aipm-llm-judge
100 / 100 · Low risk
Fine to try.
Writes an LLM-as-a-judge prompt for one failure type and the check that proves it, a labelled agreement set, agreement with human labels on a held-out split, and bias checks for position, length and self-preference. Use for "run aipm-llm-judge", "LLM as a judge", "judge prompt", "automate our eval grading", "can we trust the LLM grader", "model-graded evals", "validate our judge against humans", part of the Claude for AI Product Managers Pack by 90n.
What this skill can do
We found nothing in the files we read that matches our rules. That does not prove the skill is safe: see the limits of static scanning.
Install
Paste this into your terminal. It downloads exactly the version we scanned, not whatever is newest.
if [ -e ~/.agents/skills/'aipm-llm-judge' ] || [ -L ~/.agents/skills/'aipm-llm-judge' ]; then echo 'A skill folder named aipm-llm-judge already exists — not replacing it' >&2; false; else d=$(mktemp -d) && git clone --quiet --filter=blob:none --no-checkout 'https://github.com/polar-bear-org/claude-skills' "$d" && git -C "$d" checkout --quiet '95f01806ca43f6ee788d0579743e35853b695930' -- 'packs/ai-product-managers-pack/skills/aipm-llm-judge' && mkdir -p ~/.agents/skills && cp -R "$d"/'packs/ai-product-managers-pack/skills/aipm-llm-judge' ~/.agents/skills/'aipm-llm-judge' && 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/'aipm-llm-judge' ~/.claude/skills/'aipm-llm-judge' (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
[](https://zalavijaysinh.github.io/skilltrust/repos/polar-bear-org/claude-skills/)
Score history
| Date | Version | Score |
|---|---|---|
| 2026-11-07 | 95f0180 | 100 |