tdd

<skill id="tdd"> <purpose> Tests first, code second. No exceptions. The test defines the contract; the code fulfills it. Red (failing test) -> Green (minimal code to pass) -> Refactor (clean up while green). If you write code before tests, you're testing your bugs, not your requirements. </purpose> <prerequisite> AgentDB read-start has run. Check for existing test patterns in codebase. Identify test framework in use (Jest, Vitest, pytest, etc.). </prerequisite> <reference> Skill-specific: skills/tdd/reference/tdd-research.md </reference>

<core_principles>

1. TESTS DEFINE BEHAVIOR: Write tests from requirements, not implementation. Test what SHOULD happen.
2. MINIMAL CODE: Write only enough code to make the test pass. No anticipatory features.
3. REFACTOR WHILE GREEN: Clean up only when tests pass. Never refactor red.
4. 80% COVERAGE MINIMUM: Unit + integration + E2E. All edge cases covered.
5. ONE ASSERT PER TEST: Test names are specifications. Split if "and" appears. </core_principles>

<workflow> 1. Write user journey: "As a [role], I want [action], so that [benefit]" 2. Generate test cases from journey (happy path, edge cases, errors) 3. Run tests - they MUST fail (red) 4. Implement minimal code to pass 5. Run tests - they MUST pass (green) 6. Refactor while keeping tests green 7. Verify coverage >= 80% </workflow>

<mocking_patterns>

<!-- Supabase -->

jest.mock('@/lib/supabase', () => ({
  supabase: {
    from: jest.fn(() => ({
      select: jest.fn(() => ({
        eq: jest.fn(() => Promise.resolve({
          data: [{ id: 1, name: 'Test' }],
          error: null
        }))
      }))
    }))
  }
}))

<!-- Redis -->

jest.mock('@/lib/redis', () => ({
  searchByVector: jest.fn(() => Promise.resolve([
    { slug: 'test', similarity_score: 0.95 }
  ])),
  checkHealth: jest.fn(() => Promise.resolve({ connected: true }))
}))

<!-- OpenAI -->

jest.mock('@/lib/openai', () => ({
  generateEmbedding: jest.fn(() => Promise.resolve(
    new Array(1536).fill(0.1)
  ))
}))

</mocking_patterns>

<test_organization>

src/
  components/
    Button/
      Button.tsx
      Button.test.tsx      # Unit tests colocated
  app/
    api/
      users/
        route.ts
        route.test.ts      # Integration tests colocated
e2e/
  auth.spec.ts             # E2E tests separate
  checkout.spec.ts

</test_organization>

<anti_patterns> <block id="code_before_test">Writing code then tests validates bugs. Test first or not at all.</block> <block id="testing_implementation">Test behavior, not structure. Refactoring should not break tests.</block> <block id="multiple_asserts">One behavior per test. If you need "and", split the test.</block> <block id="brittle_selectors">Use semantic selectors ([data-testid], role), not CSS classes.</block> <block id="test_dependency">Each test must be independent. No shared state between tests.</block> </anti_patterns>

<coverage_config>

{
  "jest": {
    "coverageThreshold": {
      "global": {
        "branches": 80,
        "functions": 80,
        "lines": 80,
        "statements": 80
      }
    }
  }
}

</coverage_config>

<on_complete> agentdb write-end '{"skill":"tdd","tests_written":<N>,"coverage":"<X%>","cycle":"red->green->refactor","failures_caught":["<list>"]}'

Record test count, coverage achieved, and what edge cases the tests catch. </on_complete>

</skill>