# Write Tests

> Write colocated tests for React + Express + TypeScript code using the project's runner, conventions, mocked repository interfaces, React Testing Library, and TanStack Query patterns.

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

---


# Write tests to convention

Test `$target`. Colocate tests.

## Step 1 — Detect runner and style

```bash
bash "${CLAUDE_PROJECT_DIR}/.claude/lib/detect-toolchain.sh"
```

Open nearby tests first. Match imports, setup, mocks, naming, and
`references/test-patterns.md`.

## Step 2 — Pick the test type

- Service tests use a mocked repository interface and no DB or HTTP.
- Repository tests are integration tests only when the project already has a test DB harness.
- Query-backed hooks and components use React Testing Library with a fresh QueryClient and `retry: false`.
- Presentational components render with props and assert by role, label, or text.
- Route tests use the project's existing supertest-style pattern when one exists.

## Step 3 — Write behavior tests

Use Arrange-Act-Assert. Test observable outcomes, including meaningful error or
edge cases, not private internals.

## Step 4 — Run new tests

```bash
source "${CLAUDE_PROJECT_DIR}/.claude/lib/detect-toolchain.sh"
run_tests <new-test-path>
```

Iterate until tests pass. Report covered behaviors, not a coverage percentage.

## Checklist

- [ ] Runner and local test style were detected.
- [ ] Test file is colocated with the code under test.
- [ ] Services use a mocked repository interface.
- [ ] Query-backed UI uses a fresh QueryClient with `retry: false`.
- [ ] Assertions target user-visible behavior or boundary calls.
- [ ] New tests pass through the detected runner.