
Test Harness
Three tiers for testing Claude Code harnesses, cheapest first: unit tests for hook logic (free, no model), deterministic tests for wiring and context injection (free, scripted mock), and evals for trigger precision and behavioral shifts (paid, real model). Most harness questions never need a model. You can verify a PreToolUse hook blocks the right commands, confirm injected context reaches the prompt, or measure whether a skill description fires too often, all with the right tier. The skill installs vigiles, picks the tier that can actually answer your question, writes the test, and runs it. Start cheap and climb only when the question genuinely requires a real model.
- skills.sh installs
- 2.1K
- GitHub stars
- 15
- last updated
- Aug 21, 2026
Install
$ npx -y skills add zernie/vigiles --skill test-harness --agent claude-codeInstalls into .claude/skills of the current project.
Files
Test the Claude Code harness — the hooks, skills, settings, and CLAUDE.md that steer an agent — as the assembled machine it ships as. vigiles gives three tiers, cheapest first; this skill picks the right one, writes the test, and runs it.
The guiding rule: start at the cheapest tier that can answer the question, and climb only when it genuinely can't. Two of the three tiers need no model and no API key, so they run on every commit for free — reach for the paid real-model tier only when the question actually requires a real model.
Step 0 — Pick the tier (the judgment call)
Match what you're testing to the cheapest tier that can answer it:
Sponsors
Details
- Repository
- github.com/zernie/vigiles
- Categories
- Frontend Development, Testing & QA, Documentation, AI & Agent Building
- First seen
- Jul 14, 2026