Mandatory Testing & Linting Guard

Mandatory Testing & Linting Guard

This skill enforces the CRITICAL requirement: After EVERY code change, you MUST run npm test and npm run lint. No exceptions for "small changes" or "trivial updates".

When to Use This Skill

ALWAYS. This skill should be invoked automatically after:

• Any code modification (TypeScript, JavaScript, CSS)
• Any component creation or update
• Any service or utility change
• Any configuration file modification
• Before creating a git commit
• Before creating a pull request

Trigger phrases:

• "I've made changes to..."
• "I updated the code..."
• "I created a new component..."
• "I fixed the bug..."
• "I added a feature..."

Critical Requirement

From CLAUDE.md:

CRITICAL: After EVERY code change, you MUST run:

1. npm test - Ensure all tests pass
2. npm run lint - Verify code quality

Never proceed without both passing. No exceptions for "small changes".

Standard Workflow

Workflow Overview

Code Change → npm test → npm run lint → Commit/PR
     ↓           ↓            ↓              ↓
   Edit      875 tests    ESLint       Git commit
              MUST pass   MUST pass    (only if both pass)

Step-by-Step Process

1. After Making Code Changes

Action: Run the test suite:

npm test

Expected output:

✓ 875 tests passed
  Test Files  46 passed (46)
  Tests  875 passed (875)
  Duration  12.5s

What to check:

• All 875+ tests pass (or current total)
• No test failures or errors
• No warnings about missing coverage
• Coverage threshold (50%) is met

If tests fail:

• DO NOT proceed to linting
• Fix the failing tests first
• Re-run npm test until all pass

2. After Tests Pass

Action: Run ESLint:

npm run lint

Expected output:

✔ No ESLint warnings or errors found

What to check:

• Zero errors
• Zero warnings
• All files comply with style guide

If linting fails:

• Try auto-fix first: npm run lint:fix
• Manually fix remaining issues
• Re-run npm run lint until clean

3. Both Pass - Ready to Commit

Action: Only NOW can you proceed with:

• Creating a git commit
• Pushing to remote
• Creating a pull request
• Marking task as complete

Test Suite Overview

Current Statistics

• Total Tests: 875+
• Test Files: 46
• Coverage Threshold: 50% (statements)
• Test Framework: Vitest + React Testing Library
• Environment: happy-dom

Test Categories

Components (60% coverage):

• 40+ React component tests
• Rendering, interactions, state changes
• Accessibility checks
• Dark mode compatibility

Services (80% coverage):

• StorageManager singleton tests
• PromptManager business logic
• Validation and error handling
• Levenshtein distance algorithms

Hooks (70% coverage):

• Custom React hooks
• usePrompts, useCategories, useSettings
• Async state management

Content Scripts (40% coverage):

• Platform strategy tests
• Injection and insertion logic
• DOM manipulation

Utilities (90% coverage):

• Helper functions
• Formatters and validators
• Logger utilities

Key Test Files

Critical test locations:

src/
├── components/__tests__/         # Component tests
│   ├── PromptCard.test.tsx
│   ├── AddPromptForm.test.tsx
│   └── ... (40+ files)
├── services/__tests__/           # Service tests
│   ├── storage.test.ts
│   └── promptManager.test.ts
├── hooks/__tests__/              # Hook tests
│   └── usePrompts.test.ts
└── content/platforms/__tests__/  # Platform tests
    ├── claude-strategy.test.ts
    └── ... (platform strategies)

Common Test Errors

Error 1: Test Timeout

Symptom:

Error: Test timed out after 5000ms

Solutions:

1. Add timeout to slow async tests:

   it('loads large dataset', async () => {
     // ...
   }, 10000); // 10 second timeout
   

2. Use waitFor for async state updates:

   await waitFor(() => {
     expect(screen.getByText('Loaded')).toBeInTheDocument();
   });
   

Error 2: Mock Not Called

Symptom:

AssertionError: expected mock to have been called

Solutions:

1. Verify mock is setup correctly:

   expect(mockFn).toHaveBeenCalled();
   

2. Check you're calling the right mock:

   expect(StorageManager.getInstance().getPrompts).toHaveBeenCalled();
   

3. Clear mocks between tests:

   beforeEach(() => {
     vi.clearAllMocks();
   });
   

Error 3: DOM Element Not Found

Symptom:

Unable to find element with text: "Submit"

Solutions:

1. Use correct query method:

   // Wait for async rendering
   await screen.findByText('Submit');
   
   // Query by role
   screen.getByRole('button', { name: /submit/i });
   

2. Debug the rendered output:

   const { debug } = render(<Component />);
   debug(); // Prints current DOM
   

Error 4: Coverage Below Threshold

Symptom:

ERROR: Coverage for statements (48%) is below threshold (50%)

Solutions:

1. Generate coverage report to see gaps:

   npm run test:coverage
   open coverage/index.html
   

2. Add tests for uncovered code paths:

   • Error handling branches
   • Edge cases
   • Conditional logic

3. Check if new files need tests:

   # See which files have low coverage
   npm run test:coverage | grep -E "^[^│]*│[^│]*│[^│]*[0-9]+\.[0-9]+%"
   

Error 5: React 19 Hook Errors

Symptom:

Error: useActionState is not defined

Note: React 19 hooks (useActionState, useOptimistic) are not yet supported in Node.js test environments.

Solutions:

1. Test these hooks manually in browser
2. Skip automated tests for React 19-specific hooks
3. Use integration tests instead of unit tests
4. Document manual testing in PR description

ESLint Common Issues

Issue 1: Unused Variables

Symptom:

'variable' is defined but never used

Solutions:

1. Remove unused variables
2. Prefix with underscore if intentionally unused:

   const _unusedParam = value; // Intentionally unused
   

3. Use in function signature but not body:

   function canHandle(_element: HTMLElement): boolean {
     // _element is part of interface but not needed
   }
   

Issue 2: Missing Return Type

Symptom:

Missing return type on function

Solutions:

1. Add explicit return type:

   // Before
   async function getData() {
     return fetch('/api');
   }
   
   // After
   async function getData(): Promise<Response> {
     return fetch('/api');
   }
   

Issue 3: Any Type Usage

Symptom:

Unexpected any. Specify a different type

Solutions:

1. Use specific type instead:

   // Before
   const data: any = JSON.parse(str);
   
   // After
   const data: Prompt = JSON.parse(str);
   

2. Use unknown for truly unknown types:

   try {
     // ...
   } catch (error) {
     const err = error as Error; // Cast from unknown
   }
   

Issue 4: Missing Dependencies in useEffect

Symptom:

React Hook useEffect has missing dependencies

Solutions:

1. Add missing dependencies:

   useEffect(() => {
     loadData(id);
   }, [id]); // Add id to dependency array
   

2. If intentionally omitted, add comment:

   useEffect(() => {
     loadOnce();
   }, []); // eslint-disable-line react-hooks/exhaustive-deps
   

Issue 5: Console Statements

Symptom:

Unexpected console statement

Solution: Use the centralized Logger instead of console.*:

// ❌ DON'T
console.log('Debug message');
console.error('Error occurred');

// ✅ DO
import { Logger } from '../utils';

Logger.debug('Debug message', { component: 'MyComponent' });
Logger.error('Error occurred', toError(err), { component: 'MyComponent' });

Auto-Fix Strategies

Strategy 1: Auto-Fix Linting Issues

Action:

npm run lint:fix

This automatically fixes:

• Formatting issues
• Missing semicolons
• Import order
• Unused imports
• Simple style violations

Remaining issues must be fixed manually.

Strategy 2: Test-Driven Fixes

When tests fail, use this workflow:

1. Identify the failing test:

   npm test -- --reporter=verbose
   

2. Run only that test file:

   npm test -- PromptCard.test.tsx
   

3. Enable watch mode for rapid iteration:

   npm test -- PromptCard.test.tsx --watch
   

4. Fix the code until test passes

5. Run full suite again:

   npm test
   

Strategy 3: Coverage-Driven Testing

When coverage is low:

1. Generate coverage report:

   npm run test:coverage
   

2. Open HTML report:

   open coverage/index.html
   

3. Find red/yellow highlighted lines (uncovered code)

4. Add tests for those code paths

5. Re-run coverage check

Emergency Override Scenarios

Scenario 1: Pre-commit Hook Failure

When: Git commit fails due to test/lint errors

NEVER override with:

git commit --no-verify  # ❌ NEVER DO THIS

Instead:

1. Fix the errors properly
2. Run tests and linting again
3. Then commit normally

Scenario 2: CI/CD Pipeline Failure

When: GitHub Actions fails on PR

NEVER merge with:

• Disabling required checks
• Admin override
• Force push

Instead:

1. Pull latest changes: git pull origin main
2. Fix the errors locally
3. Run tests and linting
4. Push the fixes
5. Wait for CI to pass

Scenario 3: Urgent Hotfix Needed

When: Production bug needs immediate fix

Even for hotfixes:

1. Create the fix
2. Run npm test
3. Run npm run lint
4. Both MUST pass
5. Then deploy

Rationale: Broken tests = broken code. Never skip.

Integration with Development Workflow

Git Commit Workflow

# 1. Make your changes
vim src/components/MyComponent.tsx

# 2. MANDATORY: Run tests
npm test

# 3. MANDATORY: Run linting
npm run lint

# 4. If both pass, stage changes
git add src/components/MyComponent.tsx

# 5. Commit
git commit -m "feat: add MyComponent

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>"

# 6. Push
git push origin feature-branch

Pull Request Workflow

Before creating PR:

# 1. Ensure on latest main
git checkout main
git pull origin main

# 2. Rebase feature branch
git checkout feature-branch
git rebase main

# 3. Run full test suite
npm test

# 4. Run linting
npm run lint

# 5. Run build to ensure no build errors
npm run build

# 6. All three must pass before creating PR

In PR description, include:

## Testing Checklist

- [x] All 875+ tests pass locally
- [x] ESLint shows no errors or warnings
- [x] Coverage threshold maintained (50%+)
- [x] Manual testing completed on:
  - [ ] Chrome browser
  - [ ] AI platforms: Claude, ChatGPT, Perplexity
  - [ ] Light and dark mode

Continuous Integration

GitHub Actions Checks

The project has automated CI checks that run on every PR:

PR Checks (.github/workflows/pr-checks.yml):

- name: Run Tests
  run: npm test

- name: Run Linting
  run: npm run lint

- name: Check Coverage
  run: npm run test:coverage

All checks must pass before PR can be merged.

Pre-commit Hooks

The project uses Husky for pre-commit hooks:

{
  "husky": {
    "hooks": {
      "pre-commit": "lint-staged"
    }
  },
  "lint-staged": {
    "*.{ts,tsx}": [
      "eslint --fix",
      "vitest related --run"
    ]
  }
}

This automatically runs:

1. ESLint on changed files
2. Related tests for changed code

Quick Reference

Essential Commands

Success Indicators

Tests Pass:

✓ 875 tests passed
✓ Test Files  46 passed (46)
✓ Coverage: 55% (exceeds 50% threshold)

Linting Pass:

✔ No ESLint warnings or errors found

Failure Indicators

Tests Fail:

✖ 3 tests failed
✖ Expected: true, Received: false

Action: Fix tests before proceeding

Linting Fail:

✖ 5 problems (3 errors, 2 warnings)

Action: Run npm run lint:fix, then fix remaining issues

Enforcement Checklist

Before marking ANY task complete, verify:

• All code changes are complete
• npm test has been run
• All 875+ tests pass (or current total)
• npm run lint has been run
• Zero ESLint errors
• Zero ESLint warnings
• Coverage threshold (50%) maintained
• No console errors in test output
• Build succeeds (npm run build)

Only after ALL checkboxes are checked can you:

• Create a git commit
• Push to remote repository
• Create a pull request
• Mark the task as complete

Related Documentation

• Testing Guide: docs/TESTING.md - Comprehensive testing strategies
• Development Workflow: CLAUDE.md - Essential commands and guidelines
• Architecture: docs/ARCHITECTURE.md - System design patterns

Philosophy

Why This Matters:

1. Test Coverage: 875+ tests ensure nothing breaks
2. Code Quality: ESLint catches bugs before they ship
3. Consistency: Every developer follows the same standards
4. Confidence: Changes don't introduce regressions
5. Documentation: Tests serve as living documentation

The Rule:

"If it's too small to test, it's too small to break."

Even "trivial" changes can cause cascading failures. Always test. Always lint. No exceptions.

Last Updated: 2025-10-18 Skill Version: 1.0.0