ml-evaluation

80 / 100 · Low risk

Fine to try.

By ain3sh/.agents (GitHub) · productivity · 14 stars · scanned at commit 4ddf141

ML model evaluation and experiment tracking -- LLM benchmarking with lm-evaluation-harness (MMLU, GSM8K, HumanEval, 60+ tasks), experiment tracking and hyperparameter sweeps with Weights & Biases. Use when benchmarking models, comparing training runs, tracking experiments, or reporting evaluation results.

Works with any agent that supports SKILL.md skills.

What this skill can do

Downloads and installs other software

The author does not mention this.

Found in 6 places.

skills/ml-evaluation/SKILL.md, line 27| [lm-eval-harness](./references/lm-eval-harness.md) | Benchmark LLMs on 60+ academic tasks | `pip install lm-eval` |
skills/ml-evaluation/SKILL.md, line 28| [Weights & Biases](./references/weights-and-biases.md) | Experiment tracking, sweeps, model registry | `pip install wandb` |
skills/ml-evaluation/references/lm-eval-harness.md, line 10pip install lm-eval

Show 3 more

skills/ml-evaluation/references/lm-eval-harness.md, line 336pip install vllm
skills/ml-evaluation/references/lm-eval-harness.md, line 439pip install human-eval
skills/ml-evaluation/references/weights-and-biases.md, line 21pip install wandb

Runs commands on your computer

The author does not mention this.

Found in 3 places.

skills/ml-evaluation/SKILL.md, line 33lm_eval --model hf --model_args pretrained=meta-llama/Llama-2-7b-hf --tasks mmlu,gsm8k,hellaswag --batch_size auto
skills/ml-evaluation/references/lm-eval-harness.md, line 189os.system(f"./eval_checkpoint.sh checkpoints step-{step}")
skills/ml-evaluation/references/lm-eval-harness.md, line 206os.system(f"lm_eval --model hf --model_args pretrained={checkpoint_path} ...")

Install

Paste this into your terminal. It downloads exactly the version we scanned, not whatever is newest.

if [ -e ~/.agents/skills/'ml-evaluation' ] || [ -L ~/.agents/skills/'ml-evaluation' ]; then echo 'A skill folder named ml-evaluation already exists — not replacing it' >&2; false; else d=$(mktemp -d) && git clone --quiet --filter=blob:none --no-checkout 'https://github.com/ain3sh/.agents' "$d" && git -C "$d" checkout --quiet '4ddf141db8b6d6a225ad100c3d9589932b7163fd' -- 'skills/ml-evaluation' && mkdir -p ~/.agents/skills && cp -R "$d"/'skills/ml-evaluation' ~/.agents/skills/'ml-evaluation' && rm -rf "$d"; fi

Codex reads ~/.agents/skills directly. For Claude Code, Cursor, Copilot, Gemini CLI, Antigravity or OpenCode, also link it. For Claude Code: ln -s ~/.agents/skills/'ml-evaluation' ~/.claude/skills/'ml-evaluation' (other agents: ~/.cursor/skills, ~/.copilot/skills, ~/.gemini/skills, ~/.gemini/config/skills, ~/.config/opencode/skills). Or install through the finder skill, which does both.

Badge for your README

The badge for ain3sh/.agents shows the score of its lowest-scoring skill.

SkillTrust badge for ain3sh/.agents

[![SkillTrust](https://zalavijaysinh.github.io/skilltrust/badges/ain3sh/.agents.svg)](https://zalavijaysinh.github.io/skilltrust/repos/ain3sh/.agents/)

Score history

Date	Version	Score
2026-11-11	`4ddf141`	80