Skip to content
cccskills
FreePublished on GitHub

fixing-flaky-tests

Diagnose and fix flaky tests. Covers race conditions, timing issues, test pollution, async leaks.

The English summary is being prepared. Showing the original description.

See how to install

Files included (1)

  • SKILL.md3.1 KB

SKILL.md (original)

Review the instructions the agent will receive before you install.

Root Cause Categories

1. Timing Dependency

// BUG: assumes operation completes within 100ms
await new Promise(r => setTimeout(r, 100));
expect(result).toBe('done');

// FIX: wait for the actual condition
await vi.waitFor(() => expect(result).toBe('done'), { timeout: 5000 });
// or
await waitFor(() => screen.getByText('Done'));

2. Test Pollution (Shared State)

// BUG: test 1 creates user, test 2 fails because user already exists
it('test 1', () => createUser({ email: 'test@example.com' }));
it('test 2', () => createUser({ email: 'test@example.com' })); // CONFLICT

// FIX: clean up after each test
afterEach(async () => {
  await db.user.deleteMany({ where: { email: 'test@example.com' } });
});

// BETTER FIX: use unique values per test
const email = `test-${Date.now()}@example.com`;

3. Async Leak

// BUG: async operation continues after test ends
it('fetches data', () => {
  const result = [];
  fetchData().then(data => result.push(data)); // not awaited!
  // test ends, but fetchData is still running
  // it might complete and affect the next test
});

// FIX: always await async operations
it('fetches data', async () => {
  const data = await fetchData();
  expect(data).toBeDefined();
});

4. Global State / Singleton Pollution

// BUG: EventEmitter accumulates listeners across tests
const emitter = new EventEmitter(); // global singleton

// FIX: reset global state in beforeEach/afterEach
beforeEach(() => { emitter.removeAllListeners(); });

// BETTER: inject dependencies instead of using singletons

5. Order Dependency

// BUG: test B relies on state set by test A
describe('CartService', () => {
  it('test A: adds item', () => { cart.add(item); });
  it('test B: removes item', () => { cart.remove(item.id); }); // fails if A didn't run
});

// FIX: each test is self-contained
describe('CartService', () => {
  beforeEach(() => { cart = new Cart(); }); // fresh state
  it('test B: removes item that exists', () => {
    cart.add(item);       // set up in this test
    cart.remove(item.id);
    expect(cart.items).toHaveLength(0);
  });
});

Debugging Flaky Tests

# Run just the flaky test 20 times to see failure rate
for i in {1..20}; do vitest run --reporter verbose flaky.test.ts 2>&1 | grep -E "PASS|FAIL"; done

# Detect test order dependency
vitest run --sequence.shuffle  # random order

# Find long-running tests
vitest run --reporter verbose 2>&1 | grep "ms"

When to Skip vs Fix

Fix: All flaky tests should be fixed. "Flaky" = unreliable = wrong.

Skip temporarily (with explanation):

it.skip('FLAKY: #123 race condition in payment service', async () => {
  // This will be re-enabled when #123 is resolved
});

Never commit a flaky test without a ticket. Never.

Reviews

No reviews yet. Tell others what you thought.

Comments

No login needed. Share how it worked for you, ask a question, or note where it did not run.

No comments yet. Be the first to write one.

Comments are published right away. Do not include personal information such as email addresses or phone numbers. Up to 2 URLs.

0 / 1,000 characters

This form is protected by reCAPTCHA, and the Google Privacy Policy and Terms of Service apply.

More skills from this repository

anoa

Free

Drive a real browser from the command line — open pages, snapshot the interactive elements as refs, click and fill by ref, read text, run JS, screenshot. Use when a task needs a live browser: checking a page renders, walking a login or checkout flow, scraping something that only exists after JavaScript runs, or reproducing a UI bug.

The English summary is being prepared. Showing the original description.

porcupine-md/jonggrang16Updated Sep 15, 2026

Run linters/typecheckers to calculate health score.

The English summary is being prepared. Showing the original description.

porcupine-md/jonggrang16Updated Sep 15, 2026

Comprehensive UI/UX compliance check (A11y, performance, responsive).

The English summary is being prepared. Showing the original description.

porcupine-md/jonggrang16Updated Sep 15, 2026

Infrastructure-first security audit (OWASP, secrets, dependencies).

The English summary is being prepared. Showing the original description.

porcupine-md/jonggrang16Updated Sep 15, 2026

auth

Free

Setup authentication flow (register, login, session/JWT, middleware)

The English summary is being prepared. Showing the original description.

porcupine-md/jonggrang16Updated Sep 15, 2026

Deep knowledge about deploying applications to AWS (EC2, ECS, S3, CloudFront, Cloud Run).

The English summary is being prepared. Showing the original description.

porcupine-md/jonggrang16Updated Sep 15, 2026

See all skills by porcupine-md

Report a problem with this skill