aiskillstore/analyzing-test-quality

Analyzing Test Quality

You are an expert in test quality analysis with deep knowledge of testing principles, patterns, and metrics that apply across all testing frameworks.

Your Capabilities

1. Quality Metrics: Coverage, mutation score, test effectiveness
2. Test Patterns: AAA, GWT, fixtures, factories, page objects
3. Anti-Patterns: Flaky tests, test pollution, over-mocking
4. Maintainability: DRY, readability, test organization
5. Reliability: Determinism, isolation, independence
6. Coverage Analysis: Statement, branch, function, line coverage

When to Use This Skill

Claude should automatically invoke this skill when:

• The user asks about test quality or test effectiveness
• Code coverage reports or metrics are discussed
• Test reliability or flakiness is mentioned
• Test organization or refactoring is needed
• General test improvement is requested

How to Use This Skill

Accessing Resources

Use {baseDir} to reference files in this skill directory:

• Scripts: {baseDir}/scripts/
• Documentation: {baseDir}/references/
• Templates: {baseDir}/assets/

Available Resources

This skill includes ready-to-use resources in {baseDir}:

• references/quality-checklist.md - Printable test quality checklist with scoring guide
• assets/quality-report.template.md - Complete template for test quality assessment reports
• scripts/calculate-metrics.sh - Calculates test metrics (test count, ratios, patterns, assertions)

Test Quality Dimensions

1. Correctness

Tests accurately verify intended behavior:

• Tests match requirements
• Assertions are complete
• Edge cases are covered
• Error scenarios are tested

2. Readability

Tests are easy to understand:

• Clear naming (what is being tested)
• Proper structure (AAA/GWT pattern)
• Minimal setup noise
• Self-documenting code

3. Maintainability

Tests are easy to modify:

• DRY with appropriate helpers
• Focused tests (single responsibility)
• Proper abstraction level
• Clear dependencies

4. Reliability

Tests produce consistent results:

• No timing dependencies
• Proper isolation
• Deterministic data
• Independent execution

5. Speed

Tests run efficiently:

• Appropriate test pyramid
• Efficient setup/teardown
• Proper mocking strategy
• Parallel execution

Test Quality Checklist

Structure

• [ ] Uses AAA (Arrange-Act-Assert) or GWT pattern
• [ ] One logical assertion per test
• [ ] Descriptive test names
• [ ] Proper describe/context nesting
• [ ] Appropriate setup/teardown

Coverage

• [ ] Happy path scenarios
• [ ] Error/edge cases
• [ ] Boundary conditions
• [ ] Integration points
• [ ] Security scenarios

Reliability

• [ ] No timing dependencies
• [ ] Proper async handling
• [ ] Isolated tests (no shared state)
• [ ] Deterministic data
• [ ] Order-independent

Maintainability

• [ ] Reusable fixtures/factories
• [ ] Clear variable naming
• [ ] Focused assertions
• [ ] Appropriate abstraction
• [ ] No magic numbers/strings

Common Anti-Patterns

Test Pollution

// BAD: Shared mutable state
let count = 0;
beforeEach(() => count++);

// GOOD: Reset in setup
let count: number;
beforeEach(() => { count = 0; });

Over-Mocking

Mocking too much hides bugs and makes tests brittle.

// BAD: Mock everything - test only verifies mocks
// Jest
jest.mock('./dep1');
jest.mock('./dep2');
jest.mock('./dep3');

// Vitest
vi.mock('./dep1');
vi.mock('./dep2');
vi.mock('./dep3');

// GOOD: Mock boundaries only
// Mock external services, keep internal logic real
mock('./api'); // External service only
// Test actual business logic

Flaky Assertions

// BAD: Timing dependent
await delay(100);
expect(element).toBeVisible();

// GOOD: Wait for condition
// Testing Library
await waitFor(() => expect(element).toBeVisible());

// Playwright
await expect(element).toBeVisible();

Mystery Guest

// BAD: Hidden dependencies
test('should process', () => {
  const result = process(); // Uses global data
  expect(result).toBe(42);
});

// GOOD: Explicit setup
test('should process input', () => {
  const input = createInput({ value: 21 });
  const result = process(input);
  expect(result).toBe(42);
});

Assertion Roulette

// BAD: Multiple unrelated assertions
test('should work', () => {
  expect(user.name).toBe('John');
  expect(items.length).toBe(3);
  expect(total).toBe(100);
});

// GOOD: Focused assertions
test('should set user name', () => {
  expect(user.name).toBe('John');
});

test('should have correct item count', () => {
  expect(items).toHaveLength(3);
});

Mutation Testing

Mutation testing validates test effectiveness by modifying code and checking if tests catch the changes.

Concept

1. Mutants are created by modifying source code (changing operators, values, etc.)
2. Tests run against each mutant
3. Killed mutants = tests caught the change (good!)
4. Survived mutants = tests missed the change (weak tests)

Stryker Setup

# Install Stryker
npm install -D @stryker-mutator/core

# For specific frameworks
npm install -D @stryker-mutator/jest-runner      # Jest
npm install -D @stryker-mutator/vitest-runner    # Vitest
npm install -D @stryker-mutator/mocha-runner     # Mocha

# Initialize configuration
npx stryker init

Stryker Configuration

// stryker.conf.js
module.exports = {
  packageManager: 'npm',
  reporters: ['html', 'clear-text', 'progress'],
  testRunner: 'jest',
  coverageAnalysis: 'perTest',

  // What to mutate
  mutate: [
    'src/**/*.ts',
    '!src/**/*.test.ts',
    '!src/**/*.spec.ts',
  ],

  // Mutation types to use
  mutator: {
    excludedMutations: [
      'StringLiteral', // Skip string mutations
    ],
  },

  // Thresholds
  thresholds: {
    high: 80,
    low: 60,
    break: 50, // Fail CI if below this
  },
};

Interpreting Results

Mutation score: 85%
Killed: 170 | Survived: 30 | Timeout: 5 | No coverage: 10

High score (>80%): Tests are effective Medium score (60-80%): Some weak areas Low score (<60%): Tests need significant improvement

Common Surviving Mutations

Boundary mutations: < changed to <=

// Mutation survives if tests don't check boundary
if (value < 10) { ... }  // Changed to: value <= 10

Arithmetic mutations: + changed to -

// Mutation survives if result isn't precisely checked
return a + b;  // Changed to: a - b

Boolean mutations: && changed to ||

// Mutation survives if both conditions aren't tested
if (a && b) { ... }  // Changed to: a || b

CI Integration

# GitHub Actions
- name: Run mutation tests
  run: npx stryker run

- name: Upload Stryker report
  uses: actions/upload-artifact@v3
  with:
    name: stryker-report
    path: reports/mutation/

Coverage Metrics

Types of Coverage

• Statement: Lines executed
• Branch: Decision paths taken
• Function: Functions called
• Line: Lines covered

Coverage Thresholds

// Recommended minimums
{
  statements: 80,
  branches: 75,
  functions: 80,
  lines: 80
}

Coverage Pitfalls

• High coverage ≠ good tests
• Can miss logical errors
• Doesn't test interactions
• Can incentivize bad tests

Mutation Testing

Concept

Mutation testing modifies code to check if tests catch the changes:

• Tests should fail when code is mutated
• Surviving mutants indicate weak tests
• Higher kill rate = better tests

Types of Mutations

• Arithmetic operators (+, -, *, /)
• Comparison operators (<, >, ==)
• Boolean operators (&&, ||, !)
• Return values
• Constants

Test Pyramid

Unit Tests (Base)

• Fast execution
• Isolated components
• High coverage
• Many tests

Integration Tests (Middle)

• Component interactions
• Database/API calls
• Moderate coverage
• Medium quantity

E2E Tests (Top)

• Full user flows
• Real browser
• Critical paths only
• Few tests

Analysis Workflow

When analyzing test quality:

1. Gather Metrics

   • Run coverage report
   • Count test/code ratio
   • Measure test execution time

2. Identify Patterns

   • Check test structure
   • Look for anti-patterns
   • Assess naming quality

3. Evaluate Reliability

   • Check for flaky indicators
   • Assess isolation
   • Review async handling

4. Provide Recommendations

   • Prioritize by impact
   • Give specific examples
   • Include code samples

Examples

Example 1: Coverage Analysis

When analyzing coverage:

1. Run coverage tool
2. Identify uncovered lines
3. Prioritize critical paths
4. Suggest test cases

Example 2: Reliability Audit

When auditing for reliability:

1. Search for timing patterns
2. Check shared state usage
3. Review async assertions
4. Identify order dependencies

Important Notes

• Quality is more important than quantity
• Coverage is a starting point, not a goal
• Fast feedback enables TDD
• Readable tests serve as documentation
• Test maintenance cost should be low