type":57,"value":509},"Procedure",{"type":51,"tag":67,"props":511,"children":512},{},[513,515,524],{"type":57,"value":514},"Detailed step-by-step procedure (Steps 1–7, with all tables, code\nblocks, and gates) lives in\n",{"type":51,"tag":167,"props":516,"children":518},{"href":517},"references\u002Fprocedure.md",[519],{"type":51,"tag":73,"props":520,"children":522},{"className":521},[],[523],{"type":57,"value":517},{"type":57,"value":525},". Summary:",{"type":51,"tag":82,"props":527,"children":528},{},[529,539,549,559,569,624,666],{"type":51,"tag":86,"props":530,"children":531},{},[532,537],{"type":51,"tag":90,"props":533,"children":534},{},[535],{"type":57,"value":536},"Step 1",{"type":57,"value":538}," — Resolve active target + open source-of-truth docs.",{"type":51,"tag":86,"props":540,"children":541},{},[542,547],{"type":51,"tag":90,"props":543,"children":544},{},[545],{"type":57,"value":546},"Step 2",{"type":57,"value":548}," — Enumerate supported sensors by globbing in-tree\nper-platform camera dtbos; classify as DPHY-direct \u002F GMSL \u002F\ncustom. Never invent sensors.",{"type":51,"tag":86,"props":550,"children":551},{},[552,557],{"type":51,"tag":90,"props":553,"children":554},{},[555],{"type":57,"value":556},"Step 3 \u002F 3a",{"type":57,"value":558}," — Cross-check carrier + module support against\nDTSI, Camera Development Guide, Adaptation Guide §Camera, SoC TRM,\nModule Design Guide, schematic, and carrier pinmap. Render the\nwiring table FIRST, then issue the confirm-or-customize gate.",{"type":51,"tag":86,"props":560,"children":561},{},[562,567],{"type":51,"tag":90,"props":563,"children":564},{},[565],{"type":57,"value":566},"Step 4",{"type":57,"value":568}," (custom path only) — Batched per-sensor wiring\nquestions auto-filled from the carrier pinmap.",{"type":51,"tag":86,"props":570,"children":571},{},[572,577,579,585,587,592,594,602,604,609,611,616,617,622],{"type":51,"tag":90,"props":573,"children":574},{},[575],{"type":57,"value":576},"Step 5",{"type":57,"value":578}," — Append exactly ONE ",{"type":51,"tag":73,"props":580,"children":582},{"className":581},[],[583],{"type":57,"value":584},"\u002F* custom-bsp: camera:\u003Csensor> *\u002F",{"type":57,"value":586},"\nfragment to the composite custom overlay ",{"type":51,"tag":73,"props":588,"children":590},{"className":589},[],[591],{"type":57,"value":163},{"type":57,"value":593}," (see\n",{"type":51,"tag":167,"props":595,"children":596},{"href":169},[597],{"type":51,"tag":73,"props":598,"children":600},{"className":599},[],[601],{"type":57,"value":169},{"type":57,"value":603},").\nClone path cpp-expands the in-tree DTSI; custom path splices Step-4\nanswers + mode tables in-place. Idempotently set\n",{"type":51,"tag":73,"props":605,"children":607},{"className":606},[],[608],{"type":57,"value":271},{"type":57,"value":610}," on the composite root. Verify with\n",{"type":51,"tag":73,"props":612,"children":614},{"className":613},[],[615],{"type":57,"value":487},{"type":57,"value":383},{"type":51,"tag":73,"props":618,"children":620},{"className":619},[],[621],{"type":57,"value":183},{"type":57,"value":623}," (pre-compile single-fragment gate;\npost-compile deep-tree uniqueness gate). Commit via the\nworkflow's commit-message preview gate.",{"type":51,"tag":86,"props":625,"children":626},{},[627,632,634,640,642,648,650,656,658,664],{"type":51,"tag":90,"props":628,"children":629},{},[630],{"type":57,"value":631},"Step 6",{"type":57,"value":633}," — Verify ancillary CAM pin SFIOs (",{"type":51,"tag":73,"props":635,"children":637},{"className":636},[],[638],{"type":57,"value":639},"cam_i2c_*",{"type":57,"value":641},",\n",{"type":51,"tag":73,"props":643,"children":645},{"className":644},[],[646],{"type":57,"value":647},"extperiph\u003Cm>_clk",{"type":57,"value":649},", reset\u002FPWDN\u002FPWR_EN GPIOs) via\n",{"type":51,"tag":73,"props":651,"children":653},{"className":652},[],[654],{"type":57,"value":655},"pin_verifier.py",{"type":57,"value":657},"; route mismatches to ",{"type":51,"tag":73,"props":659,"children":661},{"className":660},[],[662],{"type":57,"value":663},"\u002Fjetson-customize-pinmux",{"type":57,"value":665},".",{"type":51,"tag":86,"props":667,"children":668},{},[669,674,676,682,684,690,692,697,699,704],{"type":51,"tag":90,"props":670,"children":671},{},[672],{"type":57,"value":673},"Step 7",{"type":57,"value":675}," — Atomic-write run-state JSON sidecar at\n",{"type":51,"tag":73,"props":677,"children":679},{"className":678},[],[680],{"type":57,"value":681},"\u003Cworkspace>\u002Ftarget-platform\u002F\u003Cprofile-stem>.jetson-customize-camera.json",{"type":57,"value":683},"\nand emit the headline, then drive the downstream next-step chain via\nsequential ",{"type":51,"tag":73,"props":685,"children":687},{"className":686},[],[688],{"type":57,"value":689},"AskUserQuestion",{"type":57,"value":691}," prompts per ",{"type":51,"tag":73,"props":693,"children":695},{"className":694},[],[696],{"type":57,"value":517},{"type":57,"value":698},"\nStep 7. ",{"type":51,"tag":90,"props":700,"children":701},{},[702],{"type":57,"value":703},"The chain is a documented workflow gate, not a clarifying\nquestion — auto-mode does NOT exempt it.",{"type":57,"value":705}," Never substitute a\nprinted \"Next step: …\" line for the prompts.",{"type":51,"tag":60,"props":707,"children":709},{"id":708},"gotchas",[710],{"type":57,"value":711},"Gotchas",{"type":51,"tag":251,"props":713,"children":714},{},[715,740,765,780,805,831,861,898,946,981,1013],{"type":51,"tag":86,"props":716,"children":717},{},[718,723,725,730,732,738],{"type":51,"tag":90,"props":719,"children":720},{},[721],{"type":57,"value":722},"Dual-fragment trap.",{"type":57,"value":724}," Contribute exactly ONE camera-tagged\n",{"type":51,"tag":73,"props":726,"children":728},{"className":727},[],[729],{"type":57,"value":155},{"type":57,"value":731}," to the composite. A second one carrying status\noverrides triggers dtc deep-merge → duplicate sibling subtrees\n(e.g. two ",{"type":51,"tag":73,"props":733,"children":735},{"className":734},[],[736],{"type":57,"value":737},"tca9546@70",{"type":57,"value":739},") → runtime first-match drops the dtsi-\nsupplied deep tree → camera silently doesn't enumerate. Gate on\nthis skill's marker only (Step 5c).",{"type":51,"tag":86,"props":741,"children":742},{},[743,756,758,763],{"type":51,"tag":90,"props":744,"children":745},{},[746,748,754],{"type":57,"value":747},"Composite root ",{"type":51,"tag":73,"props":749,"children":751},{"className":750},[],[752],{"type":57,"value":753},"compatible",{"type":57,"value":755}," is owned globally, not by this\nskill.",{"type":57,"value":757}," Don't widen from any in-tree per-sensor dtbo's\n",{"type":51,"tag":73,"props":759,"children":761},{"className":760},[],[762],{"type":57,"value":753},{"type":57,"value":764}," (devkit-SKU-gated). Fix the composite root if needed.",{"type":51,"tag":86,"props":766,"children":767},{},[768,778],{"type":51,"tag":90,"props":769,"children":770},{},[771,776],{"type":51,"tag":73,"props":772,"children":774},{"className":773},[],[775],{"type":57,"value":271},{"type":57,"value":777}," from any in-tree per-sensor dtbo.",{"type":57,"value":779}," Fixed,\ncarrier-agnostic; read once, paste onto the metadata root.",{"type":51,"tag":86,"props":781,"children":782},{},[783,795,797,803],{"type":51,"tag":90,"props":784,"children":785},{},[786,788,794],{"type":57,"value":787},"DO NOT also append the in-tree per-sensor dtbo to\n",{"type":51,"tag":73,"props":789,"children":791},{"className":790},[],[792],{"type":57,"value":793},"OVERLAY_DTB_FILE",{"type":57,"value":665},{"type":57,"value":796}," Registering both your rendered overlay AND\nthe in-tree ",{"type":51,"tag":73,"props":798,"children":800},{"className":799},[],[801],{"type":57,"value":802},"tegra\u003Csoc>-p3971-camera-\u003Csensor>-overlay.dtbo",{"type":57,"value":804},"\nproduces a phantom subdev bind that bricks camera enumeration.",{"type":51,"tag":86,"props":806,"children":807},{},[808,813,815,821,823,829],{"type":51,"tag":90,"props":809,"children":810},{},[811],{"type":57,"value":812},"Stub overlay is a known footgun.",{"type":57,"value":814}," Committing\n",{"type":51,"tag":73,"props":816,"children":818},{"className":817},[],[819],{"type":57,"value":820},"tegra-capture-vi { status=\"okay\"; num-channels=\u003CN>; }",{"type":57,"value":822}," with no\nports \u002F sensor \u002F nvcsi body bricks the camera (",{"type":51,"tag":73,"props":824,"children":826},{"className":825},[],[827],{"type":57,"value":828},"all channel init failed",{"type":57,"value":830},"). Splice the FULL sensor body via cpp + dtc.",{"type":51,"tag":86,"props":832,"children":833},{},[834,839,845,846,852,853,859],{"type":51,"tag":90,"props":835,"children":836},{},[837],{"type":57,"value":838},"Sensor mode tables must be spliced, never hand-authored.",{"type":51,"tag":73,"props":840,"children":842},{"className":841},[],[843],{"type":57,"value":844},"mode\u003CN>",{"type":57,"value":489},{"type":51,"tag":73,"props":847,"children":849},{"className":848},[],[850],{"type":57,"value":851},"sensor_modes",{"type":57,"value":489},{"type":51,"tag":73,"props":854,"children":856},{"className":855},[],[857],{"type":57,"value":858},"pixel_phase",{"type":57,"value":860}," — copy verbatim from the\nclosest in-tree DTSI.",{"type":51,"tag":86,"props":862,"children":863},{},[864,873,875,881,883,889,890,896],{"type":51,"tag":90,"props":865,"children":866},{},[867],{"type":51,"tag":73,"props":868,"children":870},{"className":869},[],[871],{"type":57,"value":872},"camera_common_regulator_get (null) ERR: -EINVAL",{"type":57,"value":874}," = missing\n",{"type":51,"tag":73,"props":876,"children":878},{"className":877},[],[879],{"type":57,"value":880},"avdd-reg",{"type":57,"value":882}," \u002F ",{"type":51,"tag":73,"props":884,"children":886},{"className":885},[],[887],{"type":57,"value":888},"iovdd-reg",{"type":57,"value":882},{"type":51,"tag":73,"props":891,"children":893},{"className":892},[],[894],{"type":57,"value":895},"dvdd-reg",{"type":57,"value":897}," strings — splice the FULL\nsensor body; always-on rails fall back to dummy regulator.",{"type":51,"tag":86,"props":899,"children":900},{},[901,921,923,929,931,936,938,944],{"type":51,"tag":90,"props":902,"children":903},{},[904,906,912,914,920],{"type":57,"value":905},"External ",{"type":51,"tag":73,"props":907,"children":909},{"className":908},[],[910],{"type":57,"value":911},"&label",{"type":57,"value":913}," refs must exist in base DTB's ",{"type":51,"tag":73,"props":915,"children":917},{"className":916},[],[918],{"type":57,"value":919},"__symbols__",{"type":57,"value":665},{"type":57,"value":922},"\nUse ",{"type":51,"tag":73,"props":924,"children":926},{"className":925},[],[927],{"type":57,"value":928},"target-path = \"\u002Ftegra-capture-vi\"",{"type":57,"value":930}," when the label is absent;\n",{"type":51,"tag":73,"props":932,"children":934},{"className":933},[],[935],{"type":57,"value":183},{"type":57,"value":937}," exits non-zero with ",{"type":51,"tag":73,"props":939,"children":941},{"className":940},[],[942],{"type":57,"value":943},"FDT_ERR_NOTFOUND",{"type":57,"value":945}," otherwise.",{"type":51,"tag":86,"props":947,"children":948},{},[949,965,967,972,974,979],{"type":51,"tag":90,"props":950,"children":951},{},[952,957,959],{"type":51,"tag":73,"props":953,"children":955},{"className":954},[],[956],{"type":57,"value":495},{"type":57,"value":958}," failure on ",{"type":51,"tag":73,"props":960,"children":962},{"className":961},[],[963],{"type":57,"value":964},"dt-bindings\u002Fgpio\u002Fgpio.h: No such file",{"type":57,"value":966}," =\nL4T source tree isn't staged. Re-run ",{"type":51,"tag":73,"props":968,"children":970},{"className":969},[],[971],{"type":57,"value":408},{"type":57,"value":973}," (Branch\nB's ",{"type":51,"tag":73,"props":975,"children":977},{"className":976},[],[978],{"type":57,"value":463},{"type":57,"value":980}," fetches the headers). Never fabricate the\nmacro expansion.",{"type":51,"tag":86,"props":982,"children":983},{},[984,989,991,997,999,1004,1006,1011],{"type":51,"tag":90,"props":985,"children":986},{},[987],{"type":57,"value":988},"No ODMDATA edit, no flash-conf edit.",{"type":57,"value":990}," Camera doesn't consume\nUPHY lanes. The carrier conf's ",{"type":51,"tag":73,"props":992,"children":994},{"className":993},[],[995],{"type":57,"value":996},"ODMDATA=\"...\"",{"type":57,"value":998}," is untouched.\n",{"type":51,"tag":73,"props":1000,"children":1002},{"className":1001},[],[1003],{"type":57,"value":199},{"type":57,"value":1005}," is owned by ",{"type":51,"tag":73,"props":1007,"children":1009},{"className":1008},[],[1010],{"type":57,"value":191},{"type":57,"value":1012}," Step\n5.0a — this skill never touches the carrier flash conf.",{"type":51,"tag":86,"props":1014,"children":1015},{},[1016,1028,1030,1036,1038,1044,1046,1051],{"type":51,"tag":90,"props":1017,"children":1018},{},[1019,1021,1027],{"type":57,"value":1020},"Don't touch the upstream BSP at ",{"type":51,"tag":73,"props":1022,"children":1024},{"className":1023},[],[1025],{"type":57,"value":1026},"\u003Cbsp_image.root_path>",{"type":57,"value":665},{"type":57,"value":1029}," All\nedits land in ",{"type":51,"tag":73,"props":1031,"children":1033},{"className":1032},[],[1034],{"type":57,"value":1035},"\u003Csource.root_path>\u002FLinux_for_Tegra\u002F",{"type":57,"value":1037}," (overlay\ntracker) and ",{"type":51,"tag":73,"props":1039,"children":1041},{"className":1040},[],[1042],{"type":57,"value":1043},"\u003Csource.root_path>\u002Fbsp_sources\u002F",{"type":57,"value":1045}," (overlay ",{"type":51,"tag":73,"props":1047,"children":1049},{"className":1048},[],[1050],{"type":57,"value":163},{"type":57,"value":1052},")\nunder the pristine + customization commit pattern.",{"type":51,"tag":60,"props":1054,"children":1056},{"id":1055},"references",[1057],{"type":57,"value":1058},"References",{"type":51,"tag":251,"props":1060,"children":1061},{},[1062,1075,1089,1103,1139,1153,1175,1190,1204,1226],{"type":51,"tag":86,"props":1063,"children":1064},{},[1065,1073],{"type":51,"tag":167,"props":1066,"children":1067},{"href":517},[1068],{"type":51,"tag":73,"props":1069,"children":1071},{"className":1070},[],[1072],{"type":57,"value":517},{"type":57,"value":1074}," — full\nstep-by-step Steps 1–7 procedure (extracted from this SKILL.md).",{"type":51,"tag":86,"props":1076,"children":1077},{},[1078,1087],{"type":51,"tag":167,"props":1079,"children":1081},{"href":1080},"references\u002Fcsi-dt-bindings.md",[1082],{"type":51,"tag":73,"props":1083,"children":1085},{"className":1084},[],[1086],{"type":57,"value":1080},{"type":57,"value":1088}," —\nCSI \u002F nvcsi \u002F vi DT binding reference notes.",{"type":51,"tag":86,"props":1090,"children":1091},{},[1092,1101],{"type":51,"tag":167,"props":1093,"children":1095},{"href":1094},"references\u002Foverlay-template.md",[1096],{"type":51,"tag":73,"props":1097,"children":1099},{"className":1098},[],[1100],{"type":57,"value":1094},{"type":57,"value":1102}," —\nguidance on the metadata-root + clone-body overlay shape.",{"type":51,"tag":86,"props":1104,"children":1105},{},[1106,1115,1117,1123,1125,1131,1132,1138],{"type":51,"tag":167,"props":1107,"children":1109},{"href":1108},"references\u002Fcamera-overlay-templates\u002F",[1110],{"type":51,"tag":73,"props":1111,"children":1113},{"className":1112},[],[1114],{"type":57,"value":1108},{"type":57,"value":1116},"\n— starter ",{"type":51,"tag":73,"props":1118,"children":1120},{"className":1119},[],[1121],{"type":57,"value":1122},".dts.tmpl",{"type":57,"value":1124}," templates: ",{"type":51,"tag":73,"props":1126,"children":1128},{"className":1127},[],[1129],{"type":57,"value":1130},"dphy-direct.dts.tmpl",{"type":57,"value":641},{"type":51,"tag":73,"props":1133,"children":1135},{"className":1134},[],[1136],{"type":57,"value":1137},"gmsl-serdes.dts.tmpl",{"type":57,"value":665},{"type":51,"tag":86,"props":1140,"children":1141},{},[1142,1151],{"type":51,"tag":167,"props":1143,"children":1145},{"href":1144},"..\u002F..\u002Fscripts\u002Fpin_verifier.py",[1146],{"type":51,"tag":73,"props":1147,"children":1149},{"className":1148},[],[1150],{"type":57,"value":1144},{"type":57,"value":1152},"\n— shared HSIO pin verifier (Step 6).",{"type":51,"tag":86,"props":1154,"children":1155},{},[1156,1165,1167,1173],{"type":51,"tag":167,"props":1157,"children":1159},{"href":1158},"..\u002F..\u002Freferences\u002Fplatform_template.yaml",[1160],{"type":51,"tag":73,"props":1161,"children":1163},{"className":1162},[],[1164],{"type":57,"value":1158},{"type":57,"value":1166},"\n— ",{"type":51,"tag":73,"props":1168,"children":1170},{"className":1169},[],[1171],{"type":57,"value":1172},"documents:",{"type":57,"value":1174}," block consumed by Step 1.",{"type":51,"tag":86,"props":1176,"children":1177},{},[1178,1188],{"type":51,"tag":167,"props":1179,"children":1181},{"href":1180},"..\u002F..\u002Fcontext\u002Fbsp-customization-workflow.md#workflow-invariants",[1182],{"type":51,"tag":73,"props":1183,"children":1185},{"className":1184},[],[1186],{"type":57,"value":1187},"..\u002F..\u002Fcontext\u002Fbsp-customization-workflow.md",{"type":57,"value":1189},"\n— overlay edit protocol.",{"type":51,"tag":86,"props":1191,"children":1192},{},[1193,1202],{"type":51,"tag":167,"props":1194,"children":1196},{"href":1195},"..\u002Fjetson-customize-pinmux\u002FSKILL.md",[1197],{"type":51,"tag":73,"props":1198,"children":1200},{"className":1199},[],[1201],{"type":57,"value":1195},{"type":57,"value":1203}," —\nsibling skill auto-invoked by Step 6 to fix HSIO pin SFIO\nmismatches (CAM I²C, MCLK, reset GPIOs).",{"type":51,"tag":86,"props":1205,"children":1206},{},[1207,1216,1218,1224],{"type":51,"tag":167,"props":1208,"children":1210},{"href":1209},"..\u002Fjetson-derive-carrier\u002FSKILL.md",[1211],{"type":51,"tag":73,"props":1212,"children":1214},{"className":1213},[],[1215],{"type":57,"value":1209},{"type":57,"value":1217},"\n— must run first; produces the carrier base overlay (the\n",{"type":51,"tag":73,"props":1219,"children":1221},{"className":1220},[],[1222],{"type":57,"value":1223},"*-dynamic.dtbo",{"type":57,"value":1225},") this skill's composite stacks after.",{"type":51,"tag":86,"props":1227,"children":1228},{},[1229,1238,1240,1246,1248,1254],{"type":51,"tag":167,"props":1230,"children":1232},{"href":1231},"..\u002Fjetson-init-source\u002FSKILL.md",[1233],{"type":51,"tag":73,"props":1234,"children":1236},{"className":1235},[],[1237],{"type":57,"value":1231},{"type":57,"value":1239}," —\nproduces the overlay tracker + ",{"type":51,"tag":73,"props":1241,"children":1243},{"className":1242},[],[1244],{"type":57,"value":1245},"bsp_sources",{"type":57,"value":1247}," repo (with the\n",{"type":51,"tag":73,"props":1249,"children":1251},{"className":1250},[],[1252],{"type":57,"value":1253},"hardware\u002Fnvidia\u002F\u003Cchip-dir>\u002F",{"type":57,"value":1255}," per-sensor DTSI tree) this skill\nreads and commits into.",{"items":1257,"total":1416},[1258,1276,1294,1305,1317,1331,1344,1358,1371,1382,1396,1405],{"slug":1259,"name":1259,"fn":1260,"description":1261,"org":1262,"tags":1263,"stars":1273,"repoUrl":1274,"updatedAt":1275},"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},[1264,1267,1270],{"name":1265,"slug":1266,"type":15},"Documentation","documentation",{"name":1268,"slug":1269,"type":15},"MCP","mcp",{"name":1271,"slug":1272,"type":15},"Search","search",21777,"https:\u002F\u002Fgithub.com\u002FNVIDIA\u002FNemoClaw","2026-07-20T06:00:01.461044",{"slug":1277,"name":1277,"fn":1278,"description":1279,"org":1280,"tags":1281,"stars":1291,"repoUrl":1292,"updatedAt":1293},"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},[1282,1285,1288],{"name":1283,"slug":1284,"type":15},"Containers","containers",{"name":1286,"slug":1287,"type":15},"Deployment","deployment",{"name":1289,"slug":1290,"type":15},"Python","python",17049,"https:\u002F\u002Fgithub.com\u002FNVIDIA\u002FMegatron-LM","2026-07-27T06:06:11.249662",{"slug":1295,"name":1295,"fn":1296,"description":1297,"org":1298,"tags":1299,"stars":1291,"repoUrl":1292,"updatedAt":1304},"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},[1300,1303],{"name":1301,"slug":1302,"type":15},"CI\u002FCD","ci-cd",{"name":1286,"slug":1287,"type":15},"2026-07-14T05:25:59.97109",{"slug":1306,"name":1306,"fn":1307,"description":1308,"org":1309,"tags":1310,"stars":1291,"repoUrl":1292,"updatedAt":1316},"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},[1311,1312,1313],{"name":1301,"slug":1302,"type":15},{"name":1286,"slug":1287,"type":15},{"name":1314,"slug":1315,"type":15},"GitHub","github","2026-07-27T06:06:12.278222",{"slug":1318,"name":1318,"fn":1319,"description":1320,"org":1321,"tags":1322,"stars":1291,"repoUrl":1292,"updatedAt":1330},"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},[1323,1326,1327],{"name":1324,"slug":1325,"type":15},"Debugging","debugging",{"name":1314,"slug":1315,"type":15},{"name":1328,"slug":1329,"type":15},"Triage","triage","2026-07-14T05:25:57.442089",{"slug":1332,"name":1332,"fn":1333,"description":1334,"org":1335,"tags":1336,"stars":1291,"repoUrl":1292,"updatedAt":1343},"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},[1337,1340],{"name":1338,"slug":1339,"type":15},"Best Practices","best-practices",{"name":1341,"slug":1342,"type":15},"Code Analysis","code-analysis","2026-07-14T05:25:56.18433",{"slug":1345,"name":1345,"fn":1346,"description":1347,"org":1348,"tags":1349,"stars":1291,"repoUrl":1292,"updatedAt":1357},"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},[1350,1353,1356],{"name":1351,"slug":1352,"type":15},"Machine Learning","machine-learning",{"name":1354,"slug":1355,"type":15},"Migration","migration",{"name":9,"slug":8,"type":15},"2026-07-17T06:07:11.777011",{"slug":1359,"name":1359,"fn":1360,"description":1361,"org":1362,"tags":1363,"stars":1291,"repoUrl":1292,"updatedAt":1370},"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},[1364,1367],{"name":1365,"slug":1366,"type":15},"QA","qa",{"name":1368,"slug":1369,"type":15},"Testing","testing","2026-07-14T05:25:53.673039",{"slug":1372,"name":1372,"fn":1373,"description":1374,"org":1375,"tags":1376,"stars":1291,"repoUrl":1292,"updatedAt":1381},"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},[1377,1378],{"name":1286,"slug":1287,"type":15},{"name":1379,"slug":1380,"type":15},"Infrastructure","infrastructure","2026-07-14T05:25:49.362534",{"slug":1383,"name":1383,"fn":1384,"description":1385,"org":1386,"tags":1387,"stars":1291,"repoUrl":1292,"updatedAt":1395},"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},[1388,1391,1392],{"name":1389,"slug":1390,"type":15},"Code Review","code-review",{"name":1314,"slug":1315,"type":15},{"name":1393,"slug":1394,"type":15},"Pull Requests","pull-requests","2026-07-14T05:26:01.226578",{"slug":1397,"name":1397,"fn":1398,"description":1399,"org":1400,"tags":1401,"stars":1291,"repoUrl":1292,"updatedAt":1404},"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},[1402,1403],{"name":1365,"slug":1366,"type":15},{"name":1368,"slug":1369,"type":15},"2026-07-14T05:25:54.928983",{"slug":1406,"name":1406,"fn":1407,"description":1408,"org":1409,"tags":1410,"stars":1291,"repoUrl":1292,"updatedAt":1415},"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},[1411,1414],{"name":1412,"slug":1413,"type":15},"Automation","automation",{"name":1301,"slug":1302,"type":15},"2026-07-30T05:29:03.275638",496,{"items":1418,"total":1514},[1419,1436,1446,1460,1470,1485,1500],{"slug":1420,"name":1420,"fn":1421,"description":1422,"org":1423,"tags":1424,"stars":23,"repoUrl":24,"updatedAt":1435},"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},[1425,1428,1431,1432],{"name":1426,"slug":1427,"type":15},"Data Analysis","data-analysis",{"name":1429,"slug":1430,"type":15},"Data Engineering","data-engineering",{"name":9,"slug":8,"type":15},{"name":1433,"slug":1434,"type":15},"Performance","performance","2026-07-14T05:28:43.176466",{"slug":1437,"name":1437,"fn":1438,"description":1439,"org":1440,"tags":1441,"stars":23,"repoUrl":24,"updatedAt":1445},"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},[1442,1443,1444],{"name":1286,"slug":1287,"type":15},{"name":1379,"slug":1380,"type":15},{"name":9,"slug":8,"type":15},"2026-07-14T05:29:06.667109",{"slug":1447,"name":1447,"fn":1448,"description":1449,"org":1450,"tags":1451,"stars":23,"repoUrl":24,"updatedAt":1459},"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},[1452,1455,1456],{"name":1453,"slug":1454,"type":15},"Agents","agents",{"name":9,"slug":8,"type":15},{"name":1457,"slug":1458,"type":15},"Research","research","2026-07-14T05:28:06.816956",{"slug":1461,"name":1461,"fn":1462,"description":1463,"org":1464,"tags":1465,"stars":23,"repoUrl":24,"updatedAt":1469},"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},[1466,1467,1468],{"name":1426,"slug":1427,"type":15},{"name":9,"slug":8,"type":15},{"name":1368,"slug":1369,"type":15},"2026-07-17T05:29:03.913266",{"slug":1471,"name":1471,"fn":1472,"description":1473,"org":1474,"tags":1475,"stars":23,"repoUrl":24,"updatedAt":1484},"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},[1476,1477,1480,1481],{"name":1412,"slug":1413,"type":15},{"name":1478,"slug":1479,"type":15},"Imaging","imaging",{"name":9,"slug":8,"type":15},{"name":1482,"slug":1483,"type":15},"Video","video","2026-07-17T05:28:53.905004",{"slug":1486,"name":1486,"fn":1487,"description":1488,"org":1489,"tags":1490,"stars":23,"repoUrl":24,"updatedAt":1499},"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},[1491,1492,1495,1496],{"name":1286,"slug":1287,"type":15},{"name":1493,"slug":1494,"type":15},"Docker","docker",{"name":9,"slug":8,"type":15},{"name":1497,"slug":1498,"type":15},"Operations","operations","2026-07-17T05:28:56.913999",{"slug":1501,"name":1501,"fn":1502,"description":1503,"org":1504,"tags":1505,"stars":23,"repoUrl":24,"updatedAt":1513},"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},[1506,1507,1510],{"name":9,"slug":8,"type":15},{"name":1508,"slug":1509,"type":15},"Quantum Computing","quantum-computing",{"name":1511,"slug":1512,"type":15},"Simulation","simulation","2026-07-14T05:26:58.898253",305]