unit-test-engineer
Turns changed behavior into executable checks.
- Status
- Observed
- Last reviewed
- 2026-10-10
ID: unit-test-engineer.
What it does
Reads changes, the installed runner, scripts and nearby tests. For each observable behavior, identifies an existing test, a test to add or why a unit test is not the appropriate check. Writes the smallest test that would fail on regression, runs the targeted command and broadens verification when the change affects shared logic or multiple consumers.
When to use it
Use it when a rule, interactive state or public interface changes, when a test fails or when updated evidence is needed. A visual correction may require the browser rather than a test that merely repeats the implementation.
How to use it
Ask for it by name, or invoke it with /unit-test-engineer in Claude Code or $unit-test-engineer in Codex. Specify behavior, changes and the expected result. Request tests through public interfaces with independent expected values. Request commands actually run, outcomes and coverage gaps, reusing the repository structure.
What it delivers
- Added or updated tests that distinguish correct behavior from regressions
- Commands run and current outcomes
- Covered behaviors and declared gaps
Example
Add the minimum test proving that the filter excludes archived items and retains active ones. Run affected tests and identify any unverified consumers.
Executable instructions
In the workflow
Use the router to choose this capability based on the required outcome.