EVALLM for AI agents

SKILL.md instructs an AI agent to check a manuscript, protocol or system description against all 84 EVALLM statements and return a structured report.

Read https://evallm.kather.ai/agent/SKILL.md and evaluate the attached manuscript against EVALLM.

SKILL.md
Evaluation instructions and all 84 consensus statements.
statements.json
The 84 statements with domain, applicability, agreement and mapping metadata.
checklist.json
The 28 checklist questions, statements and agreement values in four sections.

Install

mkdir -p .claude/skills/evallm
curl -fsSL https://evallm.kather.ai/agent/SKILL.md -o .claude/skills/evallm/SKILL.md