Test your agent's hooks before paying for evals.
What it does
Creates and runs tests for a Claude Code harness using the cheapest tier that can answer the testing question.
When to use it
Use to check hook behavior, skill triggering, injected context, or the tools and side effects produced by a run.
How to use it
Give it a project harness to test; it selects a suitable tier, writes a test, runs it, and reports the result.
What you provide
- An existing project
Uses
Good to know
- cannot be undone: Claude
Latest review
What you need · 1
Enables per-token billing for real-model evaluations as an alternative to signing in with a Claude subscription; the skill recommends unsetting it to use subscription authentication.
Access · 2
This skill
Claude
Write
Uses Claude for paid evaluations (irreversible)
vigiles
Execute
Installs vigiles for harness testing