Skip to content

ATK Custom Tests

When to Use

Use this guide when writing project-specific tests on top of ATK's catalog.

Decision

Test scope Location
Pure ATK use case (login, generic content) Already in ATK catalog — don't duplicate
Project-specific content type or workflow e2e/content/ or e2e/workflows/
Project-specific helper helpers/project/
Modification of an ATK behavior Copy the ATK test into e2e/atk-overrides/, modify; don't edit upstream

Pattern

Directory structure

tests/playwright/
├── e2e/
│   ├── atk/                    # ATK's shipped tests (copied or symlinked)
│   ├── smoke/                  # Your project's smoke set
│   ├── content/                # Project-specific content tests
│   └── workflows/              # End-to-end workflow tests
├── helpers/
│   ├── atk/                    # ATK's helpers (copied or symlinked)
│   └── project/                # Your project's helpers
└── playwright.config.js

A custom test using ATK helpers

import { test, expect } from '@playwright/test';
import { loginAsRole } from '../helpers/atk';

test.describe('custom workflow', () => {
  test.beforeEach(async ({ page }) => {
    await loginAsRole(page, 'editor');
  });

  test('editor publishes a curated landing page', async ({ page }) => {
    await page.goto('/node/add/landing_page');
    await page.fill('[data-qa-id="landing-title"]', 'Spring 2026 Campaign');
    await page.click('[data-qa-id="landing-publish"]');
    await expect(page.locator('.messages--status')).toContainText('saved');
  });
});

For Playwright fixtures, web-first assertions, locator strategies, and DDEV specifics, see Playwright for Visual Regression — the same knowledge applies to ATK functional tests, just without the toHaveScreenshot() step.

Test tagging for selective runs

// Playwright
test('login works @smoke @auth', async ({ page }) => { /* ... */ });

npx playwright test --grep "@smoke"

// Cypress
describe('login', { tags: ['@smoke', '@auth'] }, () => { /* ... */ });

npx cypress run --env grepTags="@smoke"

Common Mistakes

  • Wrong: Modifying ATK's shipped files in-place → Right: lost on next upgrade; copy and rename instead
  • Wrong: No tagging strategy → Right: can't run a smoke subset without listing every test path
  • Wrong: Helpers that only work in one runner → Right: write helpers shared across both when possible

See Also