Anki Playwright Agents Practice Exam

Question 1 Anki Playwright Agents

What is a Locator?

await expect(locator).toBeVisible(); await locator.click(); A live selector handle that re-resolves automatically and supports auto-wait + strictness; prefer locators over element handles. Create projects like { name:'ams', use:{ baseURL:'http://127.0.0.1:3437' } } and { name:'connect', use:{ baseURL:'http://127.0.0.1:3445' } }. “Only write under `/e2e/` and do not modify app code unless I explicitly approve.”

Question 2 Anki Playwright Agents

How do I test a keyboard shortcut?

Update assertions/testids first, then update screenshots with review. Record the reason (ticket/PR) so future diffs are explainable. await page.keyboard.press('Control+K') (use 'Meta+K' on macOS if applicable). Add `test.describe.configure({ timeout: 120000 })` for slower flows. npm i -D @playwright/test
npx playwright install
npx playwright test

Question 3 Anki Playwright Agents

How do I test file downloads?

const download = await page.waitForEvent('download'); await download.saveAs('tmp/file.csv'); npx playwright test --workers=1 A recorded timeline (screenshots, DOM snapshots, network, console, actions) you can inspect to debug failing/flaky tests. const card = page.getByTestId('report-card'); await card.getByRole('button', { name: 'Download' }).click();

Question 4 Anki Playwright Agents

How to handle intentional UI changes?

Update assertions/testids first, then update screenshots with review. Record the reason (ticket/PR) so future diffs are explainable. await page.getByLabel('Email').fill('user@example.com') npx playwright test --last-failed Install browsers (npx playwright install --with-deps), set CI=1, run npx playwright test, upload playwright-report/ and test-results/ artifacts.

Question 5 Anki Playwright Agents

What’s a common pitfall with running two Vite apps?

await page.waitForLoadState('domcontentloaded') (usually unnecessary—use expect() on UI instead). Ask it to: (1) add/modify tests, (2) run npx playwright test, (3) fix failures, (4) re-run until green, and (5) summarize what changed. Both may default to the same port (e.g., 8080); run one with `npm run dev -- --port 3437` to avoid conflicts. npx playwright test --max-failures=1

Question 6 Anki Playwright Agents

What is test.use() for?

Run existing suite, identify stable selectors, confirm the user flow steps, and agree on what must be asserted (URL, text, data). It sets per-file/describe options like storageState, viewport, permissions, etc. npx playwright test --list Never update snapshots unless the user explicitly wants to accept visual changes; attach diff + reasoning when proposing updates.

Question 7 Anki Playwright Agents

How do I attach extra debug info to the report?

expect(page).toHaveScreenshot({ mask: [page.locator("[data-testid=clock]")] }) to ignore volatile regions. Require placeholders for credentials and env-var injection (e.g., USERNAME/PASSWORD) with no hardcoded values. Use testInfo.attach('name', { body: Buffer.from('...'), contentType: 'text/plain' }). It often uses brittle selectors and weak assertions; refactor to getByRole/testids and add meaningful expect() checks.

Question 8 Anki Playwright Agents

How do I instruct an agent to avoid secrets?

Use context.waitForEvent('page') and keep references to each page object. In config: use: { video: 'retain-on-failure' } (or 'on' for always). Prefer storageState; don’t depend on long-lived cookies that may expire unpredictably. Require placeholders for credentials and env-var injection (e.g., USERNAME/PASSWORD) with no hardcoded values.

Question 9 Anki Playwright Agents

How do I keep terminal output concise?

Use a simpler reporter like `--reporter=line` (or configure reporter in playwright.config.ts). npx playwright test --workers=1 Set reporter to include junit in config (e.g., reporter: [['junit', { outputFile: 'results.xml' }]]). Most locator actions expect exactly one match; if multiple elements match, Playwright throws to prevent ambiguous tests.

Question 10 Anki Playwright Agents

How do I guide an AI agent to create Playwright tests I can run later?

Usually under a Playwright-managed cache; override with PLAYWRIGHT_BROWSERS_PATH=0 to install into node_modules/.cache. Specify: repo path, testDir, baseURL(s), how to start servers, login method, required assertions, and “only write under e2e/”. “List the minimum elements needing data-testid and why; wait for approval before editing app code.” page.on('requestfailed', req => console.log('FAILED', req.url(), req.failure()?.errorText))

Question 11 Anki Playwright Agents

How do I grant permissions like clipboard/geolocation?

Set permissions in context/use config: permissions: ['clipboard-read', 'clipboard-write']. use: { extraHTTPHeaders: { 'X-Test-Run': 'e2e' } } await expect(page.getByTestId('row')).toHaveCount(10) npx playwright test --project=chromium

Question 12 Anki Playwright Agents

How do I assert an element count?

npx playwright test --workers=1 await expect(page.getByTestId('row')).toHaveCount(10) await expect(page.getByRole('checkbox', { name: 'Active' })).toBeChecked() Define the assertions: URL change, key text visible, API response OK, and a DB/data effect if applicable.

Question 13 Anki Playwright Agents

Anki import: how do I force Anki to treat content as HTML?

await page.getByLabel('Role').selectOption({ label: 'Admin' }) Use locators (not element handles), avoid storing DOM nodes, and re-locate before acting. testInfo.attach('api-response.json', { body: Buffer.from(JSON.stringify(obj,null,2)), contentType: 'application/json' }) Add `#html:true` at the top; then `
` renders line breaks and `` works.

What is a Locator?

How do I test a keyboard shortcut?

How do I test file downloads?

How to handle intentional UI changes?

What’s a common pitfall with running two Vite apps?

What is test.use() for?

How do I attach extra debug info to the report?

How do I instruct an agent to avoid secrets?

How do I keep terminal output concise?

How do I guide an AI agent to create Playwright tests I can run later?

How do I grant permissions like clipboard/geolocation?

How do I assert an element count?

Anki import: how do I force Anki to treat content as HTML?

How do I disable video/trace locally for speed?

Where do Playwright tests live?

How do I create a minimal smoke suite?

What’s the difference between actionTimeout and navigationTimeout?

How do I set Playwright baseURL when ports vary?

How do I work with virtualized lists?

How do I assert the URL query params?

What’s the difference between headless and headed?

How do I share code between tests?

How do I capture screenshots on failure automatically?

Anki import: what file format does Anki accept?

How do I handle “remember me” and cookies?

How do I make an agent “run tests automatically”?

How do I run setup steps before each test?

Parallel tests without chaos?

How can one E2E suite cover both AMS and Connect?

What is “strict mode” for locators?

How do I create separate sessions for different roles?

How do I run Playwright in Docker?

Mask dynamic areas in screenshots?

How do I reduce flakiness from animations?

Why prefer getByRole()?

How do I test drag-and-drop?

How do I set extra HTTP headers for all requests?

How do I select the first matching element?

How do I verify Playwright is installed correctly?

How do I make selectors stable in Vue/Laravel apps?

How do I run teardown after each test?

How do I wait for navigation after clicking?

How do I do per-role logins (admin vs user)?

How do I force a click when an overlay blocks it?

Why shouldn’t I keep raw codegen output?

How do I authenticate once and reuse it?

How do I skip or focus tests?

What if my app uses SSO/2FA?

How do I create a storageState file?

Agent safety rule for snapshots?