jetson-print-device-info

por nvidia

Úsalo cuando necesites imprimir información del dispositivo Jetson (modelo del módulo, versión de L4T, kernel, versión del SO, modo de energía actual) desde un objetivo Jetson en ejecución. Esta es una…

npx skills add https://github.com/nvidia/skills --skill jetson-print-device-info

jetson-print-device-info

Prints a concise summary of the Jetson device this skill runs on.

This skill is intended as a reference example for the jetson-device-skills repo and the NVIDIA-wide skills CI. It runs on the Jetson target (not the host PC) and performs read-only inspection — useful as a baseline capture before running performance tests.

Purpose

Capture a baseline snapshot of a Jetson target's software stack and power mode before running performance, regression, or compatibility tests.

When to use

  • Starting a performance run and you want a captured baseline of the device's software stack and power mode.
  • Verifying that a freshly flashed Jetson matches an expected L4T / JetPack version.

Prerequisites

  • Running on a Jetson target (not the host PC).
  • Standard CLIs available: tr, cat, uname, uptime, lsb_release (or /etc/os-release).
  • nvpmodel available for power mode (optional — skill falls back gracefully).

Inputs

None. The skill reads only from the local Jetson filesystem and standard CLIs.

Instructions

Run each step in order and print the captured values into the report shown under Output format.

  1. Capture the module model and validate it is a Jetson target — exit early otherwise:
    # Device-tree strings are null-terminated, so strip NULs before printing.
    model=$(tr -d '\0' < /proc/device-tree/model 2>/dev/null)
    case "$model" in
      *Jetson*) ;;
      *) echo "Not running on a Jetson target (model: '${model:-unknown}')"; exit 1 ;;
    esac
    echo "$model"
    
  2. Extract the L4T release header line — skip the rest of the file (long list of library SHAs):
    head -1 /etc/nv_tegra_release 2>/dev/null || echo "L4T release info not found"
    # Equivalent: grep -m1 '^# R' /etc/nv_tegra_release
    
  3. Run nvpmodel -q and join its two output lines (NV Power Mode: <name> and the mode number) onto one line. paste -sd only uses the first char of its delimiter, so use awk to insert the literal / separator:
    nvpmodel -q 2>/dev/null | awk 'NR==1{a=$0; next} {print a" / "$0}' \
      || echo "nvpmodel not available"
    
  4. Print the kernel version and uptime:
    uname -r
    uptime -p
    
  5. Print the OS version (prefer lsb_release, fall back to /etc/os-release):
    lsb_release -ds 2>/dev/null || (. /etc/os-release && echo "$PRETTY_NAME") || echo "OS version not found"
    

Output format

Print a short report with these sections, one line each where possible:

Model:           <device-tree model string>
L4T release:     <release header line>
Power mode:      <nvpmodel name> / <mode number>
Kernel:          <uname -r>
Uptime:          <uptime -p>
OS version:      <lsb_release / /etc/os-release output>

Examples

Example output on an Orin AGX dev kit (L4T R36 / Ubuntu 22.04):

Model:           NVIDIA Jetson AGX Orin Developer Kit
L4T release:     # R36 (release), REVISION: 3.0
Power mode:      NV Power Mode: MAXN / 0
Kernel:          5.15.136-tegra
Uptime:          up 2 hours, 14 minutes
OS version:      Ubuntu 22.04.4 LTS

Example output on an AGX Thor (L4T R39 / Ubuntu 24.04):

Model:           NVIDIA Jetson AGX Thor Developer Kit
L4T release:     # R39 (release), REVISION: 0.0
Power mode:      NV Power Mode: 120W / 1
Kernel:          6.8.0-tegra
Uptime:          up 12 minutes
OS version:      Ubuntu 24.04 LTS

Error handling

Each command falls back to a clearly labeled "... not found" / "... not available" string if the underlying file or binary is missing — the skill never errors out mid-report. If /proc/device-tree/model is missing or does not contain a Jetson string, exit early with a clear "not running on a Jetson target" message.

Limitations

  • Read-only inspection only — does not detect GPU/CPU clocks, thermal state, or per-rail power.
  • nvpmodel output format varies between L4T versions; the skill prints it verbatim rather than parsing.

Troubleshooting

  • Error: /proc/device-tree/model: No such file or directory Cause: Running on a host PC, not a Jetson target. Solution: Run the skill on the Jetson device directly (e.g. via SSH).

  • Error: nvpmodel: command not found Cause: L4T BSP not installed, or running in a minimal container without nvpmodel. Solution: Expected on non-Jetson or stripped environments — the skill prints "nvpmodel not available" and continues.

Notes

  • Read-only. Do not change power mode, install packages, or modify any files.
  • If the skill is invoked on a host PC by mistake, /proc/device-tree/model will not contain a Jetson model string — detect that and exit with a clear message rather than printing misleading info.

Más skills de nvidia

compileiq-debug
nvidia
Úsalo cuando algo esté mal: Search() se cuelga, todas las evaluaciones devuelven INVALID_SCORE, las puntuaciones no mejoran, cada configuración devuelve el mismo número, errores de ptxas…
create-github-pr
nvidia
Crear solicitudes de extracción de GitHub usando la CLI gh. Usar cuando el usuario quiera crear un nuevo PR, enviar código para revisión o abrir una solicitud de extracción. Palabras clave de activación -…
nemoclaw-maintainer-cross-issue-sweep
nvidia
Escanea otros issues abiertos para encontrar aquellos que un PR dado también podría corregir o romper accidentalmente. Genera oportunidades de corrección adyacente y riesgos de contradicción con archivo:línea…
fhir-basics
nvidia
Enseña a los agentes cómo funcionan las APIs de FHIR R4, qué recursos están disponibles, cómo consultarlos con parámetros de búsqueda y cómo analizar correctamente todos los formatos de respuesta…
compileiq-validate-result
nvidia
Usar DESPUÉS de que una Búsqueda haya finalizado y ANTES de reclamar cualquier aceleración o enviar un ACF. Carga el CSV de dump_results, extrae los mejores K candidatos (de un solo objetivo)…
changelog-audit
nvidia
Auditar el CHANGELOG.md de Warp antes de un lanzamiento: recuperar entradas perdidas, ordenar por impacto en el usuario, refinar el lenguaje de las entradas, ajustar saltos de línea y (en modo rama de lanzamiento) incrementar comparación…
maintain-dynamic-plugins
nvidia
Mantener los cargadores de plugins dinámicos de NeMo Relay, manifiestos, SDKs nativos de Rust, protocolo de trabajador gRPC, SDK de trabajador Python, documentación, pruebas y cobertura del flujo de trabajo de lanzamiento
dgx-diagnose
nvidia
Diagnostica problemas comunes de la DGX Station GB300: fallos de CUDA, direccionamiento incorrecto de GPU, errores de contenedores vLLM/SGLang, problemas de estado MIG, errores de NVLink/Fabric Manager,…