Test Standards

Critical

• Import test and expect from fixtures/pom/test-options.ts, never from @playwright/test.

• Each test has exactly one tag from @smoke, @sanity, @regression, @e2e, @api, @destructive. Never @functional, never a tag array, never a tag on test.describe().

  // CORRECT
  test('should login', { tag: '@smoke' }, async ({ appPage }) => {});
  // WRONG
  test('should login', { tag: ['@smoke', '@e2e'] }, async () => {});

• @destructive is the heaviest tag and always wins, but only for shared/global state (locale, permissions, roles, guest access, feature flags, global settings, shared seed data). A test that creates and removes only its own data keeps its importance tag.
• Every state-mutating test has a test.afterEach() or test.afterAll() hook that reverts what it wrote, and the hook succeeds after a failed test.

• Web-first assertions only: await expect(locator).toBeVisible(), .toHaveText(), .toHaveCount(). Never page.waitForTimeout().

  // CORRECT
  await expect(appPage.username).toBeVisible();
  // WRONG
  await page.waitForTimeout(1000);

• Tests must be independent: no ordering, no shared mutable state. Setup goes in test.beforeEach() or a fixture.
• Never commit explore-only or debug specs (console.log(await page.content()), probes).
• After adding or modifying a spec, run npx playwright test <file> and confirm zero failures before reporting done.

Procedure

1. Classify: functional (tests/{area}/functional/), API (tests/{area}/api/), E2E (tests/{area}/e2e/) or setup (tests/{area}/*.setup.ts). Run ls tests/ to resolve {area}.
2. Open the nearest spec of the same type and copy its import, describe and beforeEach shape; a spec that must start signed out (login, anonymous forms) calls resetStorageState() in beforeEach; everything else starts signed in. Skeletons: references/spec-patterns.md.
3. Tag the test; when unsure which importance tag applies, read references/tags.md.
4. Split the body into test.step() blocks named GIVEN / WHEN / THEN. An API test with more than one call follows api-testing.
5. Take values from a factory or a static .ts file per data-strategy; a data-driven for...of loop sits inside describe, outside test().
6. If the test mutates state, add the cleanup hook per references/destructive-tests.md; API setup or teardown shared by several specs goes through api-testing.
7. Run npx playwright test <file> and npm run lint; on red, load debugging.

References

• references/spec-patterns.md: skeletons for a functional, data-driven, E2E and setup spec.
• references/tags.md: what each tag covers, how each suite runs, the three tagging mistakes.
• references/destructive-tests.md: the shared-state definition, the cleanup pattern, the destructive suite, suite-only failures.