← All skills

evals

85 / 100 · Low risk

Fine to try.

By grasp-technologies/grasp-agents (GitHub) · writing · 13 stars · scanned at commit e921188

How to use grasp_agents.evals and the grasp-evals CLI to measure an async function, an LLMAgent or any processor against a dataset. Covers scorers, metrics, runs, resume and rescore, comparing runs and regression gates, pairwise A/B, LLM judges with label-based validation, probes and validation gates, the labeling loop and Phoenix sync. Above all it covers the non-obvious semantics behind the numbers: how errors enter metrics, run immutability, sealed splits, example identity and pairing, and o…

Works with any agent that supports SKILL.md skills.

What this skill can do

Install

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

if [ -e ~/.agents/skills/'evals' ] || [ -L ~/.agents/skills/'evals' ]; then echo 'A skill folder named evals already exists — not replacing it' >&2; false; else d=$(mktemp -d) && git clone --quiet --filter=blob:none --no-checkout 'https://github.com/grasp-technologies/grasp-agents' "$d" && git -C "$d" checkout --quiet 'e92118807db5c49e5b5e0f4c346d9dafd8579d78' -- '.claude/skills/evals' && mkdir -p ~/.agents/skills && cp -R "$d"/'.claude/skills/evals' ~/.agents/skills/'evals' && 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/'evals' ~/.claude/skills/'evals' (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 grasp-technologies/grasp-agents shows the score of its lowest-scoring skill.

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

Score history