Comprehensive Testing

Based on Anthropic's Claude Code Best Practices and community patterns

Core Philosophy

"Claude performs best when it has a clear target to iterate against—a test case provides concrete success criteria."

Testing is not about proving code works; it's about designing code that is testable and documenting expected behavior.

---

Test Pyramid

        /\
       /  \        E2E Tests (10%)
      /----\       - Full user flows
     /      \      - Slowest, most brittle
    /--------\
   /          \    Integration Tests (20%)
  /------------\   - Component interaction
 /              \  - Real dependencies
/----------------\
       Unit Tests (70%)
       - Single function/method
       - Fast, isolated, many

Level	Speed	Scope	When to Use
Unit	<10ms	Single function	All logic
Integration	<1s	Multiple components	APIs, DB
E2E	<30s	Full flow	Critical paths

---

TDD Workflow (Anthropic Recommended)

The 6-Step Process

1. WRITE TESTS FIRST
   ↓
2. VERIFY TESTS FAIL
   ↓
3. COMMIT TEST SUITE
   ↓
4. IMPLEMENT CODE
   ↓
5. VERIFY WITH SUBAGENT
   ↓
6. COMMIT IMPLEMENTATION

Step 1: Write Tests First

Be EXPLICIT about TDD to avoid mock implementations:

"I want to implement [feature] using TDD.
First, write tests for [expected behavior] with these input/output pairs:
- Input: X → Expected: Y
- Input: A → Expected: B
Do NOT create any implementation yet."

Step 2: Verify Tests Fail

# Run tests and confirm they fail for the RIGHT reason
npm test  # or pytest, go test, etc.

# Expected: "function not found" or "undefined"
# NOT: syntax error, wrong import

Step 3: Commit Test Suite

git add tests/
git commit -m "test: Add tests for [feature] (RED phase)"

Step 4: Implement Incrementally

"Now implement the code to make these tests pass.
Do NOT modify the tests.
Run tests after each change until all pass."

Claude will enter an autonomous loop:

Write code → Run tests → Analyze failures → Adjust → Repeat

Step 5: Verify with Subagent

"Use a subagent to independently verify the implementation:
- Is it overfitting to tests?
- Are edge cases handled?
- Is the code maintainable?"

Step 6: Commit Implementation

git add src/
git commit -m "feat: Implement [feature] (GREEN phase)"

---

Test Structure Patterns

AAA Pattern (Arrange-Act-Assert)

describe('UserService', () => {
  it('should create user with valid email', async () => {
    // Arrange - Setup test data and dependencies
    const userRepo = new InMemoryUserRepository();
    const service = new UserService(userRepo);
    const input = { email: 'test@example.com', name: 'Test' };

    // Act - Execute the code under test
    const user = await service.create(input);

    // Assert - Verify the results
    expect(user.email).toBe('test@example.com');
    expect(user.id).toBeDefined();
    expect(await userRepo.findById(user.id)).toEqual(user);
  });
});

Given-When-Then Pattern

def test_order_total_with_discount():
    """
    Given an order with items totaling $100
    When a 20% discount is applied
    Then the total should be $80
    """
    # Given
    order = Order()
    order.add_item(Item(price=50))
    order.add_item(Item(price=50))

    # When
    order.apply_discount(Percentage(20))

    # Then
    assert order.total == Money(80)

---

Framework Best Practices

Jest / Vitest (JavaScript/TypeScript)

// Structure
describe('ModuleName', () => {
  describe('methodName', () => {
    it('should [expected behavior] when [condition]', () => {});
  });
});

// Setup/Teardown
beforeAll(async () => { /* one-time setup */ });
beforeEach(() => { /* per-test setup */ });
afterEach(() => { /* per-test cleanup */ });
afterAll(async () => { /* one-time cleanup */ });

// Async testing
it('handles async operations', async () => {
  const result = await asyncFunction();
  expect(result).toBe(expected);
});

// Error testing
it('throws on invalid input', () => {
  expect(() => validate(null)).toThrow('Input required');
});

// Snapshot testing (use sparingly)
it('renders correctly', () => {
  const tree = renderer.create(<Component />).toJSON();
  expect(tree).toMatchSnapshot();
});

// Table-driven tests
it.each([
  [1, 1, 2],
  [2, 2, 4],
  [0, 0, 0],
])('add(%i, %i) = %i', (a, b, expected) => {
  expect(add(a, b)).toBe(expected);
});

vitest.config.ts:

export default defineConfig({
  test: {
    globals: true,
    environment: 'node',
    coverage: {
      provider: 'v8',
      reporter: ['text', 'json', 'html'],
      thresholds: {
        lines: 80,
        branches: 70,
        functions: 80,
      },
    },
  },
});

pytest (Python)

import pytest
from mymodule import Calculator

# Fixtures for dependency injection
@pytest.fixture
def calculator():
    return Calculator()

@pytest.fixture
def database():
    db = TestDatabase()
    yield db
    db.cleanup()

# Parametrized tests
@pytest.mark.parametrize("a,b,expected", [
    (1, 1, 2),
    (2, 2, 4),
    (0, 0, 0),
    (-1, 1, 0),
])
def test_add(calculator, a, b, expected):
    assert calculator.add(a, b) == expected

# Exception testing
def test_divide_by_zero(calculator):
    with pytest.raises(ZeroDivisionError):
        calculator.divide(1, 0)

# Async testing
@pytest.mark.asyncio
async def test_async_operation():
    result = await async_function()
    assert result == expected

# Markers for categorization
@pytest.mark.slow
@pytest.mark.integration
def test_database_connection(database):
    assert database.is_connected()

conftest.py:

import pytest

@pytest.fixture(scope="session")
def database_url():
    return "postgresql://test:test@localhost/test"

@pytest.fixture(autouse=True)
def reset_database(database):
    yield
    database.rollback()

pytest.ini:

[pytest]
testpaths = tests
python_files = test_*.py
python_functions = test_*
addopts = -v --cov=src --cov-report=term-missing
markers =
    slow: marks tests as slow
    integration: marks tests as integration tests

Go Testing

package mypackage

import (
    "testing"
    "github.com/stretchr/testify/assert"
    "github.com/stretchr/testify/require"
)

func TestAdd(t *testing.T) {
    tests := []struct {
        name     string
        a, b     int
        expected int
    }{
        {"positive numbers", 1, 2, 3},
        {"zero values", 0, 0, 0},
        {"negative numbers", -1, -2, -3},
    }

    for _, tt := range tests {
        t.Run(tt.name, func(t *testing.T) {
            result := Add(tt.a, tt.b)
            assert.Equal(t, tt.expected, result)
        })
    }
}

// Table-driven with subtests
func TestUserService_Create(t *testing.T) {
    t.Run("creates user with valid input", func(t *testing.T) {
        repo := NewInMemoryRepo()
        svc := NewUserService(repo)

        user, err := svc.Create(CreateUserInput{Email: "test@example.com"})

        require.NoError(t, err)
        assert.NotEmpty(t, user.ID)
        assert.Equal(t, "test@example.com", user.Email)
    })

    t.Run("returns error for invalid email", func(t *testing.T) {
        repo := NewInMemoryRepo()
        svc := NewUserService(repo)

        _, err := svc.Create(CreateUserInput{Email: "invalid"})

        require.Error(t, err)
        assert.Contains(t, err.Error(), "invalid email")
    })
}

---

Mock Strategy

When to Mock

Scenario	Mock?	Reason
External APIs	✅ Yes	Slow, unreliable, costs money
Time/Date	✅ Yes	Non-deterministic
Random	✅ Yes	Non-deterministic
Database (unit)	✅ Yes	Slow, complex setup
Database (integration)	❌ No	Test real behavior
Your own code	⚠️ Rarely	Prefer real implementations
File system	⚠️ Depends	Use temp dirs when possible

How to Mock

// Jest - Mock module
jest.mock('./emailService', () => ({
  sendEmail: jest.fn().mockResolvedValue({ success: true }),
}));

// Jest - Mock function
const mockCallback = jest.fn();
mockCallback.mockReturnValue(42);

// Vitest - Spy
import { vi } from 'vitest';
const spy = vi.spyOn(console, 'log');

// Time mocking
beforeEach(() => {
  vi.useFakeTimers();
  vi.setSystemTime(new Date('2024-01-01'));
});

afterEach(() => {
  vi.useRealTimers();
});

# pytest - Mock with unittest.mock
from unittest.mock import Mock, patch, MagicMock

@patch('mymodule.external_api.fetch')
def test_with_mocked_api(mock_fetch):
    mock_fetch.return_value = {'data': 'mocked'}
    result = my_function()
    assert result == expected
    mock_fetch.assert_called_once_with('expected_arg')

# Fixture-based mock
@pytest.fixture
def mock_email_service():
    service = Mock()
    service.send.return_value = True
    return service

Prefer Test Doubles Over Mocks

// ❌ Heavy mocking
const mockRepo = {
  findById: jest.fn().mockResolvedValue({ id: '1', name: 'Test' }),
  save: jest.fn().mockResolvedValue(undefined),
  delete: jest.fn().mockResolvedValue(undefined),
};

// ✅ In-memory implementation (test double)
class InMemoryUserRepository implements UserRepository {
  private users: Map<string, User> = new Map();

  async findById(id: string): Promise<User | null> {
    return this.users.get(id) || null;
  }

  async save(user: User): Promise<User> {
    this.users.set(user.id, user);
    return user;
  }

  async delete(id: string): Promise<void> {
    this.users.delete(id);
  }
}

---

Extended Reference

Detailed material starting at ## Boundary Testing has been moved to reference/extended.md to keep this skill concise. Load that reference when the task requires the moved examples, command catalogs, checklists, platform details, or implementation templates.