| name | webapp-testing |
| description | Toolkit for interacting with and testing local web applications using Playwright. Supports verifying frontend functionality, debugging UI behavior, capturing browser screenshots, and viewing browser logs. Includes browser automation safety rules: rate limiting, robots.txt compliance, session handling, and PII protection.
|
| risk | medium |
| tags | ["testing","browser"] |
| created | 2026-03-07T00:00:00.000Z |
| updated | 2026-03-17T00:00:00.000Z |
Web Application Testing
Playwright-based web application testing toolkit with built-in browser automation safety rules.
How to use
When to use
Reference these guidelines when:
- testing local web applications with Playwright
- verifying frontend functionality or debugging UI behavior
- capturing browser screenshots or viewing browser logs
- building web scrapers or data extractors
- automating browser interactions (Playwright, Puppeteer, Selenium)
- fetching content from external websites
- handling extracted data that may contain PII
When NOT to use
Do NOT apply this skill when:
- the task is about unit testing (use
testing-methodology skill)
- the task is about API testing without a browser (use
testing-methodology skill)
- the task is about UI component design (use
baseline-ui or frontend-design skill)
Prefer expect-cli for Adversarial Browser Testing
For validating browser-facing code changes, prefer expect-cli over raw Playwright scripts:
EXPECT_BASE_URL=http://localhost:3000 expect-cli -m "describe what to test adversarially" -y
expect-cli (from millionco/expect) scans your git diff, generates an adversarial test plan, and executes it in a real browser with session recordings. Use it after any frontend change. See the expect skill for full documentation.
Fall back to raw Playwright scripts (below) when you need fine-grained control, custom fixtures, or integration with existing test suites.
Testing Toolkit
To test local web applications, write native Python Playwright scripts.
Helper Scripts Available:
scripts/with_server.py - Manages server lifecycle (supports multiple servers)
Always run scripts with --help first to see usage. DO NOT read the source until you try running the script first and find that a customized solution is absolutely necessary.
Decision Tree
User task → Is it static HTML?
├─ Yes → Read HTML file directly to identify selectors
│ ├─ Success → Write Playwright script using selectors
│ └─ Fails/Incomplete → Treat as dynamic (below)
│
└─ No (dynamic webapp) → Is the server already running?
├─ No → Run: python scripts/with_server.py --help
│ Then use the helper + write simplified Playwright script
│
└─ Yes → Reconnaissance-then-action:
1. Navigate and wait for networkidle
2. Take screenshot or inspect DOM
3. Identify selectors from rendered state
4. Execute actions with discovered selectors
Server Management
Single server:
python scripts/with_server.py --server "npm run dev" --port 5173 -- python your_automation.py
Multiple servers (backend + frontend):
python scripts/with_server.py \
--server "cd backend && python server.py" --port 3000 \
--server "cd frontend && npm run dev" --port 5173 \
-- python your_automation.py
Automation Script Pattern
from playwright.sync_api import sync_playwright
with sync_playwright() as p:
browser = p.chromium.launch(headless=True)
page = browser.new_page()
page.goto('http://localhost:5173')
page.wait_for_load_state('networkidle')
browser.close()
Reconnaissance-Then-Action
- Inspect rendered DOM:
page.screenshot(path='/tmp/inspect.png', full_page=True)
content = page.content()
page.locator('button').all()
- Identify selectors from inspection results
- Execute actions using discovered selectors
Common Pitfall
- Don't inspect the DOM before waiting for
networkidle on dynamic apps
- Do wait for
page.wait_for_load_state('networkidle') before inspection
Best Practices
- Use bundled scripts as black boxes —
--help first, invoke directly
- Use
sync_playwright() for synchronous scripts
- Always close the browser when done
- Use descriptive selectors:
text=, role=, CSS selectors, or IDs
- Add appropriate waits:
page.wait_for_selector() or page.wait_for_timeout()
Reference Files
- examples/element_discovery.py - Discovering buttons, links, inputs
- examples/static_html_automation.py - Using file:// URLs for local HTML
- examples/console_logging.py - Capturing console logs
Browser Automation Safety
Rules for safe, ethical browser automation and web data extraction.
Priority Levels
| Priority | Category | Impact |
|---|
| 1 | Legal and ethical | Critical |
| 2 | Rate limiting | Critical |
| 3 | Session safety | High |
| 4 | Data handling | High |
| 5 | Error resilience | Medium |
1. Legal and Ethical (Critical)
- MUST check and respect
robots.txt before scraping
- MUST respect
Crawl-delay directives
- MUST check Terms of Service for scraping restrictions
- NEVER scrape login-protected content without authorization
- NEVER circumvent anti-bot measures on sites that prohibit scraping
- MUST set a descriptive User-Agent string identifying your bot
- SHOULD provide a contact email in User-Agent or on a linked page
2. Rate Limiting (Critical)
- MUST use minimum 1-second delay between requests to the same domain
- MUST respect
Retry-After headers; implement exponential backoff
- MUST use concurrent request limits (max 2-3 per domain)
- SHOULD monitor response times; slow down if server is struggling (5xx responses)
- MUST implement circuit breaker: stop after N consecutive failures
- SHOULD use caching to avoid re-fetching unchanged content
3. Session Safety (High)
- NEVER store or log authentication cookies from external sites
- MUST clear sessions after use; do not persist browser state unnecessarily
- SHOULD use incognito/private browser contexts for isolation
- MUST rotate proxies only when authorized and necessary
- MUST handle CAPTCHAs gracefully (stop, don't bypass)
- MUST implement timeouts for all network operations
4. Data Handling (High)
- MUST identify and redact PII (emails, phone numbers, addresses) before storage
- NEVER store scraped data longer than necessary
- SHOULD log what was accessed but not the full content (for audit)
- MUST sanitize extracted content before processing (prevent injection)
- MUST respect copyright; do not redistribute copyrighted content
- SHOULD store data with access controls; encrypt at rest if sensitive
5. Error Resilience (Medium)
- MUST handle network errors, timeouts, and partial page loads gracefully
- MUST implement retry with exponential backoff (max 3 retries)
- SHOULD validate extracted data shape before processing
- SHOULD log errors with context for debugging
- SHOULD use health checks to verify automation is working correctly
- SHOULD set up monitoring for long-running automation tasks
Cross-references
- testing-methodology skill: Test strategy, unit/integration testing
- baseline-ui skill: UI constraints, responsive design patterns
- security-review skill: Security audit patterns
- SECURITY_PLAYBOOK.md: Broader security rules