[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill-nvidia-jetson-headless-mode":3,"mdc-2q3vg8-key":34,"related-repo-nvidia-jetson-headless-mode":1250,"related-org-nvidia-jetson-headless-mode":1351},{"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-headless-mode","configure Jetson headless mode","Plan and apply safe Jetson headless-mode changes to reclaim GUI and daemon memory.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":9},"nvidia","NVIDIA","https:\u002F\u002Fpexgzepcugksgbtrxkhf.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Forg-logos\u002Fnvidia.png",[12,16,19,22],{"name":13,"slug":14,"type":15},"Performance","performance","tag",{"name":17,"slug":18,"type":15},"Memory","memory",{"name":20,"slug":21,"type":15},"Deployment","deployment",{"name":9,"slug":8,"type":15},2473,"https:\u002F\u002Fgithub.com\u002FNVIDIA\u002Fskills","2026-07-14T05:26:27.670412","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-headless-mode","---\nname: jetson-headless-mode\ndescription: Plan and apply safe Jetson headless-mode changes to reclaim GUI and daemon memory.\nversion: 0.0.1\nlicense: \"Apache-2.0\"\nmetadata:\n  author: \"Jetson Team\"\n  tags: [jetson, headless, memory]\n  languages: [bash]\n  data-classification: public\n---\n\n# Jetson Headless Mode\n\nPlan-then-apply for safe, reversible user-space memory reclamation: switch the default systemd target away from `graphical.target` and disable a curated set of non-essential daemons. This is the highest-yield, lowest-risk memory win on Jetson.\n\n## Purpose\n\nBuild a user-approved headless-mode plan from live audit data, then apply only safe, reversible user-space changes that reduce desktop and daemon memory use on Jetson.\n\n## When to use\n\n- \"Free as much memory as possible — I don't need the GUI.\"\n- \"I'm shipping this Jetson as an inference appliance \u002F edge node.\"\n- After `jetson-memory-audit` shows `default_systemd_target=graphical.target` or shows `gdm3` \u002F `lightdm` \u002F `sddm` active on a system the user describes as headless.\n\n## When NOT to use\n\n- The user needs the local desktop, display output, kiosk UI, or any X\u002FWayland session. In that case, do not recommend disabling the graphical target or display manager; use `jetson-memory-audit` for a read-only view and suggest non-GUI memory options instead.\n- You do not have current audit data. Run `jetson-memory-audit` first, or ask the user for its output, before proposing changes or estimating savings.\n\nUse live device data as the source of truth. Jetson family, SKU\u002Fvariant, memory totals, active display services, and savings estimates must come from `jetson-diagnostic\u002Fscripts\u002Fdetect_jetson.sh`, `audit.json`, or a fresh `jetson-memory-audit` run. If a value is not available, say it is unknown instead of guessing. The savings numbers below are upper bounds; the real delta is whatever a before\u002Fafter audit reports.\n\n## Prerequisites\n\n- Start from a current `jetson-memory-audit` JSON snapshot.\n- Confirm the user does not need the local desktop, display output, kiosk UI, or X\u002FWayland session.\n- Mutating changes require `sudo` and explicit user approval; dry-run first unless approval was already given in the same prompt.\n- Run on the Jetson host or in a host-visible sandbox with access to systemd state.\n\n## Available Scripts\n\n| Script | Purpose | Arguments |\n|--------|---------|-----------|\n| `scripts\u002Fplan.sh` | Reads a memory audit JSON and emits a plan containing safe, reversible recommendations. | `--audit PATH` or `--audit -`, plus `--human`. |\n| `scripts\u002Fapply.sh` | Prints or applies the safe commands from a plan JSON. Dry-run by default. | `--plan PATH` or `--plan -`, `--apply`, `--reboot`, `--drop-caches`. |\n\nIf your agent runtime supports `run_script`, use it to run `scripts\u002Fplan.sh` and `scripts\u002Fapply.sh` and summarize the returned output. Otherwise run the scripts with `bash` from the repository root.\n\n## Instructions\n\n1. Run `scripts\u002Fplan.sh` to read `audit.json` (from `jetson-memory-audit`) and emit a plan with only `safety: safe` knobs (target switch, display managers, audio, print, modem, etc.).\n2. Show the plan to the user and confirm.\n3. Run `scripts\u002Fapply.sh --plan plan.json` for a **dry run**. Re-run with `--apply` to execute. Add `--drop-caches` to flush the page cache afterward, or `--reboot` to take effect immediately.\n4. Re-run `jetson-memory-audit\u002Fscripts\u002Faudit.sh` to verify the actual delta.\n\n## Expected workflow\n\nUse the scripts for estimates and application so recommendations are based on the current device state rather than the static upper-bound table alone.\n\n- For \"what would headless save\", \"estimate\", \"plan\", or production planning prompts, run `scripts\u002Fplan.sh --audit \u003Caudit.json>` and report `estimated_total_savings_mb`, the top `recommendations[*].knob`, and whether any display manager or `graphical.target` is active. Do not run `apply.sh`.\n- For prompts where the user explicitly says to apply headless mode now, run `scripts\u002Fapply.sh --plan \u003Cplan.json>` once as a dry run first. If the user has already approved mutation in the same prompt, re-run the same command with `--apply` and mention the reversible command(s).\n- If direct execution fails in an agent runtime, invoke scripts with `bash {baseDir}\u002Fscripts\u002F\u003Cscript-name> ...`. Do not try to `chmod` installed skill files.\n\n## Plan \u002F apply contract\n\n- `plan.sh` emits the same JSON shape as `jetson-inference-mem-tune\u002Fscripts\u002Frecommend.py`: an array of `recommendations` with `{layer, knob, estimated_savings_mb, safety, command, reversible_command, rationale}`.\n- `apply.sh` filters entries to `safety == \"safe\"` with a non-empty `command`, then re-checks the filtered safety marker in the shell loop before execution. Anything else, such as kernel command-line changes, device-tree changes, or accuracy tradeoffs, is out of scope for this skill.\n- Default mode is **dry-run**. `--apply` is required to mutate the system.\n\n## Knobs covered\n\n| Knob                            | Action                                       | Estimated savings | Reversible? |\n|---------------------------------|----------------------------------------------|-------------------|-------------|\n| `disable-graphical-target`      | `systemctl set-default multi-user.target`    | up to 865 MB      | yes         |\n| `stop-gdm3` \u002F `gdm` \u002F `lightdm` \u002F `sddm` \u002F `display-manager` | `systemctl disable --now \u003Csvc>` | ~200 MB \u002F svc | yes |\n| `stop-pulseaudio`               | disable audio daemon                         | ~8 MB             | yes         |\n| `stop-bluetooth`                | disable Bluetooth stack                      | ~6 MB             | yes         |\n| `stop-ModemManager`             | disable WWAN manager                         | ~4 MB             | yes         |\n| `stop-cups` \u002F `stop-cups-browsed` | disable print stack                        | ~5 \u002F ~3 MB        | yes         |\n| `stop-snapd`                    | disable Snap daemon                          | ~30 MB            | yes         |\n| `stop-whoopsie` \u002F `kerneloops`  | disable crash reporters                      | ~4 \u002F ~2 MB        | yes         |\n| `stop-avahi-daemon`             | disable mDNS                                 | ~3 MB             | yes         |\n| `stop-unattended-upgrades` \u002F `packagekit` | disable background package work    | ~6 \u002F ~8 MB        | yes         |\n\n## Do NOT disable these services\n\n- `nvargus-daemon` — required for any libargus camera pipeline.\n- `nvgetty.service` — serial console; disabling can lock you out of recovery.\n- `nvpmodel` — power-mode service; required for clock\u002Fpower tuning.\n- `containerd` \u002F `docker` — leave on if you run containers (most inference workloads do).\n- `nvfb` \u002F `nvdisplay`-related kernel services — tied to boot-time display configuration, so this skill does not change them.\n\n## Safety\n\n- Does not edit `\u002Fboot\u002Fextlinux\u002Fextlinux.conf`, the device tree, or boot-time memory reservations.\n- Does not disable services it does not have an explicit entry for (no blanket \"disable everything not whitelisted\").\n- Every applied change has a documented `reversible_command`. Re-running the plan with the reverts is sufficient to restore.\n- Dry-run by default. `--apply` is the only way to mutate.\n- Report only device facts and savings figures that came from live detection or audit output.\n\n## Cross-platform behavior\n\nThe same set of knobs applies to every Jetson family in the matrix above. The script reads `JETSON_GENERATION` \u002F `JETSON_PRODUCT_LINE` \u002F `JETSON_VARIANT` from `jetson-diagnostic\u002Fscripts\u002Fdetect_jetson.sh` (and still exports legacy `JETSON_SKU`) so the agent can attribute the savings correctly in its summary, but it does not branch on product line.\n",{"data":35,"body":45},{"name":4,"description":6,"version":36,"license":26,"metadata":37},"0.0.1",{"author":38,"tags":39,"languages":42,"data-classification":44},"Jetson Team",[40,41,18],"jetson","headless",[43],"bash","public",{"type":46,"children":47},"root",[48,56,71,78,83,89,148,154,181,209,215,253,259,390,424,430,527,533,538,625,631,713,719,1067,1073,1145,1151,1202,1208],{"type":49,"tag":50,"props":51,"children":52},"element","h1",{"id":4},[53],{"type":54,"value":55},"text","Jetson Headless Mode",{"type":49,"tag":57,"props":58,"children":59},"p",{},[60,62,69],{"type":54,"value":61},"Plan-then-apply for safe, reversible user-space memory reclamation: switch the default systemd target away from ",{"type":49,"tag":63,"props":64,"children":66},"code",{"className":65},[],[67],{"type":54,"value":68},"graphical.target",{"type":54,"value":70}," and disable a curated set of non-essential daemons. This is the highest-yield, lowest-risk memory win on Jetson.",{"type":49,"tag":72,"props":73,"children":75},"h2",{"id":74},"purpose",[76],{"type":54,"value":77},"Purpose",{"type":49,"tag":57,"props":79,"children":80},{},[81],{"type":54,"value":82},"Build a user-approved headless-mode plan from live audit data, then apply only safe, reversible user-space changes that reduce desktop and daemon memory use on Jetson.",{"type":49,"tag":72,"props":84,"children":86},{"id":85},"when-to-use",[87],{"type":54,"value":88},"When to use",{"type":49,"tag":90,"props":91,"children":92},"ul",{},[93,99,104],{"type":49,"tag":94,"props":95,"children":96},"li",{},[97],{"type":54,"value":98},"\"Free as much memory as possible — I don't need the GUI.\"",{"type":49,"tag":94,"props":100,"children":101},{},[102],{"type":54,"value":103},"\"I'm shipping this Jetson as an inference appliance \u002F edge node.\"",{"type":49,"tag":94,"props":105,"children":106},{},[107,109,115,117,123,125,131,133,139,140,146],{"type":54,"value":108},"After ",{"type":49,"tag":63,"props":110,"children":112},{"className":111},[],[113],{"type":54,"value":114},"jetson-memory-audit",{"type":54,"value":116}," shows ",{"type":49,"tag":63,"props":118,"children":120},{"className":119},[],[121],{"type":54,"value":122},"default_systemd_target=graphical.target",{"type":54,"value":124}," or shows ",{"type":49,"tag":63,"props":126,"children":128},{"className":127},[],[129],{"type":54,"value":130},"gdm3",{"type":54,"value":132}," \u002F ",{"type":49,"tag":63,"props":134,"children":136},{"className":135},[],[137],{"type":54,"value":138},"lightdm",{"type":54,"value":132},{"type":49,"tag":63,"props":141,"children":143},{"className":142},[],[144],{"type":54,"value":145},"sddm",{"type":54,"value":147}," active on a system the user describes as headless.",{"type":49,"tag":72,"props":149,"children":151},{"id":150},"when-not-to-use",[152],{"type":54,"value":153},"When NOT to use",{"type":49,"tag":90,"props":155,"children":156},{},[157,169],{"type":49,"tag":94,"props":158,"children":159},{},[160,162,167],{"type":54,"value":161},"The user needs the local desktop, display output, kiosk UI, or any X\u002FWayland session. In that case, do not recommend disabling the graphical target or display manager; use ",{"type":49,"tag":63,"props":163,"children":165},{"className":164},[],[166],{"type":54,"value":114},{"type":54,"value":168}," for a read-only view and suggest non-GUI memory options instead.",{"type":49,"tag":94,"props":170,"children":171},{},[172,174,179],{"type":54,"value":173},"You do not have current audit data. Run ",{"type":49,"tag":63,"props":175,"children":177},{"className":176},[],[178],{"type":54,"value":114},{"type":54,"value":180}," first, or ask the user for its output, before proposing changes or estimating savings.",{"type":49,"tag":57,"props":182,"children":183},{},[184,186,192,194,200,202,207],{"type":54,"value":185},"Use live device data as the source of truth. Jetson family, SKU\u002Fvariant, memory totals, active display services, and savings estimates must come from ",{"type":49,"tag":63,"props":187,"children":189},{"className":188},[],[190],{"type":54,"value":191},"jetson-diagnostic\u002Fscripts\u002Fdetect_jetson.sh",{"type":54,"value":193},", ",{"type":49,"tag":63,"props":195,"children":197},{"className":196},[],[198],{"type":54,"value":199},"audit.json",{"type":54,"value":201},", or a fresh ",{"type":49,"tag":63,"props":203,"children":205},{"className":204},[],[206],{"type":54,"value":114},{"type":54,"value":208}," run. If a value is not available, say it is unknown instead of guessing. The savings numbers below are upper bounds; the real delta is whatever a before\u002Fafter audit reports.",{"type":49,"tag":72,"props":210,"children":212},{"id":211},"prerequisites",[213],{"type":54,"value":214},"Prerequisites",{"type":49,"tag":90,"props":216,"children":217},{},[218,230,235,248],{"type":49,"tag":94,"props":219,"children":220},{},[221,223,228],{"type":54,"value":222},"Start from a current ",{"type":49,"tag":63,"props":224,"children":226},{"className":225},[],[227],{"type":54,"value":114},{"type":54,"value":229}," JSON snapshot.",{"type":49,"tag":94,"props":231,"children":232},{},[233],{"type":54,"value":234},"Confirm the user does not need the local desktop, display output, kiosk UI, or X\u002FWayland session.",{"type":49,"tag":94,"props":236,"children":237},{},[238,240,246],{"type":54,"value":239},"Mutating changes require ",{"type":49,"tag":63,"props":241,"children":243},{"className":242},[],[244],{"type":54,"value":245},"sudo",{"type":54,"value":247}," and explicit user approval; dry-run first unless approval was already given in the same prompt.",{"type":49,"tag":94,"props":249,"children":250},{},[251],{"type":54,"value":252},"Run on the Jetson host or in a host-visible sandbox with access to systemd state.",{"type":49,"tag":72,"props":254,"children":256},{"id":255},"available-scripts",[257],{"type":54,"value":258},"Available Scripts",{"type":49,"tag":260,"props":261,"children":262},"table",{},[263,286],{"type":49,"tag":264,"props":265,"children":266},"thead",{},[267],{"type":49,"tag":268,"props":269,"children":270},"tr",{},[271,277,281],{"type":49,"tag":272,"props":273,"children":274},"th",{},[275],{"type":54,"value":276},"Script",{"type":49,"tag":272,"props":278,"children":279},{},[280],{"type":54,"value":77},{"type":49,"tag":272,"props":282,"children":283},{},[284],{"type":54,"value":285},"Arguments",{"type":49,"tag":287,"props":288,"children":289},"tbody",{},[290,335],{"type":49,"tag":268,"props":291,"children":292},{},[293,303,308],{"type":49,"tag":294,"props":295,"children":296},"td",{},[297],{"type":49,"tag":63,"props":298,"children":300},{"className":299},[],[301],{"type":54,"value":302},"scripts\u002Fplan.sh",{"type":49,"tag":294,"props":304,"children":305},{},[306],{"type":54,"value":307},"Reads a memory audit JSON and emits a plan containing safe, reversible recommendations.",{"type":49,"tag":294,"props":309,"children":310},{},[311,317,319,325,327,333],{"type":49,"tag":63,"props":312,"children":314},{"className":313},[],[315],{"type":54,"value":316},"--audit PATH",{"type":54,"value":318}," or ",{"type":49,"tag":63,"props":320,"children":322},{"className":321},[],[323],{"type":54,"value":324},"--audit -",{"type":54,"value":326},", plus ",{"type":49,"tag":63,"props":328,"children":330},{"className":329},[],[331],{"type":54,"value":332},"--human",{"type":54,"value":334},".",{"type":49,"tag":268,"props":336,"children":337},{},[338,347,352],{"type":49,"tag":294,"props":339,"children":340},{},[341],{"type":49,"tag":63,"props":342,"children":344},{"className":343},[],[345],{"type":54,"value":346},"scripts\u002Fapply.sh",{"type":49,"tag":294,"props":348,"children":349},{},[350],{"type":54,"value":351},"Prints or applies the safe commands from a plan JSON. Dry-run by default.",{"type":49,"tag":294,"props":353,"children":354},{},[355,361,362,368,369,375,376,382,383,389],{"type":49,"tag":63,"props":356,"children":358},{"className":357},[],[359],{"type":54,"value":360},"--plan PATH",{"type":54,"value":318},{"type":49,"tag":63,"props":363,"children":365},{"className":364},[],[366],{"type":54,"value":367},"--plan -",{"type":54,"value":193},{"type":49,"tag":63,"props":370,"children":372},{"className":371},[],[373],{"type":54,"value":374},"--apply",{"type":54,"value":193},{"type":49,"tag":63,"props":377,"children":379},{"className":378},[],[380],{"type":54,"value":381},"--reboot",{"type":54,"value":193},{"type":49,"tag":63,"props":384,"children":386},{"className":385},[],[387],{"type":54,"value":388},"--drop-caches",{"type":54,"value":334},{"type":49,"tag":57,"props":391,"children":392},{},[393,395,401,403,408,410,415,417,422],{"type":54,"value":394},"If your agent runtime supports ",{"type":49,"tag":63,"props":396,"children":398},{"className":397},[],[399],{"type":54,"value":400},"run_script",{"type":54,"value":402},", use it to run ",{"type":49,"tag":63,"props":404,"children":406},{"className":405},[],[407],{"type":54,"value":302},{"type":54,"value":409}," and ",{"type":49,"tag":63,"props":411,"children":413},{"className":412},[],[414],{"type":54,"value":346},{"type":54,"value":416}," and summarize the returned output. Otherwise run the scripts with ",{"type":49,"tag":63,"props":418,"children":420},{"className":419},[],[421],{"type":54,"value":43},{"type":54,"value":423}," from the repository root.",{"type":49,"tag":72,"props":425,"children":427},{"id":426},"instructions",[428],{"type":54,"value":429},"Instructions",{"type":49,"tag":431,"props":432,"children":433},"ol",{},[434,468,473,514],{"type":49,"tag":94,"props":435,"children":436},{},[437,439,444,446,451,453,458,460,466],{"type":54,"value":438},"Run ",{"type":49,"tag":63,"props":440,"children":442},{"className":441},[],[443],{"type":54,"value":302},{"type":54,"value":445}," to read ",{"type":49,"tag":63,"props":447,"children":449},{"className":448},[],[450],{"type":54,"value":199},{"type":54,"value":452}," (from ",{"type":49,"tag":63,"props":454,"children":456},{"className":455},[],[457],{"type":54,"value":114},{"type":54,"value":459},") and emit a plan with only ",{"type":49,"tag":63,"props":461,"children":463},{"className":462},[],[464],{"type":54,"value":465},"safety: safe",{"type":54,"value":467}," knobs (target switch, display managers, audio, print, modem, etc.).",{"type":49,"tag":94,"props":469,"children":470},{},[471],{"type":54,"value":472},"Show the plan to the user and confirm.",{"type":49,"tag":94,"props":474,"children":475},{},[476,477,483,485,491,493,498,500,505,507,512],{"type":54,"value":438},{"type":49,"tag":63,"props":478,"children":480},{"className":479},[],[481],{"type":54,"value":482},"scripts\u002Fapply.sh --plan plan.json",{"type":54,"value":484}," for a ",{"type":49,"tag":486,"props":487,"children":488},"strong",{},[489],{"type":54,"value":490},"dry run",{"type":54,"value":492},". Re-run with ",{"type":49,"tag":63,"props":494,"children":496},{"className":495},[],[497],{"type":54,"value":374},{"type":54,"value":499}," to execute. Add ",{"type":49,"tag":63,"props":501,"children":503},{"className":502},[],[504],{"type":54,"value":388},{"type":54,"value":506}," to flush the page cache afterward, or ",{"type":49,"tag":63,"props":508,"children":510},{"className":509},[],[511],{"type":54,"value":381},{"type":54,"value":513}," to take effect immediately.",{"type":49,"tag":94,"props":515,"children":516},{},[517,519,525],{"type":54,"value":518},"Re-run ",{"type":49,"tag":63,"props":520,"children":522},{"className":521},[],[523],{"type":54,"value":524},"jetson-memory-audit\u002Fscripts\u002Faudit.sh",{"type":54,"value":526}," to verify the actual delta.",{"type":49,"tag":72,"props":528,"children":530},{"id":529},"expected-workflow",[531],{"type":54,"value":532},"Expected workflow",{"type":49,"tag":57,"props":534,"children":535},{},[536],{"type":54,"value":537},"Use the scripts for estimates and application so recommendations are based on the current device state rather than the static upper-bound table alone.",{"type":49,"tag":90,"props":539,"children":540},{},[541,584,604],{"type":49,"tag":94,"props":542,"children":543},{},[544,546,552,554,560,562,568,570,575,577,583],{"type":54,"value":545},"For \"what would headless save\", \"estimate\", \"plan\", or production planning prompts, run ",{"type":49,"tag":63,"props":547,"children":549},{"className":548},[],[550],{"type":54,"value":551},"scripts\u002Fplan.sh --audit \u003Caudit.json>",{"type":54,"value":553}," and report ",{"type":49,"tag":63,"props":555,"children":557},{"className":556},[],[558],{"type":54,"value":559},"estimated_total_savings_mb",{"type":54,"value":561},", the top ",{"type":49,"tag":63,"props":563,"children":565},{"className":564},[],[566],{"type":54,"value":567},"recommendations[*].knob",{"type":54,"value":569},", and whether any display manager or ",{"type":49,"tag":63,"props":571,"children":573},{"className":572},[],[574],{"type":54,"value":68},{"type":54,"value":576}," is active. Do not run ",{"type":49,"tag":63,"props":578,"children":580},{"className":579},[],[581],{"type":54,"value":582},"apply.sh",{"type":54,"value":334},{"type":49,"tag":94,"props":585,"children":586},{},[587,589,595,597,602],{"type":54,"value":588},"For prompts where the user explicitly says to apply headless mode now, run ",{"type":49,"tag":63,"props":590,"children":592},{"className":591},[],[593],{"type":54,"value":594},"scripts\u002Fapply.sh --plan \u003Cplan.json>",{"type":54,"value":596}," once as a dry run first. If the user has already approved mutation in the same prompt, re-run the same command with ",{"type":49,"tag":63,"props":598,"children":600},{"className":599},[],[601],{"type":54,"value":374},{"type":54,"value":603}," and mention the reversible command(s).",{"type":49,"tag":94,"props":605,"children":606},{},[607,609,615,617,623],{"type":54,"value":608},"If direct execution fails in an agent runtime, invoke scripts with ",{"type":49,"tag":63,"props":610,"children":612},{"className":611},[],[613],{"type":54,"value":614},"bash {baseDir}\u002Fscripts\u002F\u003Cscript-name> ...",{"type":54,"value":616},". Do not try to ",{"type":49,"tag":63,"props":618,"children":620},{"className":619},[],[621],{"type":54,"value":622},"chmod",{"type":54,"value":624}," installed skill files.",{"type":49,"tag":72,"props":626,"children":628},{"id":627},"plan-apply-contract",[629],{"type":54,"value":630},"Plan \u002F apply contract",{"type":49,"tag":90,"props":632,"children":633},{},[634,668,694],{"type":49,"tag":94,"props":635,"children":636},{},[637,643,645,651,653,659,661,667],{"type":49,"tag":63,"props":638,"children":640},{"className":639},[],[641],{"type":54,"value":642},"plan.sh",{"type":54,"value":644}," emits the same JSON shape as ",{"type":49,"tag":63,"props":646,"children":648},{"className":647},[],[649],{"type":54,"value":650},"jetson-inference-mem-tune\u002Fscripts\u002Frecommend.py",{"type":54,"value":652},": an array of ",{"type":49,"tag":63,"props":654,"children":656},{"className":655},[],[657],{"type":54,"value":658},"recommendations",{"type":54,"value":660}," with ",{"type":49,"tag":63,"props":662,"children":664},{"className":663},[],[665],{"type":54,"value":666},"{layer, knob, estimated_savings_mb, safety, command, reversible_command, rationale}",{"type":54,"value":334},{"type":49,"tag":94,"props":669,"children":670},{},[671,676,678,684,686,692],{"type":49,"tag":63,"props":672,"children":674},{"className":673},[],[675],{"type":54,"value":582},{"type":54,"value":677}," filters entries to ",{"type":49,"tag":63,"props":679,"children":681},{"className":680},[],[682],{"type":54,"value":683},"safety == \"safe\"",{"type":54,"value":685}," with a non-empty ",{"type":49,"tag":63,"props":687,"children":689},{"className":688},[],[690],{"type":54,"value":691},"command",{"type":54,"value":693},", then re-checks the filtered safety marker in the shell loop before execution. Anything else, such as kernel command-line changes, device-tree changes, or accuracy tradeoffs, is out of scope for this skill.",{"type":49,"tag":94,"props":695,"children":696},{},[697,699,704,706,711],{"type":54,"value":698},"Default mode is ",{"type":49,"tag":486,"props":700,"children":701},{},[702],{"type":54,"value":703},"dry-run",{"type":54,"value":705},". ",{"type":49,"tag":63,"props":707,"children":709},{"className":708},[],[710],{"type":54,"value":374},{"type":54,"value":712}," is required to mutate the system.",{"type":49,"tag":72,"props":714,"children":716},{"id":715},"knobs-covered",[717],{"type":54,"value":718},"Knobs covered",{"type":49,"tag":260,"props":720,"children":721},{},[722,748],{"type":49,"tag":264,"props":723,"children":724},{},[725],{"type":49,"tag":268,"props":726,"children":727},{},[728,733,738,743],{"type":49,"tag":272,"props":729,"children":730},{},[731],{"type":54,"value":732},"Knob",{"type":49,"tag":272,"props":734,"children":735},{},[736],{"type":54,"value":737},"Action",{"type":49,"tag":272,"props":739,"children":740},{},[741],{"type":54,"value":742},"Estimated savings",{"type":49,"tag":272,"props":744,"children":745},{},[746],{"type":54,"value":747},"Reversible?",{"type":49,"tag":287,"props":749,"children":750},{},[751,782,838,864,890,916,949,975,1008,1034],{"type":49,"tag":268,"props":752,"children":753},{},[754,763,772,777],{"type":49,"tag":294,"props":755,"children":756},{},[757],{"type":49,"tag":63,"props":758,"children":760},{"className":759},[],[761],{"type":54,"value":762},"disable-graphical-target",{"type":49,"tag":294,"props":764,"children":765},{},[766],{"type":49,"tag":63,"props":767,"children":769},{"className":768},[],[770],{"type":54,"value":771},"systemctl set-default multi-user.target",{"type":49,"tag":294,"props":773,"children":774},{},[775],{"type":54,"value":776},"up to 865 MB",{"type":49,"tag":294,"props":778,"children":779},{},[780],{"type":54,"value":781},"yes",{"type":49,"tag":268,"props":783,"children":784},{},[785,820,829,834],{"type":49,"tag":294,"props":786,"children":787},{},[788,794,795,801,802,807,808,813,814],{"type":49,"tag":63,"props":789,"children":791},{"className":790},[],[792],{"type":54,"value":793},"stop-gdm3",{"type":54,"value":132},{"type":49,"tag":63,"props":796,"children":798},{"className":797},[],[799],{"type":54,"value":800},"gdm",{"type":54,"value":132},{"type":49,"tag":63,"props":803,"children":805},{"className":804},[],[806],{"type":54,"value":138},{"type":54,"value":132},{"type":49,"tag":63,"props":809,"children":811},{"className":810},[],[812],{"type":54,"value":145},{"type":54,"value":132},{"type":49,"tag":63,"props":815,"children":817},{"className":816},[],[818],{"type":54,"value":819},"display-manager",{"type":49,"tag":294,"props":821,"children":822},{},[823],{"type":49,"tag":63,"props":824,"children":826},{"className":825},[],[827],{"type":54,"value":828},"systemctl disable --now \u003Csvc>",{"type":49,"tag":294,"props":830,"children":831},{},[832],{"type":54,"value":833},"~200 MB \u002F svc",{"type":49,"tag":294,"props":835,"children":836},{},[837],{"type":54,"value":781},{"type":49,"tag":268,"props":839,"children":840},{},[841,850,855,860],{"type":49,"tag":294,"props":842,"children":843},{},[844],{"type":49,"tag":63,"props":845,"children":847},{"className":846},[],[848],{"type":54,"value":849},"stop-pulseaudio",{"type":49,"tag":294,"props":851,"children":852},{},[853],{"type":54,"value":854},"disable audio daemon",{"type":49,"tag":294,"props":856,"children":857},{},[858],{"type":54,"value":859},"~8 MB",{"type":49,"tag":294,"props":861,"children":862},{},[863],{"type":54,"value":781},{"type":49,"tag":268,"props":865,"children":866},{},[867,876,881,886],{"type":49,"tag":294,"props":868,"children":869},{},[870],{"type":49,"tag":63,"props":871,"children":873},{"className":872},[],[874],{"type":54,"value":875},"stop-bluetooth",{"type":49,"tag":294,"props":877,"children":878},{},[879],{"type":54,"value":880},"disable Bluetooth stack",{"type":49,"tag":294,"props":882,"children":883},{},[884],{"type":54,"value":885},"~6 MB",{"type":49,"tag":294,"props":887,"children":888},{},[889],{"type":54,"value":781},{"type":49,"tag":268,"props":891,"children":892},{},[893,902,907,912],{"type":49,"tag":294,"props":894,"children":895},{},[896],{"type":49,"tag":63,"props":897,"children":899},{"className":898},[],[900],{"type":54,"value":901},"stop-ModemManager",{"type":49,"tag":294,"props":903,"children":904},{},[905],{"type":54,"value":906},"disable WWAN manager",{"type":49,"tag":294,"props":908,"children":909},{},[910],{"type":54,"value":911},"~4 MB",{"type":49,"tag":294,"props":913,"children":914},{},[915],{"type":54,"value":781},{"type":49,"tag":268,"props":917,"children":918},{},[919,935,940,945],{"type":49,"tag":294,"props":920,"children":921},{},[922,928,929],{"type":49,"tag":63,"props":923,"children":925},{"className":924},[],[926],{"type":54,"value":927},"stop-cups",{"type":54,"value":132},{"type":49,"tag":63,"props":930,"children":932},{"className":931},[],[933],{"type":54,"value":934},"stop-cups-browsed",{"type":49,"tag":294,"props":936,"children":937},{},[938],{"type":54,"value":939},"disable print stack",{"type":49,"tag":294,"props":941,"children":942},{},[943],{"type":54,"value":944},"~5 \u002F ~3 MB",{"type":49,"tag":294,"props":946,"children":947},{},[948],{"type":54,"value":781},{"type":49,"tag":268,"props":950,"children":951},{},[952,961,966,971],{"type":49,"tag":294,"props":953,"children":954},{},[955],{"type":49,"tag":63,"props":956,"children":958},{"className":957},[],[959],{"type":54,"value":960},"stop-snapd",{"type":49,"tag":294,"props":962,"children":963},{},[964],{"type":54,"value":965},"disable Snap daemon",{"type":49,"tag":294,"props":967,"children":968},{},[969],{"type":54,"value":970},"~30 MB",{"type":49,"tag":294,"props":972,"children":973},{},[974],{"type":54,"value":781},{"type":49,"tag":268,"props":976,"children":977},{},[978,994,999,1004],{"type":49,"tag":294,"props":979,"children":980},{},[981,987,988],{"type":49,"tag":63,"props":982,"children":984},{"className":983},[],[985],{"type":54,"value":986},"stop-whoopsie",{"type":54,"value":132},{"type":49,"tag":63,"props":989,"children":991},{"className":990},[],[992],{"type":54,"value":993},"kerneloops",{"type":49,"tag":294,"props":995,"children":996},{},[997],{"type":54,"value":998},"disable crash reporters",{"type":49,"tag":294,"props":1000,"children":1001},{},[1002],{"type":54,"value":1003},"~4 \u002F ~2 MB",{"type":49,"tag":294,"props":1005,"children":1006},{},[1007],{"type":54,"value":781},{"type":49,"tag":268,"props":1009,"children":1010},{},[1011,1020,1025,1030],{"type":49,"tag":294,"props":1012,"children":1013},{},[1014],{"type":49,"tag":63,"props":1015,"children":1017},{"className":1016},[],[1018],{"type":54,"value":1019},"stop-avahi-daemon",{"type":49,"tag":294,"props":1021,"children":1022},{},[1023],{"type":54,"value":1024},"disable mDNS",{"type":49,"tag":294,"props":1026,"children":1027},{},[1028],{"type":54,"value":1029},"~3 MB",{"type":49,"tag":294,"props":1031,"children":1032},{},[1033],{"type":54,"value":781},{"type":49,"tag":268,"props":1035,"children":1036},{},[1037,1053,1058,1063],{"type":49,"tag":294,"props":1038,"children":1039},{},[1040,1046,1047],{"type":49,"tag":63,"props":1041,"children":1043},{"className":1042},[],[1044],{"type":54,"value":1045},"stop-unattended-upgrades",{"type":54,"value":132},{"type":49,"tag":63,"props":1048,"children":1050},{"className":1049},[],[1051],{"type":54,"value":1052},"packagekit",{"type":49,"tag":294,"props":1054,"children":1055},{},[1056],{"type":54,"value":1057},"disable background package work",{"type":49,"tag":294,"props":1059,"children":1060},{},[1061],{"type":54,"value":1062},"~6 \u002F ~8 MB",{"type":49,"tag":294,"props":1064,"children":1065},{},[1066],{"type":54,"value":781},{"type":49,"tag":72,"props":1068,"children":1070},{"id":1069},"do-not-disable-these-services",[1071],{"type":54,"value":1072},"Do NOT disable these services",{"type":49,"tag":90,"props":1074,"children":1075},{},[1076,1087,1098,1109,1127],{"type":49,"tag":94,"props":1077,"children":1078},{},[1079,1085],{"type":49,"tag":63,"props":1080,"children":1082},{"className":1081},[],[1083],{"type":54,"value":1084},"nvargus-daemon",{"type":54,"value":1086}," — required for any libargus camera pipeline.",{"type":49,"tag":94,"props":1088,"children":1089},{},[1090,1096],{"type":49,"tag":63,"props":1091,"children":1093},{"className":1092},[],[1094],{"type":54,"value":1095},"nvgetty.service",{"type":54,"value":1097}," — serial console; disabling can lock you out of recovery.",{"type":49,"tag":94,"props":1099,"children":1100},{},[1101,1107],{"type":49,"tag":63,"props":1102,"children":1104},{"className":1103},[],[1105],{"type":54,"value":1106},"nvpmodel",{"type":54,"value":1108}," — power-mode service; required for clock\u002Fpower tuning.",{"type":49,"tag":94,"props":1110,"children":1111},{},[1112,1118,1119,1125],{"type":49,"tag":63,"props":1113,"children":1115},{"className":1114},[],[1116],{"type":54,"value":1117},"containerd",{"type":54,"value":132},{"type":49,"tag":63,"props":1120,"children":1122},{"className":1121},[],[1123],{"type":54,"value":1124},"docker",{"type":54,"value":1126}," — leave on if you run containers (most inference workloads do).",{"type":49,"tag":94,"props":1128,"children":1129},{},[1130,1136,1137,1143],{"type":49,"tag":63,"props":1131,"children":1133},{"className":1132},[],[1134],{"type":54,"value":1135},"nvfb",{"type":54,"value":132},{"type":49,"tag":63,"props":1138,"children":1140},{"className":1139},[],[1141],{"type":54,"value":1142},"nvdisplay",{"type":54,"value":1144},"-related kernel services — tied to boot-time display configuration, so this skill does not change them.",{"type":49,"tag":72,"props":1146,"children":1148},{"id":1147},"safety",[1149],{"type":54,"value":1150},"Safety",{"type":49,"tag":90,"props":1152,"children":1153},{},[1154,1167,1172,1185,1197],{"type":49,"tag":94,"props":1155,"children":1156},{},[1157,1159,1165],{"type":54,"value":1158},"Does not edit ",{"type":49,"tag":63,"props":1160,"children":1162},{"className":1161},[],[1163],{"type":54,"value":1164},"\u002Fboot\u002Fextlinux\u002Fextlinux.conf",{"type":54,"value":1166},", the device tree, or boot-time memory reservations.",{"type":49,"tag":94,"props":1168,"children":1169},{},[1170],{"type":54,"value":1171},"Does not disable services it does not have an explicit entry for (no blanket \"disable everything not whitelisted\").",{"type":49,"tag":94,"props":1173,"children":1174},{},[1175,1177,1183],{"type":54,"value":1176},"Every applied change has a documented ",{"type":49,"tag":63,"props":1178,"children":1180},{"className":1179},[],[1181],{"type":54,"value":1182},"reversible_command",{"type":54,"value":1184},". Re-running the plan with the reverts is sufficient to restore.",{"type":49,"tag":94,"props":1186,"children":1187},{},[1188,1190,1195],{"type":54,"value":1189},"Dry-run by default. ",{"type":49,"tag":63,"props":1191,"children":1193},{"className":1192},[],[1194],{"type":54,"value":374},{"type":54,"value":1196}," is the only way to mutate.",{"type":49,"tag":94,"props":1198,"children":1199},{},[1200],{"type":54,"value":1201},"Report only device facts and savings figures that came from live detection or audit output.",{"type":49,"tag":72,"props":1203,"children":1205},{"id":1204},"cross-platform-behavior",[1206],{"type":54,"value":1207},"Cross-platform behavior",{"type":49,"tag":57,"props":1209,"children":1210},{},[1211,1213,1219,1220,1226,1227,1233,1235,1240,1242,1248],{"type":54,"value":1212},"The same set of knobs applies to every Jetson family in the matrix above. The script reads ",{"type":49,"tag":63,"props":1214,"children":1216},{"className":1215},[],[1217],{"type":54,"value":1218},"JETSON_GENERATION",{"type":54,"value":132},{"type":49,"tag":63,"props":1221,"children":1223},{"className":1222},[],[1224],{"type":54,"value":1225},"JETSON_PRODUCT_LINE",{"type":54,"value":132},{"type":49,"tag":63,"props":1228,"children":1230},{"className":1229},[],[1231],{"type":54,"value":1232},"JETSON_VARIANT",{"type":54,"value":1234}," from ",{"type":49,"tag":63,"props":1236,"children":1238},{"className":1237},[],[1239],{"type":54,"value":191},{"type":54,"value":1241}," (and still exports legacy ",{"type":49,"tag":63,"props":1243,"children":1245},{"className":1244},[],[1246],{"type":54,"value":1247},"JETSON_SKU",{"type":54,"value":1249},") so the agent can attribute the savings correctly in its summary, but it does not branch on product line.",{"items":1251,"total":1350},[1252,1267,1279,1293,1305,1322,1336],{"slug":1253,"name":1253,"fn":1254,"description":1255,"org":1256,"tags":1257,"stars":23,"repoUrl":24,"updatedAt":1266},"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},[1258,1261,1264,1265],{"name":1259,"slug":1260,"type":15},"Data Analysis","data-analysis",{"name":1262,"slug":1263,"type":15},"Data Engineering","data-engineering",{"name":9,"slug":8,"type":15},{"name":13,"slug":14,"type":15},"2026-07-14T05:28:43.176466",{"slug":1268,"name":1268,"fn":1269,"description":1270,"org":1271,"tags":1272,"stars":23,"repoUrl":24,"updatedAt":1278},"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},[1273,1274,1277],{"name":20,"slug":21,"type":15},{"name":1275,"slug":1276,"type":15},"Infrastructure","infrastructure",{"name":9,"slug":8,"type":15},"2026-07-14T05:29:06.667109",{"slug":1280,"name":1280,"fn":1281,"description":1282,"org":1283,"tags":1284,"stars":23,"repoUrl":24,"updatedAt":1292},"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},[1285,1288,1289],{"name":1286,"slug":1287,"type":15},"Agents","agents",{"name":9,"slug":8,"type":15},{"name":1290,"slug":1291,"type":15},"Research","research","2026-07-14T05:28:06.816956",{"slug":1294,"name":1294,"fn":1295,"description":1296,"org":1297,"tags":1298,"stars":23,"repoUrl":24,"updatedAt":1304},"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},[1299,1300,1301],{"name":1259,"slug":1260,"type":15},{"name":9,"slug":8,"type":15},{"name":1302,"slug":1303,"type":15},"Testing","testing","2026-07-17T05:29:03.913266",{"slug":1306,"name":1306,"fn":1307,"description":1308,"org":1309,"tags":1310,"stars":23,"repoUrl":24,"updatedAt":1321},"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},[1311,1314,1317,1318],{"name":1312,"slug":1313,"type":15},"Automation","automation",{"name":1315,"slug":1316,"type":15},"Imaging","imaging",{"name":9,"slug":8,"type":15},{"name":1319,"slug":1320,"type":15},"Video","video","2026-07-17T05:28:53.905004",{"slug":1323,"name":1323,"fn":1324,"description":1325,"org":1326,"tags":1327,"stars":23,"repoUrl":24,"updatedAt":1335},"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},[1328,1329,1331,1332],{"name":20,"slug":21,"type":15},{"name":1330,"slug":1124,"type":15},"Docker",{"name":9,"slug":8,"type":15},{"name":1333,"slug":1334,"type":15},"Operations","operations","2026-07-17T05:28:56.913999",{"slug":1337,"name":1337,"fn":1338,"description":1339,"org":1340,"tags":1341,"stars":23,"repoUrl":24,"updatedAt":1349},"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},[1342,1343,1346],{"name":9,"slug":8,"type":15},{"name":1344,"slug":1345,"type":15},"Quantum Computing","quantum-computing",{"name":1347,"slug":1348,"type":15},"Simulation","simulation","2026-07-14T05:26:58.898253",305,{"items":1352,"total":1503},[1353,1371,1387,1398,1410,1424,1437,1451,1462,1471,1485,1494],{"slug":1354,"name":1354,"fn":1355,"description":1356,"org":1357,"tags":1358,"stars":1368,"repoUrl":1369,"updatedAt":1370},"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},[1359,1362,1365],{"name":1360,"slug":1361,"type":15},"Documentation","documentation",{"name":1363,"slug":1364,"type":15},"MCP","mcp",{"name":1366,"slug":1367,"type":15},"Search","search",21777,"https:\u002F\u002Fgithub.com\u002FNVIDIA\u002FNemoClaw","2026-07-20T06:00:01.461044",{"slug":1372,"name":1372,"fn":1373,"description":1374,"org":1375,"tags":1376,"stars":1384,"repoUrl":1385,"updatedAt":1386},"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},[1377,1380,1381],{"name":1378,"slug":1379,"type":15},"Containers","containers",{"name":20,"slug":21,"type":15},{"name":1382,"slug":1383,"type":15},"Python","python",17049,"https:\u002F\u002Fgithub.com\u002FNVIDIA\u002FMegatron-LM","2026-07-27T06:06:11.249662",{"slug":1388,"name":1388,"fn":1389,"description":1390,"org":1391,"tags":1392,"stars":1384,"repoUrl":1385,"updatedAt":1397},"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},[1393,1396],{"name":1394,"slug":1395,"type":15},"CI\u002FCD","ci-cd",{"name":20,"slug":21,"type":15},"2026-07-14T05:25:59.97109",{"slug":1399,"name":1399,"fn":1400,"description":1401,"org":1402,"tags":1403,"stars":1384,"repoUrl":1385,"updatedAt":1409},"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},[1404,1405,1406],{"name":1394,"slug":1395,"type":15},{"name":20,"slug":21,"type":15},{"name":1407,"slug":1408,"type":15},"GitHub","github","2026-07-27T06:06:12.278222",{"slug":1411,"name":1411,"fn":1412,"description":1413,"org":1414,"tags":1415,"stars":1384,"repoUrl":1385,"updatedAt":1423},"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},[1416,1419,1420],{"name":1417,"slug":1418,"type":15},"Debugging","debugging",{"name":1407,"slug":1408,"type":15},{"name":1421,"slug":1422,"type":15},"Triage","triage","2026-07-14T05:25:57.442089",{"slug":1425,"name":1425,"fn":1426,"description":1427,"org":1428,"tags":1429,"stars":1384,"repoUrl":1385,"updatedAt":1436},"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},[1430,1433],{"name":1431,"slug":1432,"type":15},"Best Practices","best-practices",{"name":1434,"slug":1435,"type":15},"Code Analysis","code-analysis","2026-07-14T05:25:56.18433",{"slug":1438,"name":1438,"fn":1439,"description":1440,"org":1441,"tags":1442,"stars":1384,"repoUrl":1385,"updatedAt":1450},"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},[1443,1446,1449],{"name":1444,"slug":1445,"type":15},"Machine Learning","machine-learning",{"name":1447,"slug":1448,"type":15},"Migration","migration",{"name":9,"slug":8,"type":15},"2026-07-17T06:07:11.777011",{"slug":1452,"name":1452,"fn":1453,"description":1454,"org":1455,"tags":1456,"stars":1384,"repoUrl":1385,"updatedAt":1461},"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},[1457,1460],{"name":1458,"slug":1459,"type":15},"QA","qa",{"name":1302,"slug":1303,"type":15},"2026-07-14T05:25:53.673039",{"slug":1463,"name":1463,"fn":1464,"description":1465,"org":1466,"tags":1467,"stars":1384,"repoUrl":1385,"updatedAt":1470},"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},[1468,1469],{"name":20,"slug":21,"type":15},{"name":1275,"slug":1276,"type":15},"2026-07-14T05:25:49.362534",{"slug":1472,"name":1472,"fn":1473,"description":1474,"org":1475,"tags":1476,"stars":1384,"repoUrl":1385,"updatedAt":1484},"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},[1477,1480,1481],{"name":1478,"slug":1479,"type":15},"Code Review","code-review",{"name":1407,"slug":1408,"type":15},{"name":1482,"slug":1483,"type":15},"Pull Requests","pull-requests","2026-07-14T05:26:01.226578",{"slug":1486,"name":1486,"fn":1487,"description":1488,"org":1489,"tags":1490,"stars":1384,"repoUrl":1385,"updatedAt":1493},"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},[1491,1492],{"name":1458,"slug":1459,"type":15},{"name":1302,"slug":1303,"type":15},"2026-07-14T05:25:54.928983",{"slug":1495,"name":1495,"fn":1496,"description":1497,"org":1498,"tags":1499,"stars":1384,"repoUrl":1385,"updatedAt":1502},"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},[1500,1501],{"name":1312,"slug":1313,"type":15},{"name":1394,"slug":1395,"type":15},"2026-07-30T05:29:03.275638",496]