test-gap-audit

Purpose

Shared policy: Follow instructions/coordination.instructions.md for precedence, invocation, delegation, and handoffs. Apply instructions/workflow.instructions.md for proportional work and verification.

The gap report is conversational by default. Persist it only when the user, repository policy, or a parent workflow explicitly requests a durable artifact; otherwise return the risk-prioritized findings without writing files.

You are conducting a read-only audit of existing code for dangerous untested paths.

Your goals are to:

• Find gaps — identify code paths, branches, and error scenarios that lack test coverage.
• Assess risk — prioritize gaps by the damage they can cause when they fail silently.
• Evaluate quality — flag existing tests that pass but don't actually verify meaningful behavior.
• Report next steps — produce a risk-prioritized gap report and recommend test-gap-fill for approved test generation.

---

When to use this skill

Use this skill whenever:

• A bug shipped that "should have been caught by tests" — retroactively identify the gap.
• Before a major refactoring — ensure the safety net exists before you lean on it.
• During maintenance cycles — periodic test health check.
• The user asks to "find untested code", "improve test coverage", or "audit tests for [module]".
• After a production incident — identify what tests would have prevented it.
• You need the first step in the chain: test-gap-audit → approval → test-gap-fill.

Do not use this skill for:

• Planning tests for new features — use test-strategy instead.
• Running, debugging, generating, or editing tests — use test-gap-fill after approved gaps.
• Code review — use git-commit-review or code-reviewer agent.

Difference from test-strategy: The test-strategy skill is forward-looking (plans tests for new work). This skill is retroactive — it audits existing code that may have been in production for months without adequate coverage.