developing-test-driven

name	developing-test-driven
description	AI agent practices test-first development with the Red-Green-Refactor cycle for confident, well-designed code. Use when implementing features, fixing bugs, or establishing testing practices.
category	methodology
related_skills	["methodology/test-enforcement","methodology/test-task-generation","testing/vitest","testing/pytest"]
related_commands	["/dev:tdd","/dev:feature-tested","/dev:test-write","/quality:verify-done"]

Developing Test-Driven

Quick Start

1. Red - Write a failing test that defines desired behavior
2. Green - Write minimal code to make the test pass
3. Refactor - Improve code while keeping tests green
4. Repeat - Continue with next behavior (1-5 minute cycles)

Features

Feature	Description	Guide
Red-Green-Refactor	Core TDD cycle	Fail -> Pass -> Improve -> Repeat
Test Patterns	Effective test structures	Arrange-Act-Assert, Given-When-Then
Test Fixtures	Reusable test setup	beforeEach, factories, builders
Mocking Strategies	Isolate dependencies	Inject deps, mock boundaries only
Parameterized Tests	Test multiple inputs	it.each for input variations
Coverage Analysis	Verify thoroughness	Statements, branches, functions

Common Patterns

// RED-GREEN-REFACTOR CYCLE

// RED: Write failing test
it('rejects passwords shorter than 8 characters', () => {
  const result = validatePassword('short');
  expect(result.valid).toBe(false);
  expect(result.errors).toContain('Password must be at least 8 characters');
});

// GREEN: Minimal implementation
function validatePassword(password: string) {
  const errors = [];
  if (password.length < 8) {
    errors.push('Password must be at least 8 characters');
  }
  return { valid: errors.length === 0, errors };
}

// REFACTOR: Extract rules (tests stay green)
const rules = [
  { test: (p) => p.length >= 8, message: 'Must be 8+ chars' },
  { test: (p) => /[A-Z]/.test(p), message: 'Must have uppercase' },
];
function validatePassword(password: string) {
  const errors = rules.filter(r => !r.test(password)).map(r => r.message);
  return { valid: errors.length === 0, errors };
}

// ARRANGE-ACT-ASSERT Pattern
it('increases total when item added', () => {
  // ARRANGE
  const cart = new ShoppingCart();
  const item = { id: '1', price: 10.00 };

  // ACT
  cart.addItem(item);

  // ASSERT
  expect(cart.total).toBe(10.00);
});

// PARAMETERIZED TESTS
it.each([
  ['[e-mail]', true],
  ['[e-mail]', true],
  ['invalid', false],
  ['@nodomain.com', false],
])('validates %s as %s', (email, expected) => {
  expect(isValidEmail(email)).toBe(expected);
});

// BUILDER PATTERN for test data
const user = new UserBuilder().admin().inactive().build();

# Mocking Guidelines
MOCK (external boundaries):
- External APIs (payment, email)
- Third-party services
- System clock, random
- Network requests

DON'T MOCK (your code):
- Pure functions
- Data transformations
- Business logic
- Internal services

# Coverage Analysis
| Type | Target | Notes |
|------|--------|-------|
| Statements | 80%+ | Code executed |
| Branches | 70%+ | If/else paths |
| Functions | 90%+ | All functions called |

REMEMBER: 100% coverage != well-tested
- Test edge cases
- Test error paths
- Test boundary conditions

Best Practices

Do	Avoid
Write test BEFORE implementation	Writing implementation first
Keep tests focused on ONE behavior	Testing multiple things in one test
Use descriptive test names	Generic names like "test1"
Run tests frequently (every few minutes)	Long gaps between test runs
Refactor only when tests green	Refactoring with failing tests
Test behavior, not implementation	Testing private methods directly
Keep cycles short (1-5 minutes)	30+ minute cycles
Mock only external dependencies	Over-mocking your own code

Command Integration

Using TDD Command

# Start TDD workflow
/dev:tdd "user authentication"

# TDD with specific coverage
/dev:tdd "payment processing" --coverage 95

# TDD with specific test types
/dev:tdd "API endpoint" --test-types unit,integration,contract

Configuration

# .omgkit/workflow.yaml
testing:
  enabled: true
  enforcement:
    level: strict  # TDD recommends strict
  coverage_gates:
    unit:
      minimum: 90
      target: 95

Related Skills

• methodology/test-enforcement - Enforce test completion
• methodology/test-task-generation - Auto-generate test tasks
• methodology/testing-anti-patterns - Avoid common test mistakes
• testing/vitest - Vitest testing framework
• testing/playwright - E2E testing with Playwright