Web Application Testing

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. These scripts can be very large and thus pollute your context window. They exist to be called directly as black-box scripts rather than ingested into your context window.

Browser Tools (Direct Chrome DevTools Control)

The bin/browser-tools utility provides lightweight, context-rot-proof browser automation using the Chrome DevTools Protocol directly (no MCP overhead).

Available Commands:

# Launch browser and get connection details
bin/browser-tools start [--port PORT] [--headless]

# Navigate to URL
bin/browser-tools nav <url> [--wait-for {load|networkidle|domcontentloaded}]

# Evaluate JavaScript
bin/browser-tools eval "<javascript code>"

# Take screenshot
bin/browser-tools screenshot <output.png> [--full-page] [--selector CSS_SELECTOR]

# Interactive element picker (returns selectors)
bin/browser-tools pick

# Get console logs
bin/browser-tools console [--level {log|warn|error|all}]

# Search page content
bin/browser-tools search "<query>" [--case-sensitive]

# Extract page content (markdown, links, text)
bin/browser-tools content [--format {markdown|links|text}]

# Get/set cookies
bin/browser-tools cookies [--set NAME=VALUE] [--domain DOMAIN]

# Inspect element details
bin/browser-tools inspect <css-selector>

# Terminate browser session
bin/browser-tools kill

When to Use Browser-Tools vs Playwright:

✅ Use browser-tools when:

• Quick inspection/debugging of running apps
• Need to identify selectors interactively (pick command)
• Extracting page content for analysis
• Running simple JavaScript snippets
• Monitoring console logs in real-time
• Taking quick screenshots without writing scripts

✅ Use Playwright when:

• Complex multi-step automation workflows
• Need full test assertion framework
• Handling multi-step forms
• Database integration (Supabase utilities)
• Comprehensive test coverage
• Generating test reports

Example Workflow:

# 1. Start browser pointed at your app
bin/browser-tools start --port 9222

# 2. Navigate to the page
bin/browser-tools nav http://localhost:3000

# 3. Use interactive picker to find selectors
bin/browser-tools pick
# Click on elements in the browser, get their selectors

# 4. Inspect specific elements
bin/browser-tools inspect "button.submit"

# 5. Take screenshot for documentation
bin/browser-tools screenshot /tmp/page.png --full-page

# 6. Check console for errors
bin/browser-tools console --level error

# 7. Clean up
bin/browser-tools kill

Benefits:

• No MCP overhead: Direct Chrome DevTools Protocol communication
• Context-rot proof: No dependency on evolving MCP specifications
• Interactive picker: Visual element selection for selector discovery
• Lightweight: Faster than full Playwright for simple tasks
• Pre-compiled binary: Ready to use, no compilation needed (auto-compiled via create-rule.js)

Decision Tree: Choosing Your Approach

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

Example: Using with_server.py

To start a server, run --help first, then use the helper:

Single server:

python scripts/with_server.py --server "npm run dev" --port 3000 -- python your_automation.py

Multiple servers (e.g., backend + frontend):

python scripts/with_server.py \
  --server "cd backend && python server.py" --port 8000 \
  --server "cd frontend && npm run dev" --port 3000 \
  -- python your_automation.py

To create an automation script, include only Playwright logic (servers are managed automatically):

from playwright.sync_api import sync_playwright

APP_PORT = 3000  # Match the port from --port argument

with sync_playwright() as p:
    browser = p.chromium.launch(headless=True) # Always launch chromium in headless mode
    page = browser.new_page()
    page.goto(f'http://localhost:{APP_PORT}') # Server already running and ready
    page.wait_for_load_state('networkidle') # CRITICAL: Wait for JS to execute
    # ... your automation logic
    browser.close()

Reconnaissance-Then-Action Pattern

1. Inspect rendered DOM:

   page.screenshot(path='/tmp/inspect.png', full_page=True)
   content = page.content()
   page.locator('button').all()
   

2. Identify selectors from inspection results

3. 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 - To accomplish a task, consider whether one of the scripts available in scripts/ can help. These scripts handle common, complex workflows reliably without cluttering the context window. Use --help to see usage, then 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()

Utility Modules

The skill now includes comprehensive utilities for common testing patterns:

UI Interactions (utils/ui_interactions.py)

Handle common UI patterns automatically:

from utils.ui_interactions import (
    dismiss_cookie_banner,
    dismiss_modal,
    click_with_header_offset,
    force_click_if_needed,
    wait_for_no_overlay,
    wait_for_stable_dom
)

# Dismiss cookie consent
dismiss_cookie_banner(page)

# Close welcome modal
dismiss_modal(page, modal_identifier="Welcome")

# Click button behind fixed header
click_with_header_offset(page, 'button#submit', header_height=100)

# Try click with force fallback
force_click_if_needed(page, 'button#action')

# Wait for loading overlays to disappear
wait_for_no_overlay(page)

# Wait for DOM to stabilize
wait_for_stable_dom(page)

Smart Form Filling (utils/form_helpers.py)

Intelligently handle form variations:

from utils.form_helpers import (
    SmartFormFiller,
    handle_multi_step_form,
    auto_fill_form
)

# Works with both "Full Name" and "First/Last Name" fields
filler = SmartFormFiller()
filler.fill_name_field(page, "Jane Doe")
filler.fill_email_field(page, "jane@example.com")
filler.fill_password_fields(page, "SecurePass123!")
filler.fill_phone_field(page, "+447700900123")
filler.fill_date_field(page, "1990-01-15", field_hint="birth")

# Auto-fill entire form
results = auto_fill_form(page, {
    'email': 'test@example.com',
    'password': 'Pass123!',
    'full_name': 'Test User',
    'phone': '+447700900123',
    'date_of_birth': '1990-01-15'
})

# Handle multi-step forms
steps = [
    {'fields': {'email': 'test@example.com', 'password': 'Pass123!'}, 'checkbox': True},
    {'fields': {'full_name': 'Test User', 'date_of_birth': '1990-01-15'}},
    {'complete': True}
]
handle_multi_step_form(page, steps)

Supabase Testing (utils/supabase.py)

Database operations for Supabase-based apps:

from utils.supabase import SupabaseTestClient, quick_cleanup

# Initialize client
client = SupabaseTestClient(
    url="https://project.supabase.co",
    service_key="your-service-role-key",
    db_password="your-db-password"
)

# Create test user
user_id = client.create_user("test@example.com", "password123")

# Create invite code
client.create_invite_code("TEST2024", code_type="general")

# Bypass email verification
client.confirm_email(user_id)

# Cleanup after test
client.cleanup_related_records(user_id)
client.delete_user(user_id)

# Quick cleanup helper
quick_cleanup("test@example.com", "db_password", "https://project.supabase.co")

Advanced Wait Strategies (utils/wait_strategies.py)

Better alternatives to simple sleep():

from utils.wait_strategies import (
    wait_for_api_call,
    wait_for_element_stable,
    smart_navigation_wait,
    combined_wait
)

# Wait for specific API response
response = wait_for_api_call(page, '**/api/profile**')

# Wait for element to stop moving
wait_for_element_stable(page, '.dropdown-menu', stability_ms=1000)

# Smart navigation with URL check
page.click('button#login')
smart_navigation_wait(page, expected_url_pattern='**/dashboard**')

# Comprehensive wait (network + DOM + overlays)
combined_wait(page)

Smart Selectors (utils/smart_selectors.py) ⭐ NEW

Automatically try multiple selector strategies to find elements, reducing test brittleness:

from utils.smart_selectors import SelectorStrategies

# Find and fill email field (tries 7 different selector strategies)
success = SelectorStrategies.smart_fill(page, 'email', 'test@example.com')
# Output: ✓ Found field via placeholder: input[placeholder*="email" i]
#         ✓ Filled 'email' with value

# Find and click button (tries 8 different strategies)
success = SelectorStrategies.smart_click(page, 'Sign In')
# Output: ✓ Found button via case-insensitive text: button:text-matches("Sign In", "i")
#         ✓ Clicked 'Sign In' button

# Manual control - find input field selector
selector = SelectorStrategies.find_input_field(page, 'password')
if selector:
    page.fill(selector, 'my-password')

# Manual control - find button selector
selector = SelectorStrategies.find_button(page, 'Submit')
if selector:
    page.click(selector)

# Try custom selectors with fallback
selectors = ['button#submit', 'button.submit-btn', 'input[type="submit"]']
selector = SelectorStrategies.find_any_element(page, selectors)

Selector Strategies Tried (in order):

For input fields:

1. Test IDs: [data-testid*="email"]
2. ARIA labels: input[aria-label*="email"]
3. Placeholder: input[placeholder*="email"]
4. Name attribute: input[name*="email"]
5. Type attribute: input[type="email"]
6. ID exact: #email
7. ID partial: input[id*="email"]

For buttons:

1. Test IDs: [data-testid*="sign-in"]
2. Name attribute: button[name*="Sign In"]
3. Exact text: button:has-text("Sign In")
4. Case-insensitive: button:text-matches("Sign In", "i")
5. Link as button: a:has-text("Sign In")
6. Submit input: input[type="submit"][value*="Sign In"]
7. Role button: [role="button"]:has-text("Sign In")

When to use:

• ✅ Forms where HTML structure changes frequently
• ✅ Third-party components with unpredictable selectors
• ✅ Multi-application test suites
• ❌ Performance-critical tests (adds 5s timeout per strategy)

Performance:

• Timeout per strategy: 5 seconds (configurable)
• Max timeout: 10 seconds across all strategies
• Typical: First strategy works (1-2 seconds)

Browser Configuration (utils/browser_config.py) ⭐ NEW

Auto-configure browser context for testing environments:

from utils.browser_config import BrowserConfig

# Auto-detect CSP bypass for localhost
context = BrowserConfig.create_test_context(
    browser,
    'http://localhost:3000'
)
# Output:
# ============================================================
#   Browser Context Configuration
# ============================================================
#   Base URL: http://localhost:3000
#   🔓 CSP bypass: ENABLED (testing on localhost)
#   ⚠️  HTTPS errors: IGNORED (self-signed certs OK)
#   📐 Viewport: 1280x720
# ============================================================

# Production testing (no CSP bypass)
context = BrowserConfig.create_test_context(
    browser,
    'https://production.example.com'
)
# Output: 🔒 CSP bypass: DISABLED (production mode)

# Mobile device emulation
context = BrowserConfig.create_mobile_context(
    browser,
    device='iPhone 12',
    base_url='http://localhost:3000'
)
# Output: 📱 Mobile context: iPhone 12
#         Viewport: {'width': 390, 'height': 844}
#         🔓 CSP bypass: ENABLED

# Manual override
context = BrowserConfig.create_test_context(
    browser,
    'http://localhost:3000',
    bypass_csp=False,  # Force disable even for localhost
    record_video=True,  # Record test session
    extra_http_headers={'Authorization': 'Bearer token'}
)

Features:

• Auto CSP detection: Enables bypass for localhost, disables for production
• Self-signed certs: Ignores HTTPS errors by default
• Consistent viewport: 1280x720 default for reproducible tests
• Mobile emulation: Built-in device profiles
• Video recording: Optional session recording
• Verbose logging: See exactly what's configured

When to use:

• ✅ Every test (replaces manual browser.new_context())
• ✅ Testing on localhost with CSP restrictions
• ✅ Mobile-responsive testing
• ✅ Need consistent browser configuration across tests

CSP Monitoring (utils/ui_interactions.py) ⭐ NEW

Auto-detect and suggest fixes for Content Security Policy violations:

from utils.ui_interactions import setup_page_with_csp_handling
from utils.browser_config import BrowserConfig

context = BrowserConfig.create_test_context(browser, 'http://localhost:7160')
page = context.new_page()

# Enable CSP violation monitoring
setup_page_with_csp_handling(page)

page.goto('http://localhost:7160')

# If CSP violation occurs, you'll see:
# ======================================================================
# ⚠️  CSP VIOLATION DETECTED
# ======================================================================
# Message: Refused to execute inline script because it violates...
#
# 💡 SUGGESTION:
#    For localhost testing, use:
#
#    from utils.browser_config import BrowserConfig
#    context = BrowserConfig.create_test_context(
#        browser, 'http://localhost:3000'
#    )
#    # Auto-enables CSP bypass for localhost
#
#    Or manually:
#    context = browser.new_context(bypass_csp=True)
# ======================================================================

When to use:

• ✅ Debugging why tests fail on localhost
• ✅ Identifying CSP configuration issues
• ✅ Verifying CSP bypass is working
• ❌ Production testing (CSP should be enforced)

Complete Examples

Multi-Step Registration

See examples/multi_step_registration.py for a complete example showing:

• Database setup (invite codes)
• Cookie banner dismissal
• Multi-step form automation
• Email verification bypass
• Login flow
• Dashboard verification
• Cleanup

Run it:

python examples/multi_step_registration.py

Using the Webapp-Testing Subagent

A specialized subagent is available for testing automation. Use it to keep your main conversation focused on development:

You: "Use webapp-testing agent to register test@example.com and verify the parent role switch works"

Main Agent: [Launches webapp-testing subagent]

Webapp-Testing Agent: [Runs complete automation, returns results]

Benefits:

• Keeps main context clean
• Specialized for Playwright automation
• Access to all skill utilities
• Automatic screenshot capture
• Clear result reporting

Reference Files

• examples/ - Examples showing common patterns:

  • element_discovery.py - Discovering buttons, links, and inputs on a page
  • static_html_automation.py - Using file:// URLs for local HTML
  • console_logging.py - Capturing console logs during automation
  • multi_step_registration.py - Complete registration flow example (NEW)

• utils/ - Reusable utility modules (NEW):

  • ui_interactions.py - Cookie banners, modals, overlays, stable waits
  • form_helpers.py - Smart form filling, multi-step automation
  • supabase.py - Database operations for Supabase apps
  • wait_strategies.py - Advanced waiting patterns