Agent Evals

Measure an agent instead of vibing it: fixed task sets, seeded runs, and a rubric written before the results are seen.

gulmezeren2-byte Updated 1 repo stars
skilldoctor checking a 60-skill example home: the listing budget bar at 10,096 of 8,000 characters (126%), then findings — budget exceeded, a description past t

File contents

agent-evals

Measure an agent instead of vibing it: fixed task sets, seeded runs, and a rubric written before the results are seen.

gulmezeren2-byte/claude-skills-doctor commit a7ad4c58ee

Frequently asked questions

npx get-skill add gulmezeren2-byte/agent-evals