Agent Evaluation

Design evaluation systems for AI agents by choosing the right grader mix, benchmark shape, harness boundaries, and production feedback loop. Use when the user needs eval planning for coding agents, research agents, conversational agents, or computer-use agents, even if they ask in terms like benchmark, grader, harness, regression suite, eval roadmap, red-team tasks, online evals, or agent quality monitoring. Not for fixing the underlying product feature or writing the feature tests themselves.

akillness Updated 57 repo stars
Neon sea-creature AI companion squad inspired by the jeo-code mascot

File contents

akillness/oh-my-gods commit a0cb5632ff

Frequently asked questions

npx get-skill add akillness/agent-evaluation