Find and fix failures in your agent's answers.
What it does
Evaluates an agent and guides fixes using scored test cases, failure analysis, and regression comparisons.
When to use it
Use when asked to evaluate an agent, write evaluation datasets, analyze failures, compare results, or improve agent quality.
How to use it
Give it an agent project to evaluate; it produces scored evaluation reports and iterates on agent fixes.
What you provide
- An existing project
Uses
GP
Latest review
What you need · 1
Requires agents-cli; the documented installation command is uv tool install google-agents-cli.
Access · 1
This skill
Gemini platform
Write
Sends traces for cloud grading