| name | write-tests |
| description | Author tests for a code change AND run mutation testing (Stryker / mutmut / go-mutesting) to verify the suite kills mutants. Each test is mutation-resistant by design — catches at least one plausible mutation (boolean, conditional, arithmetic, boundary, off-by-one, return-value). Auto-detects the repo's runner; skips when no test setup exists. Use after fixes land or for any 'cover these files' request. Triggers: write tests, add tests, test coverage for, unit tests for, test this, mutation testing, stryker, mutmut, kill mutants, 'tests for this change', spec these files. |
| argument-hint | <featureId> |
You are a test author. Write green coverage AND verify the suite kills mutants.
Read context
!`browzer get-step WRITE_TESTS --id $ARGUMENTS 2>/dev/null || echo "(no prior WRITE_TESTS step — first run)"`
$ARGUMENTS is the feature id passed by the orchestrator (e.g. feat-20260507-preamble-staging-migration); it is also the directory name under docs/browzer/.
The blob lists every changed file from completed TASK_NN steps and the relevant test runner the orchestrator probed.
Process
Dispatch the test author with subagent_type: browzer:tester, model: sonnet, effort: high.
- Pre-flight. Detect the repo's runner (vitest, jest, pytest, go test, ...). If no test infra exists, set
skipped: true with rationale and stop.
- Author tests scoped to the changed files. Cover happy path + edge cases + boundary conditions.
- Mutation testing. Run Stryker (JS/TS), mutmut (Python), or go-mutesting (Go). Mutation-resistant principles in
references/mutation-principles.md. Six mutation categories MUST be killed: boolean, conditional, arithmetic, boundary, off-by-one, return-value.
- Iterate on surviving mutants until each is killed or formally documented as unreachable.
Produce
Write docs/browzer/<feat>/staging/WRITE_TESTS.json.
Shape reference: see template.md (auto-generated from the workflow CUE schema). Do not paste schema-claiming JSON into this body.
When skipped: true, the only required field is rationale.
Persistence
The autosave hook persists staging/WRITE_TESTS.json automatically on write. Recommended flags when manually invoking save-step:
--quiet --async — WRITE_TESTS is not load-bearing for the next phase; fire-and-forget after the mutation score is known.
On validation failure, re-run with --hint-fixes for worked examples of valid values.
Done when
- File exists at
docs/browzer/<feat>/staging/WRITE_TESTS.json.
- Every survived mutant carries an explicit
rationale.
- The autosave hook validates and persists.
Return one line: write-tests: <N> tests added; mutation score=<score>.
