e2e

E2E — End-to-End Testing

Activate When

• User invokes /godmode:e2e
• User says "end to end", "end-to-end", "e2e test", "user flow test", "checkout flow"
• User says "E2E test", "browser test", "integration test"
• User asks about Playwright, Cypress, or Selenium
• User needs to fix flaky E2E tests
• Ship skill needs user-flow validation

Workflow

Step 1: Assess E2E Test State

# Find existing E2E tests
find . -name "*.e2e.*" -o -name "*.spec.*" \
  | grep -i -E "e2e|playwright|cypress" \
  | grep -v node_modules

# Check for test framework config
ls playwright.config.* cypress.config.* wdio.conf.* \
  2>/dev/null

# Count existing test files
find e2e/ tests/ -name "*.spec.ts" 2>/dev/null | wc -l

E2E TEST STATE:
Framework: <Playwright | Cypress | Selenium | none>
Config file: <path or none>
Test count: <N>
Page objects: <N> (or none)
Flaky tests: <N known flaky | unknown>

IF test count == 0: scaffold from scratch
IF test count > 0 AND flaky > 3: prioritize remediation
IF framework == none: default to Playwright

Step 2: Design Test Architecture

e2e/
├── tests/               # By feature
│   ├── auth/
│   │   ├── login.spec.ts
│   │   └── registration.spec.ts
│   ├── checkout/
│   │   ├── cart.spec.ts
│   │   └── payment.spec.ts
│   └── dashboard/
│       └── navigation.spec.ts
├── pages/               # Page Object Models
│   ├── base.page.ts
│   ├── login.page.ts
│   └── dashboard.page.ts
└── fixtures/
    └── test-data.ts

Step 3: Page Object Model

All locators live in page objects. Tests call methods only.

Step 4: Write E2E Tests

Simulate real user journeys (login, add to cart, checkout), not individual page checks.

Step 5: Flakiness Remediation

FLAKINESS DIAGNOSIS:
| Symptom              | Fix                          |
|----------------------|------------------------------|
| Element not found    | Use auto-waiting locators    |
| Timeout on nav       | waitForLoadState, bump limit |
| Stale element ref    | Re-query after action        |
| Test order deps      | Isolate data in beforeEach   |

ANTI-FLAKINESS RULES:
- Never use fixed sleep — use auto-waiting assertions
- Never depend on execution order — independent tests
- Never use CSS selectors — use data-testid, roles
- Always seed test data via API, not through the UI
- Always retry failed tests in CI (investigate repeats)
- Always record traces/screenshots on failure

THRESHOLDS:
  Pass rate target: 100% over 10 consecutive runs
  Max flaky tolerance: 0 (zero flaky tests in suite)
  Retry budget in CI: 2 retries per test max