Engineering QATestingCoverage

Senior QA

Test generation, coverage analysis, E2E testing, and testing patterns.

$skillfish add ~skills/senior-qa

Real Examples Get Started Models & Context

Real Conversations

See how the Senior QA skill responds to actual requests. These are representative examples of what your AI agent will produce.

You

Building a new payment processing service in Node.js. It integrates Stripe, sends webhooks to merchants, and writes to Postgres. What's the testing strategy?

Agent

Testing Pyramid for Payment Service:

Unit tests (fast, no I/O) — 70% of suite:

// Test business logic in isolation — mock Stripe and DB
describe('calculateProcessingFee', () => {
  it('applies 2.9% + $0.30 for card payments under $100', () => {
    expect(calculateProcessingFee({ amount: 5000, method: 'card' })).toBe(175); // cents
  });
  it('applies 0.8% capped at $5 for ACH', () => {
    expect(calculateProcessingFee({ amount: 100000, method: 'ach' })).toBe(500);
  });
});

Integration tests — 20% of suite:

// Test Stripe integration against test mode (no mocks)
it('creates a PaymentIntent and returns client_secret', async () => {
  const stripe = new Stripe(process.env.STRIPE_TEST_KEY);
  const intent = await createPaymentIntent({ amount: 2000, currency: 'usd' });
  expect(intent.client_secret).toMatch(/^pi_.*_secret_/);
});

// Test DB writes with a real test database (testcontainers)
it('persists payment record after successful charge', async () => {
  await processPayment({ chargeId: 'ch_test_123', amount: 2000 });
  const record = await db.payments.findByChargeId('ch_test_123');
  expect(record.status).toBe('succeeded');
});

E2E / contract tests — 10% of suite:

Use Stripe's test webhook CLI to simulate webhook delivery: stripe trigger payment_intent.succeeded
Test idempotency: send the same webhook event twice, assert the DB record is created exactly once

Critical edge cases to cover:

Webhook signature validation (reject tampered events)
Network timeout on Stripe call (should retry with exponential backoff, not double-charge)
DB write failure after successful Stripe charge (reconciliation path)

You

We have 12 Jest tests that fail ~20% of the time in CI. They all test async operations — API calls, setTimeout callbacks, and event emitters. How do I fix them?

Agent

Flaky async tests have 3 root causes — diagnose first:

Pattern 1: Real timers in tests (most common)

// FLAKY: depends on actual wall clock time
setTimeout(() => { doSomething(); }, 100);
await new Promise(resolve => setTimeout(resolve, 150)); // race on slow CI

// FIXED: fake timers
jest.useFakeTimers();
setTimeout(() => { doSomething(); }, 100);
jest.advanceTimersByTime(100);  // deterministic
await Promise.resolve(); // flush microtasks
jest.useRealTimers();

Pattern 2: Event emitter — not waiting for emission

// FLAKY: event might fire before .on() is registered
emitter.emit('done', result);
const result = await new Promise(r => emitter.on('done', r)); // missed it

// FIXED: register listener before triggering
const resultPromise = new Promise(r => emitter.on('done', r));
emitter.emit('done', result);
const data = await resultPromise;

Pattern 3: Shared state between tests (test pollution)

// Add to every test file that mutates global/module state:
afterEach(() => {
  jest.clearAllMocks();  // clear mock call counts
  jest.resetModules();   // reset require cache if module state is shared
});

Find the flaky tests systematically:

# Run the test file 20 times and report failures
for i in {1..20}; do npx jest --testPathPattern=flaky.test.ts --silent; done

CI-specific: Add --runInBand to eliminate parallelism as a variable while debugging — if tests pass consistently with --runInBand, shared state is the culprit.

Get Started

How to install and use this skill in your preferred environment.

Skills are designed for AI coding agents (Claude Code, Cursor, Windsurf) and IDE-based workflows where the agent can read files, run scripts, and act on your codebase. Web-based AI can use the knowledge and frameworks, but won't have tool access.

Models & Context

Which AI models and context windows work best with this skill.

Recommended Models

Best

Claude Opus 4 Claude Sonnet 4 GPT-4.1 Gemini 2.5 Pro Grok 3 Kimi K2

Good

Claude Haiku 4.5 GPT-4.1 mini Gemini 2.5 Flash Grok 3 mini

Larger models produce more detailed, production-ready outputs.

Context Window

This skill's SKILL.md is typically 3–10 KB — fits in any modern context window.

8K Skill only

32K+ Skill + conversation

100K+ Skill + references + codebase

All current frontier models (Claude, GPT, Gemini) support 100K+ context. Use the full window for complex multi-service work.

Pro tips for best results

1

Be specific

Include numbers — users, budget, RPS — so the skill can size the architecture.

2

Share constraints

Compliance needs, team size, and existing stack all improve the output.

3

Iterate

Start with a high-level design, then ask follow-ups for IaC, cost analysis, or security review.

4

Combine skills

Pair with companion skills below for end-to-end coverage.

Ready to try Senior QA?

Install the skill and start getting expert-level guidance in your workflow — any agent, any IDE.

$skillfish add ~skills/senior-qa

← Browse all 169 skills