Field-tested methodology and concrete recipes for training and operating large-scale LLM/VLM/multi-modal models end to end - choosing and benchmarking accelerators, storage and network; SLURM/Kubernetes orchestration; maximizing training throughput and fitting models in memory; diagnosing and surviv
evaluate-cluster
An agent skill for Claude Code, Cursor, Codex and other AI agents, from stas00/ml-engineering. Scored by fixed code checks, no AI.
CCode 67/100Builder 0/10
Evaluates an ML GPU cluster for a cloud trial or acceptance test: environment dump, isolated newest PyTorch, matmul FLOPS (MAMF/MSMF) on every GPU while the others compute, intra-node all-reduce bandwidth and per-call latency of every collective, the same inter-node on every node you were given (omit those sections if there is only one node), fio on local disk and shared FS, dated markdown report. Use when the user asks to evaluate a cluster, kick the tires on trial nodes, run cluster acceptance, or measure GPU/network/storage. Canonical copy: https://github.com/stas00/ml-engineering/blob/master/skills/evaluate-cluster/SKILL.md
Install
npx skills add stas00/ml-engineering --skill evaluate-cluster
Where do I paste this?
Open a terminal in your project folder, paste the command and press Enter. It needs Node.js on your computer.
Run it in your project. The skills CLI asks which agents to install it for (Claude Code, Cursor, Codex and more); add -g to install it for your user instead.
No terminal? Copy SKILL.md and give it to your AI as instructions.
SKILL.md
A skill is instructions and scripts your agent will follow, so read it before you install it. Its text loads here from GitHub; you can also read it on GitHub.
More skills from stas00/ml-engineering
Code 55/100
No description.
Code 25/100