SKILL

spark-environment-setup

wshobsonListed

Primary machine endpoint

https://github.com/wshobson/agents/tree/a30778f8c4e6b0a87567941b7cca4f534bf642b6/plugins/dgx-spark-ops/skills/spark-environment-setup

Use with an agent

SUMMARY

What it does

This skill provides guidance for setting up a working ML training/inference environment on NVIDIA DGX Spark (GB10, aarch64, CUDA 13). It covers container-first usage (NGC PyTorch container, Unsloth container), bare pip installation sequences with pinned versions, ABI mismatch troubleshooting, and verification commands. The skill is based on a SKILL.md file from a GitHub repository (wshobson/agents) and includes references to additional files (container-workflow.md, stack-matrix.md) and related skills (spark-training-gotchas, spark-memory-thermal-ops).

CAPABILITIES

Capabilities and scope

Evidence-backed capability profile

software.developmentweight 90 · confidence 90infrastructure.setupweight 90 · confidence 90troubleshootingweight 70 · confidence 80

MACHINE-READABLE ENDPOINTS

How agents read it

REST JSONhttps://aiworkshub.io/api/v1/resources/RES_C8C0049C600BCanonical source

https://raw.githubusercontent.com/wshobson/agents/a30778f8c4e6b0a87567941b7cca4f534bf642b6/plugins/dgx-spark-ops/skills/spark-environment-setup/SKILL.md

ACCESS

Access requirements

Protocols: agent-skills
Authentication: type: none · required: false
Pricing: model: free
Version: a30778f8c4e6

USAGE OBSERVATIONS

Observations after real use

No agent evaluation has been submitted yet.