Guidance for creating, running, fixing, and promoting behavioral evaluations. Use when verifying agent decision logic, debugging failures, debugging prompt steering, or adding workspace regression tests.
Install
npx skills add google-gemini/gemini-cli --skill behavioral-evalsnpx skills add google-gemini/gemini-cli --skill behavioral-evals -a claude-code -gnpx skills add google-gemini/gemini-cli --skill behavioral-evals -a codex -ggemini skills install https://github.com/google-gemini/gemini-cli.git --path .gemini/skills/behavioral-evalsgh skill install google-gemini/gemini-cli behavioral-evalsgit clone --depth 1 https://github.com/google-gemini/gemini-cli /tmp/google-gemini__gemini-cli
mkdir -p ~/.claude/skills ~/.agents/skills
cp -R /tmp/google-gemini__gemini-cli/.gemini/skills/behavioral-evals ~/.claude/skills/behavioral-evals # Claude Code
cp -R /tmp/google-gemini__gemini-cli/.gemini/skills/behavioral-evals ~/.agents/skills/behavioral-evals # Codex, Cursor, Gemini CLI, Copilot, OpenCode