# Test

> Add, update, or repair tests for behavior, regressions, and test-first development. Use when the user asks for tests, TDD, red-green-refactor, regression coverage, test repair, or behavior-driven implementation.

- Skill: `hunvreus/test` (Agent Skill, multi-file: 2 files)
- Install (CLI): `npx skillmds@latest add hunvreus/test`
- Raw SKILL.md: https://api.skillmd.com/api/skills/hunvreus/test/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- Author: hunvreus (https://skillmd.com/u/hunvreus)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/hunvreus/test

---


# Test

## Input

- A behavior, bug, module, feature, failing test, or uncovered path to test.
- Use explicit input first; otherwise infer from context, recent edits, selected files, or branch.
- Safest default: identify public behavior and highest-risk paths before writing tests.

## Workflow

1. **Define behavior**. Identify the public interface and the behavior that matters to the user.
2. **Pick the first slice**. Choose one narrow behavior that proves the path end to end.
3. **Write or repair tests**. Prefer behavior-facing tests over implementation-detail tests.
4. **Use TDD when requested**. Red: write one failing test. Green: write the smallest implementation that passes. Repeat by slice.
5. **Refactor when green**. Simplify duplication, names, and module shape after tests pass.
6. **Validate**. Run relevant tests, typecheck, lint, or build commands.

## Output

- Behaviors covered
- Tests added, changed, or repaired
- Implementation changed, if any
- Validation run
- Remaining test gaps

## Guardrails

- Do not write all tests before all implementation when using TDD.
- Prefer tests through public interfaces over private functions and internal mocks.
- Do not refactor while tests are red.
- Keep test scope proportional to behavior risk.