test-gap-detector

Test Gap Detector

Activate this skill before implementing a feature or fix. Get back a ranked list of coverage gaps and the minimum viable tests to close them.

Gap Categories

Category	What It Means
missing test file	Source file changed but no *.test.* counterpart exists
untested error path	A catch, else, or error branch has no test
untested branch	An if/else or switch arm has no test exercising it
no integration test	A service-to-service or API call has no integration test
no regression test	A previously-failed path has no regression test guarding it

Workflow

1. List all files to be changed
2. For each file, check if a test file exists (*.test.ts, *.spec.ts, __tests__/)
3. For files with tests, scan for untested branches:
   • Count if, else, catch, switch statements
   • Cross-reference with test file to see which paths are exercised
4. Check ~/.fd-plan/<slug>/.codebase/FAILURES.json for prior failures on these paths — flag as regression gap if no regression test exists
5. For external calls (fetch, db.query, sendEmail), check for integration test coverage
6. Rank gaps by risk (auth > payment > data > logic > UI)
7. Produce minimum viable test set (top 3–5 tests)

Output Format

## Test Gap Report

### Gap Summary
| File | Gap Type | Risk | Suggested Test |
|------|----------|------|----------------|

### Minimum Viable Test Set (top 5)
1. **[test name]** — [file.test.ts]
   Tests: [what it validates]
   ```typescript
   it('[test name]', async () => {
     // test skeleton
   })

Coverage Verdict: GOOD | GAPS FOUND | CRITICAL GAPS

## Guidance

- A test skeleton is always better than no test — write it even if incomplete
- Auth and payment paths with no regression test = CRITICAL GAP, block merge
- Do not add tests just to hit a coverage number — add tests for real risk paths