dgx-spark-ops

v1.0.0

NVIDIA DGX Spark (GB10 Grace Blackwell) environment operations: aarch64/CUDA-13 stack setup, training gotcha preflights, and unified-memory/thermal management for local ML workloads

CodexClaude Code3 skills

by Seth HobsonMIT38.6kupdated 4 days ago

Source

$ git clone https://github.com/wshobson/agents

Clone the source, then follow the repository's marketplace instructions for your runtime. The plugin root is plugins/dgx-spark-ops/ inside the repository.

Layout

plugins/dgx-spark-ops/

├── .codex-plugin/plugin.json
├── .claude-plugin/plugin.json
├── skills/spark-environment-setup/SKILL.md
├── skills/spark-memory-thermal-ops/SKILL.md
└── skills/spark-training-gotchas/SKILL.md

Skills3

spark-environment-setupskills/spark-environment-setup/SKILL.md

Set up a working ML training/inference environment on NVIDIA DGX Spark (GB10, aarch64, CUDA 13). Use when installing PyTorch/Unsloth/TRL/vLLM on DGX Spark, hitting libcudart or wheel-ABI errors on aarch64, or choosing between NGC containers and bare pip installs.

spark-memory-thermal-opsskills/spark-memory-thermal-ops/SKILL.md

Manage unified memory and thermals during long-running ML jobs on NVIDIA DGX Spark. Use when planning memory headroom for a training run on GB10, when a job OOMs on unified memory, or when monitoring temperature and power during multi-hour training.

spark-training-gotchasskills/spark-training-gotchas/SKILL.md

Preflight and diagnose the ten known failure modes for ML training on NVIDIA DGX Spark. Use when a training run on DGX Spark fails to start, OOMs below the 128GB limit, slows down mid-run, or before any multi-hour training job on GB10.

Manifests2

plugins/dgx-spark-ops/.codex-plugin/plugin.json

{
  "name": "dgx-spark-ops",
  "version": "1.0.0",
  "description": "NVIDIA DGX Spark (GB10 Grace Blackwell) environment operations: aarch64/CUDA-13 stack setup, training gotcha preflights, and unified-memory/thermal management for local ML workloads",
  "skills": "./skills/",
  "author": {
    "name": "Seth Hobson",
    "email": "seth@major7apps.com"
  },
  "license": "MIT",
  "interface": {
    "displayName": "Dgx Spark Ops",
    "shortDescription": "NVIDIA DGX Spark (GB10 Grace Blackwell) environment operations: aarch64/CUDA-13 stack setup, training gotcha…",
    "category": "Coding"
  }
}

plugins/dgx-spark-ops/.claude-plugin/plugin.json

{
  "name": "dgx-spark-ops",
  "version": "1.0.0",
  "description": "NVIDIA DGX Spark (GB10 Grace Blackwell) environment operations: aarch64/CUDA-13 stack setup, training gotcha preflights, and unified-memory/thermal management for local ML workloads",
  "author": {
    "name": "Seth Hobson",
    "email": "seth@major7apps.com"
  },
  "license": "MIT"
}