Create and run agent evals to verify AI output quality. Use when the user asks to test agent outputs, create eval cases, or benchmark skill performance.
Instalación
Instalar con Codex o Claude Copia este prompt, pégalo en Codex, Claude u otro asistente, y deja que revise la página de la skill y la instale por ti.
Create and run agent evals to verify AI output quality. Use when the user asks to test agent outputs, create eval cases, or benchmark skill performance.
Define, run, and report on evaluation cases for agent skills to confirm output quality and correctness.
When To Use
Use this skill when the user asks to:
Create eval cases for a skill
Test whether an agent produces correct output
Benchmark skill performance
Validate agent routing behavior
Inputs
Skill name
Test prompts (natural language)
Expected outputs (structured description)
Relevant files for context
Workflow
Identify the skill or agent being evaluated
Write 2–5 eval cases covering core use cases and edge cases
Define expected output for each case (structured, not exact match)
Save as JSON in evals/
Run eval prompts and compare outputs to expected
Report pass/fail with notes
Output Format
{"skill_name":"[name]","evals":[{"id":1,"name":"[eval-name]","prompt":"[test prompt]","expected_output":"[description of correct output]","files":[]}]}
