By name: test-fix description: Diagnose and minimally fix a failing test, then re-run. allowed-tools: [read, edit, bash]
Test fix
When the user reports a failing test (or a test suite failure), work in this order. Resist the temptation to rewrite large chunks of code.
1. Reproduce
Run the failing test in isolation first. For Go: go test -run TestThing ./pkg/.... For Node: npm test -- --run TestThing. The
goal is a fast feedback loop, not the whole suite.
2. Read the failure carefully
- What was expected vs got?
- Which line of test code triggered the failure?
- Is it a logic error in the implementation, an outdated test expectation, or environmental drift (e.g. timezone, locale, file system case-sensitivity)?
3. Choose the smallest possible fix
- If the test is wrong → update the assertion only. Don't restructure.
- If the implementation is wrong → patch the smallest function that fixes the failing case. Avoid drive-by refactors.
- If the test depends on behaviour you need to change deliberately → ask the user before changing the assertion.
4. Re-run
The same isolated test, then the immediate suite, then the full test command if applicable. Stop the moment you see green; don't keep poking.
5. Report
Write 2–3 sentences: what was wrong, what changed, and which command verifies the fix. Include the diff inline.