gusgad/service-agent-by-sap / .agents/skills/playwright
playwright
VerifiedUse when testing or validating the client UI end-to-end - verifying a page renders, a form submits, navigation works, or checking visual/responsive behavior in a real browser. Use when the user asks to "test", "check", or "verify" something in the app that requires clicking through the UI, not just unit tests.
Playwright Browser Automation
Write and run a small, purpose-built Playwright script for the task at hand rather than reaching for a generic pre-built test. Most browser-verification tasks in this repo are one-off: does this page render, does this form submit, does this flow redirect correctly.
This Repo's Setup
client/— Angular app, served vianpm start(ng serve --host 0.0.0.0 --port 4200) →http://localhost:4200agent/— Express API the client talks to, vianpm run dev→http://localhost:3000- Neither package currently depends on Playwright. Install it once per machine as a devDependency before first use:
(Or install globally in a scratch directory if you'd rather not touchnpm install -D @playwright/test --prefix client npx --prefix client playwright install chromiumclient/package.jsonfor an ad-hoc check.)
Workflow
- Confirm the dev server is actually running before writing a script — don't assume port 4200 (and 3000, if the flow touches the API) is up. Start it if it isn't (
npm startinclient/,npm run devinagent/), or ask for the URL if testing something already deployed. - Write the script to a scratch path (this session's scratchpad directory, or
/tmp), never intoclient/— these are throwaway verification scripts, not part of the app. - Parameterize the URL as a constant at the top of the script so it's obvious what's being tested.
- Run with a visible browser by default (
headless: false) when a human is watching / debugging; useheadless: truefor a quick unattended check or when running from a non-interactive context. - Always wait on real signals (
waitForSelector,waitForURL,waitForLoadState('networkidle')), never a fixedwaitForTimeoutas the only guard — Angular's change detection and API round-trips make fixed sleeps flaky.
Minimal Script Template
// scratch/playwright-check.js
const { chromium } = require('playwright');
const TARGET_URL = 'http://localhost:4200'; // parameterize, don't hardcode inline below
(async () => {
const browser = await chromium.launch({ headless: false });
const page = await browser.newPage();
try {
await page.goto(TARGET_URL, { waitUntil: 'networkidle' });
console.log('Title:', await page.title());
// ... actions specific to the task ...
await page.screenshot({ path: 'scratch/screenshot.png', fullPage: true });
} finally {
await browser.close();
}
})();
Run with: node scratch/playwright-check.js
Common Patterns
Fill and submit a form (e.g. a job-request form in client/):
await page.goto(`${TARGET_URL}/requests/new`);
await page.fill('input[formcontrolname="name"]', 'Test Job');
await page.selectOption('select[formcontrolname="type"]', 'http');
await page.click('button[type="submit"]');
await page.waitForSelector('.success-message, [role="alert"]');
Check a list/status page renders data from the API:
await page.goto(`${TARGET_URL}/requests`);
await page.waitForLoadState('networkidle');
const rowCount = await page.locator('table tbody tr').count();
console.log('Rows rendered:', rowCount);
Responsive check across viewports:
const viewports = [
{ name: 'desktop', width: 1440, height: 900 },
{ name: 'tablet', width: 768, height: 1024 },
{ name: 'mobile', width: 375, height: 667 },
];
for (const v of viewports) {
await page.setViewportSize({ width: v.width, height: v.height });
await page.goto(TARGET_URL, { waitUntil: 'networkidle' });
await page.screenshot({ path: `scratch/${v.name}.png`, fullPage: true });
}
Multi-tenant check (this app is multi-tenant — verify tenant isolation when relevant):
// Set whatever header/localStorage/cookie this app uses to select tenant context
await page.goto(TARGET_URL);
await page.evaluate((tenantId) => localStorage.setItem('tenantId', tenantId), 'tenant-a');
await page.reload();
Confirm the actual mechanism (header vs. localStorage vs. route param) by reading how client/src sets tenant context before assuming this pattern.
Selectors
Prefer, in order: role= / accessible name, text=, a stable data-testid if present, Angular formcontrolname/component selectors, then CSS as a last resort. Avoid brittle nth-child selectors.
Tips
- Use
slowMo: 100when a human needs to actually watch the run happen - Always wrap the page interaction in
try/finallysobrowser.close()runs even on failure console.logprogress as you go — output is the only feedback while headless- Don't commit scratch scripts or screenshots to the repo
Troubleshooting
| Symptom | Likely cause |
|---|---|
browserType.launch: Executable doesn't exist | Run npx playwright install chromium |
| Element not found immediately after navigation | Angular hasn't finished rendering — add waitForSelector or waitForLoadState('networkidle') |
| Works locally, flaky in CI | Fixed timeouts instead of condition-based waits — replace waitForTimeout with a real wait condition |
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| finishing-a-development-branchobra/superpowersUse when implementation is complete, all tests pass, and you need to decide how to integrate the work | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| systematic-debuggingobra/superpowersUse when encountering any bug, test failure, or unexpected behavior, before proposing fixes | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| test-driven-developmentobra/superpowersUse when implementing any feature or bugfix, before writing implementation code | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| skill-creatoranthropics/skillsCreate new skills, modify and improve existing skills, and measure skill performance. Use when users want to create a skill from scratch, edit, or optimize an existing skill, run evals to test a skill, benchmark skill performance with variance analysis, or optimize a skill's description for better triggering accuracy. | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |