Initial commit

This commit is contained in:
Zhongwei Li
2025-11-30 09:03:44 +08:00
commit 51eac46940
41 changed files with 5371 additions and 0 deletions

View File

@@ -0,0 +1,737 @@
# Advanced Examples
Complex workflows and advanced usage patterns for chrome-devtools.
## Table of Contents
- [E2E Testing Workflows](#e2e-testing-workflows)
- [Web Scraping Patterns](#web-scraping-patterns)
- [Performance Optimization](#performance-optimization)
- [CI/CD Integration](#cicd-integration)
- [Cross-Browser Testing Simulation](#cross-browser-testing-simulation)
- [Advanced JavaScript Extraction](#advanced-javascript-extraction)
---
## E2E Testing Workflows
### Complete User Registration Flow
Test full registration process with validation:
```bash
#!/bin/bash
# test_registration.sh
echo "Starting registration flow test..."
# 1. Open registration page
node scripts/new_page.js --url https://example.com/register
# 2. Wait for page load
node scripts/wait_for.js --text "Create Account" --timeout 10000
# 3. Get form structure
node scripts/take_snapshot.js --filePath registration_form.txt
# 4. Fill registration form (update UIDs from snapshot)
node scripts/fill_form.js --elements '[
{"uid":"input_firstname","value":"John"},
{"uid":"input_lastname","value":"Doe"},
{"uid":"input_email","value":"john.doe@example.com"},
{"uid":"input_password","value":"SecurePass123!"},
{"uid":"input_password_confirm","value":"SecurePass123!"},
{"uid":"select_country","value":"United States"}
]'
# 5. Screenshot before submission
node scripts/take_screenshot.js --filePath before_submit.png
# 6. Accept terms checkbox
node scripts/click.js --uid checkbox_terms
# 7. Submit form
node scripts/click.js --uid button_register
# 8. Wait for success message
if node scripts/wait_for.js --text "Registration successful" --timeout 15000; then
echo "✓ Registration succeeded"
# 9. Capture success state
node scripts/take_screenshot.js --filePath registration_success.png
node scripts/take_snapshot.js --filePath success_page.txt
# 10. Check for welcome email notification
node scripts/wait_for.js --text "Verification email sent" --timeout 5000
echo "✓ Test passed: Registration flow complete"
exit 0
else
echo "✗ Registration failed"
# Capture error state
node scripts/take_screenshot.js --filePath registration_error.png
node scripts/list_console_messages.js --types error > console_errors.txt
echo "✗ Test failed: See error screenshots and logs"
exit 1
fi
```
### Shopping Cart Checkout Flow
Test complete e-commerce checkout:
```bash
#!/bin/bash
# test_checkout.sh
echo "Testing checkout flow..."
# 1. Navigate to product page
node scripts/new_page.js --url https://example.com/products/widget-123
# 2. Wait and verify product loads
node scripts/wait_for.js --text "Add to Cart" --timeout 10000
node scripts/take_snapshot.js > product_page.txt
# 3. Add to cart
node scripts/click.js --uid button_add_to_cart
node scripts/wait_for.js --text "Added to cart" --timeout 5000
# 4. Go to cart
node scripts/click.js --uid link_view_cart
node scripts/wait_for.js --text "Shopping Cart" --timeout 5000
# 5. Verify cart contents
node scripts/take_snapshot.js --filePath cart_contents.txt
node scripts/evaluate_script.js --function "() => document.querySelector('.cart-total').textContent" > cart_total.txt
# 6. Proceed to checkout
node scripts/click.js --uid button_checkout
node scripts/wait_for.js --text "Shipping Information" --timeout 10000
# 7. Fill shipping info
node scripts/fill_form.js --elements '[
{"uid":"input_address","value":"123 Main Street"},
{"uid":"input_city","value":"San Francisco"},
{"uid":"input_state","value":"CA"},
{"uid":"input_zip","value":"94102"}
]'
# 8. Continue to payment
node scripts/click.js --uid button_continue
node scripts/wait_for.js --text "Payment Method" --timeout 10000
# 9. Select payment method
node scripts/click.js --uid radio_credit_card
# 10. Fill payment info (test mode)
node scripts/fill_form.js --elements '[
{"uid":"input_card_number","value":"4242424242424242"},
{"uid":"input_card_expiry","value":"12/25"},
{"uid":"input_card_cvc","value":"123"}
]'
# 11. Place order
node scripts/click.js --uid button_place_order
node scripts/wait_for.js --text "Order Confirmed" --timeout 20000
# 12. Capture confirmation
node scripts/take_screenshot.js --filePath order_confirmation.png
node scripts/evaluate_script.js --function "() => ({orderNumber: document.querySelector('.order-number').textContent, total: document.querySelector('.order-total').textContent})" > order_details.json
echo "✓ Checkout flow completed successfully"
```
---
## Web Scraping Patterns
### Multi-Page Data Extraction
Scrape data across paginated results:
```bash
#!/bin/bash
# scrape_listings.sh
BASE_URL="https://example.com/listings"
OUTPUT_FILE="all_listings.json"
MAX_PAGES=10
echo "[" > $OUTPUT_FILE
for page in $(seq 1 $MAX_PAGES); do
echo "Scraping page $page..."
# Navigate to page
if [ $page -eq 1 ]; then
node scripts/new_page.js --url "$BASE_URL"
else
node scripts/navigate_page.js --url "$BASE_URL?page=$page"
fi
# Wait for listings to load
node scripts/wait_for.js --text "listings found" --timeout 10000
# Extract listings data
LISTINGS=$(node scripts/evaluate_script.js --function "() => Array.from(document.querySelectorAll('.listing')).map(item => ({
title: item.querySelector('.title').textContent.trim(),
price: item.querySelector('.price').textContent.trim(),
location: item.querySelector('.location').textContent.trim(),
url: item.querySelector('a').href
}))")
# Append to output (remove outer brackets)
echo "$LISTINGS" | jq '.[]' >> $OUTPUT_FILE
# Check if there's a next page
HAS_NEXT=$(node scripts/evaluate_script.js --function "() => document.querySelector('.pagination .next') !== null")
if [ "$HAS_NEXT" != "true" ]; then
echo "Reached last page at page $page"
break
fi
# Add comma separator
if [ $page -lt $MAX_PAGES ]; then
echo "," >> $OUTPUT_FILE
fi
# Be polite: wait between requests
sleep 2
done
echo "]" >> $OUTPUT_FILE
echo "✓ Scraped listings saved to $OUTPUT_FILE"
# Summary statistics
TOTAL=$(cat $OUTPUT_FILE | jq 'length')
echo "Total listings extracted: $TOTAL"
```
### Dynamic Content Scraping with Infinite Scroll
Extract data from infinite scroll pages:
```bash
#!/bin/bash
# scrape_infinite_scroll.sh
echo "Scraping infinite scroll content..."
# Open page
node scripts/new_page.js --url https://example.com/feed
# Wait for initial content
node scripts/wait_for.js --text "Post" --timeout 10000
# Initialize
PREVIOUS_COUNT=0
SCROLL_ATTEMPTS=0
MAX_SCROLLS=20
OUTPUT_FILE="feed_items.json"
while [ $SCROLL_ATTEMPTS -lt $MAX_SCROLLS ]; do
# Count current items
CURRENT_COUNT=$(node scripts/evaluate_script.js --function "() => document.querySelectorAll('.feed-item').length")
echo "Scroll $SCROLL_ATTEMPTS: Found $CURRENT_COUNT items"
# Check if we got new items
if [ "$CURRENT_COUNT" -eq "$PREVIOUS_COUNT" ]; then
echo "No new items loaded, reached end"
break
fi
PREVIOUS_COUNT=$CURRENT_COUNT
# Scroll to bottom
node scripts/evaluate_script.js --function "() => window.scrollTo(0, document.body.scrollHeight)"
# Wait for new content to load
sleep 2
SCROLL_ATTEMPTS=$((SCROLL_ATTEMPTS + 1))
done
# Extract all items
echo "Extracting all $CURRENT_COUNT items..."
node scripts/evaluate_script.js --function "() => Array.from(document.querySelectorAll('.feed-item')).map(item => ({
author: item.querySelector('.author').textContent.trim(),
content: item.querySelector('.content').textContent.trim(),
timestamp: item.querySelector('.timestamp').textContent.trim(),
likes: item.querySelector('.likes').textContent.trim()
}))" > $OUTPUT_FILE
echo "✓ Extracted $CURRENT_COUNT items to $OUTPUT_FILE"
```
---
## Performance Optimization
### Automated Performance Regression Testing
Compare performance across builds:
```bash
#!/bin/bash
# performance_regression_test.sh
BASELINE_URL="https://staging.example.com"
CURRENT_URL="https://production.example.com"
echo "Running performance regression tests..."
# Test baseline (staging)
echo "Testing baseline: $BASELINE_URL"
node scripts/new_page.js --url "$BASELINE_URL"
node scripts/performance_start_trace.js --reload true --autoStop false
node scripts/wait_for.js --text "Ready" --timeout 30000
node scripts/performance_stop_trace.js > baseline_trace.json
BASELINE_LCP=$(cat baseline_trace.json | jq -r '.metrics.LCP.value')
BASELINE_FID=$(cat baseline_trace.json | jq -r '.metrics.FID.value')
BASELINE_CLS=$(cat baseline_trace.json | jq -r '.metrics.CLS.value')
echo "Baseline: LCP=$BASELINE_LCP, FID=$BASELINE_FID, CLS=$BASELINE_CLS"
# Test current (production)
echo "Testing current: $CURRENT_URL"
node scripts/navigate_page.js --url "$CURRENT_URL"
node scripts/performance_start_trace.js --reload true --autoStop false
node scripts/wait_for.js --text "Ready" --timeout 30000
node scripts/performance_stop_trace.js > current_trace.json
CURRENT_LCP=$(cat current_trace.json | jq -r '.metrics.LCP.value')
CURRENT_FID=$(cat current_trace.json | jq -r '.metrics.FID.value')
CURRENT_CLS=$(cat current_trace.json | jq -r '.metrics.CLS.value')
echo "Current: LCP=$CURRENT_LCP, FID=$CURRENT_FID, CLS=$CURRENT_CLS"
# Compare (allow 10% regression threshold)
THRESHOLD=1.1
FAILED=false
LCP_RATIO=$(echo "scale=2; $CURRENT_LCP / $BASELINE_LCP" | bc)
if (( $(echo "$LCP_RATIO > $THRESHOLD" | bc -l) )); then
echo "✗ LCP regression detected: ${CURRENT_LCP}ms vs ${BASELINE_LCP}ms"
FAILED=true
else
echo "✓ LCP within threshold"
fi
FID_RATIO=$(echo "scale=2; $CURRENT_FID / $BASELINE_FID" | bc)
if (( $(echo "$FID_RATIO > $THRESHOLD" | bc -l) )); then
echo "✗ FID regression detected: ${CURRENT_FID}ms vs ${BASELINE_FID}ms"
FAILED=true
else
echo "✓ FID within threshold"
fi
CLS_RATIO=$(echo "scale=2; $CURRENT_CLS / $BASELINE_CLS" | bc)
if (( $(echo "$CLS_RATIO > $THRESHOLD" | bc -l) )); then
echo "✗ CLS regression detected: ${CURRENT_CLS} vs ${BASELINE_CLS}"
FAILED=true
else
echo "✓ CLS within threshold"
fi
if [ "$FAILED" = true ]; then
echo "Performance regression test FAILED"
exit 1
else
echo "Performance regression test PASSED"
exit 0
fi
```
### Network Performance Testing Matrix
Test performance across network conditions:
```bash
#!/bin/bash
# network_performance_matrix.sh
URL="https://example.com"
OUTPUT_DIR="performance_results"
mkdir -p $OUTPUT_DIR
# Define network profiles
declare -A NETWORKS
NETWORKS["fast"]='{"downloadThroughput":10000000,"uploadThroughput":10000000,"latency":10}'
NETWORKS["4g"]='{"downloadThroughput":1600000,"uploadThroughput":750000,"latency":150}'
NETWORKS["3g"]='{"downloadThroughput":180000,"uploadThroughput":84000,"latency":562}'
NETWORKS["slow-3g"]='{"downloadThroughput":50000,"uploadThroughput":50000,"latency":2000}'
echo "Testing performance across network conditions..."
# Open page once
node scripts/new_page.js --url "$URL"
for profile in "${!NETWORKS[@]}"; do
echo "Testing network profile: $profile"
# Apply network emulation
node scripts/emulate.js --networkConditions "${NETWORKS[$profile]}"
# Run performance trace
node scripts/performance_start_trace.js --reload true --autoStop false
node scripts/wait_for.js --text "Ready" --timeout 60000
node scripts/performance_stop_trace.js > "$OUTPUT_DIR/${profile}_trace.json"
# Extract key metrics
LCP=$(cat "$OUTPUT_DIR/${profile}_trace.json" | jq -r '.metrics.LCP.value')
FID=$(cat "$OUTPUT_DIR/${profile}_trace.json" | jq -r '.metrics.FID.value')
CLS=$(cat "$OUTPUT_DIR/${profile}_trace.json" | jq -r '.metrics.CLS.value')
echo "$profile: LCP=${LCP}ms, FID=${FID}ms, CLS=$CLS"
# Capture screenshot
node scripts/take_screenshot.js --filePath "$OUTPUT_DIR/${profile}_loaded.png"
done
# Reset to normal
node scripts/emulate.js --cpuThrottlingRate 1
echo "✓ Performance testing complete. Results in $OUTPUT_DIR/"
```
---
## CI/CD Integration
### GitHub Actions Workflow
```yaml
# .github/workflows/e2e-tests.yml
name: E2E Tests
on:
push:
branches: [ main, develop ]
pull_request:
branches: [ main ]
jobs:
e2e-tests:
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v3
- name: Setup Node.js
uses: actions/setup-node@v3
with:
node-version: '18'
- name: Install mcp2skill-tools
run: |
npm install -g mcp2skill-tools
- name: Start mcp2rest
run: |
mcp2rest start
mcp2rest add chrome-devtools
- name: Install skill dependencies
run: |
cd ~/.claude/skills/mcp-chrome-devtools/scripts
npm install
- name: Run E2E tests
run: |
cd tests
bash test_registration.sh
bash test_checkout.sh
bash test_search.sh
- name: Upload test artifacts
if: always()
uses: actions/upload-artifact@v3
with:
name: test-results
path: |
**/*.png
**/*.txt
**/*.json
- name: Stop mcp2rest
if: always()
run: mcp2rest stop
```
### Jenkins Pipeline
```groovy
// Jenkinsfile
pipeline {
agent any
environment {
NODE_VERSION = '18'
}
stages {
stage('Setup') {
steps {
sh 'npm install -g mcp2skill-tools'
sh 'mcp2rest start'
sh 'mcp2rest add chrome-devtools'
}
}
stage('Install Dependencies') {
steps {
dir('~/.claude/skills/mcp-chrome-devtools/scripts') {
sh 'npm install'
}
}
}
stage('E2E Tests') {
parallel {
stage('Registration Flow') {
steps {
sh 'bash tests/test_registration.sh'
}
}
stage('Checkout Flow') {
steps {
sh 'bash tests/test_checkout.sh'
}
}
stage('Search Flow') {
steps {
sh 'bash tests/test_search.sh'
}
}
}
}
stage('Performance Tests') {
steps {
sh 'bash tests/performance_regression_test.sh'
}
}
}
post {
always {
archiveArtifacts artifacts: '**/*.png,**/*.txt,**/*.json', allowEmptyArchive: true
sh 'mcp2rest stop'
}
failure {
emailext (
subject: "E2E Tests Failed: ${env.JOB_NAME} - ${env.BUILD_NUMBER}",
body: "Check console output at ${env.BUILD_URL}",
to: "team@example.com"
)
}
}
}
```
---
## Cross-Browser Testing Simulation
### Device Emulation Matrix
```bash
#!/bin/bash
# device_matrix_test.sh
URL="https://example.com"
OUTPUT_DIR="device_screenshots"
mkdir -p $OUTPUT_DIR
# Define device viewports
declare -A DEVICES
DEVICES["desktop_1080"]="1920 1080"
DEVICES["desktop_720"]="1280 720"
DEVICES["laptop"]="1366 768"
DEVICES["ipad_portrait"]="768 1024"
DEVICES["ipad_landscape"]="1024 768"
DEVICES["iphone_portrait"]="375 667"
DEVICES["iphone_landscape"]="667 375"
DEVICES["android_portrait"]="360 640"
DEVICES["android_landscape"]="640 360"
echo "Testing across device viewports..."
# Open page
node scripts/new_page.js --url "$URL"
for device in "${!DEVICES[@]}"; do
echo "Testing device: $device"
# Parse dimensions
read -r width height <<< "${DEVICES[$device]}"
# Resize window
node scripts/resize_page.js --width $width --height $height
# Wait for reflow
sleep 2
# Capture screenshot
node scripts/take_screenshot.js --fullPage true --filePath "$OUTPUT_DIR/${device}.png"
# Capture snapshot for debugging
node scripts/take_snapshot.js --filePath "$OUTPUT_DIR/${device}_snapshot.txt"
done
echo "✓ Device matrix testing complete. Screenshots in $OUTPUT_DIR/"
```
---
## Advanced JavaScript Extraction
### Extract Structured Data from SPA
```bash
#!/bin/bash
# extract_spa_data.sh
echo "Extracting data from Single Page Application..."
# Open SPA
node scripts/new_page.js --url https://example.com/dashboard
# Wait for app to initialize
node scripts/wait_for.js --text "Dashboard" --timeout 15000
# Wait for data to load (check for spinner to disappear)
node scripts/evaluate_script.js --function "() => new Promise(resolve => {
const checkSpinner = setInterval(() => {
if (!document.querySelector('.loading-spinner')) {
clearInterval(checkSpinner);
resolve(true);
}
}, 100);
})"
# Extract React/Vue component state
APP_STATE=$(node scripts/evaluate_script.js --function "() => {
// Try to access React DevTools state
const root = document.querySelector('#root');
const reactInternals = root._reactRootContainer || root._reactInternalFiber;
// Or access Vue instance
const vueInstance = root.__vue__;
// Extract data from global state (Redux, Vuex, etc.)
return {
redux: window.__REDUX_DEVTOOLS_EXTENSION__ ? window.store.getState() : null,
vuex: window.__VUEX_DEVTOOLS_GLOBAL_HOOK__ ? window.$store.state : null,
// Or extract directly from DOM
domData: Array.from(document.querySelectorAll('.data-item')).map(item => ({
id: item.dataset.id,
title: item.querySelector('.title').textContent,
value: item.querySelector('.value').textContent
}))
};
}")
echo "$APP_STATE" | jq '.' > app_state.json
echo "✓ Extracted application state to app_state.json"
```
### Monitor Real-Time Updates
```bash
#!/bin/bash
# monitor_realtime_updates.sh
echo "Monitoring real-time updates..."
# Open page with real-time data
node scripts/new_page.js --url https://example.com/live-feed
# Set up monitoring script
node scripts/evaluate_script.js --function "() => {
window.updates = [];
window.observer = new MutationObserver(mutations => {
mutations.forEach(mutation => {
if (mutation.type === 'childList') {
mutation.addedNodes.forEach(node => {
if (node.classList && node.classList.contains('update-item')) {
window.updates.push({
timestamp: new Date().toISOString(),
content: node.textContent,
type: node.dataset.type
});
}
});
}
});
});
const container = document.querySelector('.updates-container');
window.observer.observe(container, { childList: true, subtree: true });
return 'Monitoring started';
}"
# Wait for updates to accumulate
echo "Collecting updates for 60 seconds..."
sleep 60
# Retrieve collected updates
UPDATES=$(node scripts/evaluate_script.js --function "() => window.updates")
echo "$UPDATES" | jq '.' > realtime_updates.json
# Clean up
node scripts/evaluate_script.js --function "() => { window.observer.disconnect(); return 'Monitoring stopped'; }"
echo "✓ Collected updates saved to realtime_updates.json"
```
---
## Complex Workflow Orchestration
### Multi-User Simulation
```bash
#!/bin/bash
# simulate_multi_user.sh
echo "Simulating multi-user interactions..."
# User 1: Admin workflow
(
echo "User 1: Admin actions"
node scripts/new_page.js --url https://example.com/admin
node scripts/wait_for.js --text "Admin Panel" --timeout 10000
# ... admin actions ...
) &
# User 2: Customer workflow
(
echo "User 2: Customer actions"
node scripts/new_page.js --url https://example.com/shop
node scripts/wait_for.js --text "Products" --timeout 10000
# ... customer actions ...
) &
# Wait for both users to complete
wait
echo "✓ Multi-user simulation complete"
```
This advanced examples reference provides production-ready patterns for complex automation, testing, and data extraction workflows.

View File

@@ -0,0 +1,772 @@
# Complete Tool Reference
Alphabetical listing of all 26 chrome-devtools tools with full parameters and examples.
## Table of Contents
- [click](#click)
- [close_page](#close_page)
- [drag](#drag)
- [emulate](#emulate)
- [evaluate_script](#evaluate_script)
- [fill](#fill)
- [fill_form](#fill_form)
- [get_console_message](#get_console_message)
- [get_network_request](#get_network_request)
- [handle_dialog](#handle_dialog)
- [hover](#hover)
- [list_console_messages](#list_console_messages)
- [list_network_requests](#list_network_requests)
- [list_pages](#list_pages)
- [navigate_page](#navigate_page)
- [new_page](#new_page)
- [performance_analyze_insight](#performance_analyze_insight)
- [performance_start_trace](#performance_start_trace)
- [performance_stop_trace](#performance_stop_trace)
- [press_key](#press_key)
- [resize_page](#resize_page)
- [select_page](#select_page)
- [take_screenshot](#take_screenshot)
- [take_snapshot](#take_snapshot)
- [upload_file](#upload_file)
- [wait_for](#wait_for)
---
## click
Clicks on an element identified by its UID.
**Group:** Element Interaction
**Required Parameters:**
- `--uid UID` - Element unique identifier from snapshot
**Optional Parameters:**
- `--dblClick BOOLEAN` - Double click instead of single click (default: false)
**Examples:**
```bash
# Single click
node scripts/click.js --uid button_submit_abc123
# Double click
node scripts/click.js --uid file_icon_xyz789 --dblClick true
```
**Related Tools:** hover, fill, take_snapshot
---
## close_page
Closes a page by its index. Cannot close the last remaining page.
**Group:** Page Management
**Required Parameters:**
- `--pageIdx NUMBER` - Index of page to close (from list_pages)
**Examples:**
```bash
node scripts/close_page.js --pageIdx 2
```
**Notes:** Always keep at least one page open. Close pages from highest to lowest index to avoid index shifting.
**Related Tools:** list_pages, new_page, select_page
---
## drag
Drags one element onto another element.
**Group:** Element Interaction
**Required Parameters:**
- `--from-uid UID` - UID of element to drag
- `--to-uid UID` - UID of target drop zone
**Examples:**
```bash
node scripts/drag.js --from-uid task_item_abc --to-uid dropzone_xyz
```
**Related Tools:** hover, click, take_snapshot
---
## emulate
Emulates network conditions and CPU throttling.
**Group:** Performance Analysis
**Optional Parameters:**
- `--networkConditions JSON` - Network throttling config
- `--cpuThrottlingRate NUMBER` - CPU throttling multiplier (1 = normal)
**Network Conditions Format:**
```json
{
"downloadThroughput": 50000,
"uploadThroughput": 50000,
"latency": 2000
}
```
**Examples:**
```bash
# Slow 3G
node scripts/emulate.js --networkConditions '{"downloadThroughput":50000,"uploadThroughput":50000,"latency":2000}'
# 4x CPU slowdown
node scripts/emulate.js --cpuThrottlingRate 4
# Combined
node scripts/emulate.js --networkConditions '{"downloadThroughput":100000,"uploadThroughput":50000,"latency":100}' --cpuThrottlingRate 2
# Reset to normal
node scripts/emulate.js --cpuThrottlingRate 1
```
**Common Presets:**
- Slow 3G: 50KB/s down, 50KB/s up, 2000ms latency
- Fast 3G: 180KB/s down, 84KB/s up, 562ms latency
- 4G: 1.6MB/s down, 750KB/s up, 150ms latency
**Related Tools:** performance_start_trace, performance_stop_trace
---
## evaluate_script
Executes JavaScript in the page context and returns JSON-serializable results.
**Group:** Performance Analysis
**Required Parameters:**
- `--function STRING` - JavaScript function as string
**Optional Parameters:**
- `--args JSON_ARRAY` - Function arguments as JSON array
**Examples:**
```bash
# Simple extraction
node scripts/evaluate_script.js --function "() => document.title"
# With parameters
node scripts/evaluate_script.js --function "(x, y) => x + y" --args '[5, 3]'
# Extract array of data
node scripts/evaluate_script.js --function "() => Array.from(document.querySelectorAll('h2')).map(h => h.textContent)"
# Complex object extraction
node scripts/evaluate_script.js --function "() => ({title: document.title, links: document.querySelectorAll('a').length})"
```
**Important:** Return values must be JSON-serializable (no DOM nodes, functions, or circular references).
**Related Tools:** take_snapshot, wait_for
---
## fill
Types text into inputs/textareas or selects dropdown options.
**Group:** Element Interaction
**Required Parameters:**
- `--uid UID` - Element unique identifier
- `--value STRING` - Text to type or option to select
**Examples:**
```bash
# Text input
node scripts/fill.js --uid input_username_abc --value "john.doe"
# Textarea
node scripts/fill.js --uid textarea_comment_def --value "This is a long comment"
# Dropdown select
node scripts/fill.js --uid select_country_ghi --value "United States"
```
**Related Tools:** fill_form, click, take_snapshot
---
## fill_form
Fills multiple form fields at once.
**Group:** Element Interaction
**Required Parameters:**
- `--elements JSON_ARRAY` - Array of {uid, value} objects
**Format:**
```json
[
{"uid": "input_email", "value": "test@example.com"},
{"uid": "input_password", "value": "secret123"}
]
```
**Examples:**
```bash
node scripts/fill_form.js --elements '[{"uid":"input_email","value":"test@example.com"},{"uid":"input_password","value":"secret123"}]'
# Multi-line for readability (bash)
node scripts/fill_form.js --elements '[
{"uid":"input_name","value":"John Doe"},
{"uid":"input_email","value":"john@example.com"},
{"uid":"select_country","value":"USA"}
]'
```
**Related Tools:** fill, click, take_snapshot
---
## get_console_message
Retrieves a specific console message by ID.
**Group:** Inspection & Debugging
**Required Parameters:**
- `--msgid STRING` - Message ID from list_console_messages
**Examples:**
```bash
node scripts/get_console_message.js --msgid msg_abc123
```
**Related Tools:** list_console_messages
---
## get_network_request
Gets details of a specific network request.
**Group:** Inspection & Debugging
**Optional Parameters:**
- `--reqid STRING` - Request ID from list_network_requests (if omitted, returns currently selected request in DevTools)
**Examples:**
```bash
# Specific request
node scripts/get_network_request.js --reqid req_abc123
# Currently selected in DevTools
node scripts/get_network_request.js
```
**Output includes:** URL, method, status code, headers, request/response body, timing.
**Related Tools:** list_network_requests
---
## handle_dialog
Handles browser dialogs (alert, confirm, prompt).
**Group:** Performance Analysis
**Required Parameters:**
- `--action STRING` - Either "accept" or "dismiss"
**Optional Parameters:**
- `--promptText STRING` - Text to enter in prompt dialog (only for prompts)
**Examples:**
```bash
# Accept alert
node scripts/handle_dialog.js --action accept
# Dismiss confirm
node scripts/handle_dialog.js --action dismiss
# Accept prompt with input
node scripts/handle_dialog.js --action accept --promptText "My Answer"
```
**Notes:** Dialog must already be open when calling this tool.
**Related Tools:** click, wait_for
---
## hover
Hovers the mouse over an element.
**Group:** Element Interaction
**Required Parameters:**
- `--uid UID` - Element unique identifier
**Examples:**
```bash
node scripts/hover.js --uid tooltip_trigger_abc
# Common pattern: hover then click
node scripts/hover.js --uid menu_trigger_xyz
node scripts/click.js --uid menu_item_settings
```
**Use Cases:** Triggering tooltips, dropdown menus, hover effects.
**Related Tools:** click, take_snapshot
---
## list_console_messages
Lists console messages (logs, warnings, errors) from the current page.
**Group:** Inspection & Debugging
**Optional Parameters:**
- `--pageSize NUMBER` - Number of messages per page
- `--pageIdx NUMBER` - Page index for pagination (0-based)
- `--types STRING` - Comma-separated types: log,warn,error,info
- `--includePreservedMessages BOOLEAN` - Include messages from before current navigation
**Examples:**
```bash
# All messages
node scripts/list_console_messages.js
# Only errors
node scripts/list_console_messages.js --types error
# Errors and warnings
node scripts/list_console_messages.js --types error,warn
# Paginated
node scripts/list_console_messages.js --pageSize 10 --pageIdx 0
```
**Related Tools:** get_console_message
---
## list_network_requests
Lists network requests made by the current page.
**Group:** Inspection & Debugging
**Optional Parameters:**
- `--pageSize NUMBER` - Number of requests per page
- `--pageIdx NUMBER` - Page index for pagination (0-based)
- `--resourceTypes STRING` - Comma-separated types: fetch,xhr,document,script,stylesheet,image
- `--includePreservedRequests BOOLEAN` - Include requests from before current navigation
**Examples:**
```bash
# All requests
node scripts/list_network_requests.js
# Only API calls
node scripts/list_network_requests.js --resourceTypes fetch,xhr
# Only scripts and stylesheets
node scripts/list_network_requests.js --resourceTypes script,stylesheet
# Paginated
node scripts/list_network_requests.js --pageSize 20 --pageIdx 0
```
**Related Tools:** get_network_request
---
## list_pages
Lists all open browser pages with their indices and URLs.
**Group:** Page Management
**No Parameters**
**Examples:**
```bash
node scripts/list_pages.js
```
**Output Example:**
```
Page 0: https://example.com (selected)
Page 1: https://google.com
Page 2: https://github.com
```
**Related Tools:** new_page, select_page, close_page
---
## navigate_page
Navigates the currently selected page.
**Group:** Page Management
**Optional Parameters:**
- `--url STRING` - URL to navigate to
- `--type STRING` - Navigation type: navigate, reload, back, forward
- `--ignoreCache BOOLEAN` - Bypass cache on reload
- `--timeout NUMBER` - Navigation timeout in milliseconds
**Examples:**
```bash
# Navigate to URL
node scripts/navigate_page.js --url https://example.com/page2
# Reload page
node scripts/navigate_page.js --type reload
# Reload ignoring cache
node scripts/navigate_page.js --type reload --ignoreCache true
# Go back
node scripts/navigate_page.js --type back
# Go forward
node scripts/navigate_page.js --type forward
```
**Related Tools:** new_page, wait_for
---
## new_page
Creates a new browser page (tab).
**Group:** Page Management
**Required Parameters:**
- `--url STRING` - URL to open
**Optional Parameters:**
- `--timeout NUMBER` - Page load timeout in milliseconds (default: 30000)
**Examples:**
```bash
node scripts/new_page.js --url https://example.com
# With longer timeout
node scripts/new_page.js --url https://slow-site.com --timeout 60000
```
**Notes:** New page becomes the selected page automatically.
**Related Tools:** list_pages, select_page, close_page
---
## performance_analyze_insight
Gets detailed information about a specific performance insight.
**Group:** Performance Analysis
**Required Parameters:**
- `--insightSetId STRING` - Insight set ID from performance trace
- `--insightName STRING` - Name of specific insight to analyze
**Examples:**
```bash
node scripts/performance_analyze_insight.js --insightSetId set_abc123 --insightName LargestContentfulPaint
node scripts/performance_analyze_insight.js --insightSetId set_abc123 --insightName CumulativeLayoutShift
```
**Common Insight Names:**
- LargestContentfulPaint
- FirstContentfulPaint
- CumulativeLayoutShift
- TotalBlockingTime
- TimeToInteractive
**Related Tools:** performance_start_trace, performance_stop_trace
---
## performance_start_trace
Starts performance trace recording.
**Group:** Performance Analysis
**Required Parameters:**
- `--reload BOOLEAN` - Reload page before tracing
- `--autoStop BOOLEAN` - Automatically stop when page loads
**Examples:**
```bash
# Start with page reload
node scripts/performance_start_trace.js --reload true --autoStop false
# Start without reload
node scripts/performance_start_trace.js --reload false --autoStop true
```
**Notes:** Only one trace can be active at a time. Call performance_stop_trace to get results.
**Related Tools:** performance_stop_trace, performance_analyze_insight
---
## performance_stop_trace
Stops the active performance trace and returns results.
**Group:** Performance Analysis
**No Parameters**
**Examples:**
```bash
node scripts/performance_stop_trace.js
```
**Output includes:**
- Core Web Vitals scores (LCP, FID, CLS)
- Performance insights
- Insight set IDs for detailed analysis
- Timing metrics
**Related Tools:** performance_start_trace, performance_analyze_insight
---
## press_key
Presses a key or key combination.
**Group:** Element Interaction
**Required Parameters:**
- `--key STRING` - Key or key combination to press
**Examples:**
```bash
# Single key
node scripts/press_key.js --key Enter
node scripts/press_key.js --key Tab
node scripts/press_key.js --key Escape
# Arrow keys
node scripts/press_key.js --key ArrowDown
node scripts/press_key.js --key ArrowUp
# Key combinations (use + separator)
node scripts/press_key.js --key "Control+S"
node scripts/press_key.js --key "Control+Shift+P"
node scripts/press_key.js --key "Alt+F4"
```
**Common Keys:** Enter, Tab, Escape, Space, Backspace, Delete, ArrowUp, ArrowDown, ArrowLeft, ArrowRight
**Modifiers:** Control, Shift, Alt, Meta (Command on Mac)
**Related Tools:** fill, click
---
## resize_page
Resizes the browser window to specific dimensions.
**Group:** Page Management
**Required Parameters:**
- `--width NUMBER` - Width in pixels
- `--height NUMBER` - Height in pixels
**Examples:**
```bash
# Desktop
node scripts/resize_page.js --width 1920 --height 1080
# Laptop
node scripts/resize_page.js --width 1366 --height 768
# Tablet
node scripts/resize_page.js --width 768 --height 1024
# Mobile
node scripts/resize_page.js --width 375 --height 667
```
**Common Dimensions:**
- Desktop: 1920x1080, 1440x900, 1366x768
- Tablet: 768x1024 (iPad), 600x960
- Mobile: 375x667 (iPhone), 360x640 (Android)
**Related Tools:** take_screenshot, emulate
---
## select_page
Switches the active context to a specific page.
**Group:** Page Management
**Required Parameters:**
- `--pageIdx NUMBER` - Page index from list_pages
**Examples:**
```bash
node scripts/select_page.js --pageIdx 0
node scripts/select_page.js --pageIdx 2
```
**Notes:** All subsequent commands operate on the selected page.
**Related Tools:** list_pages, new_page
---
## take_screenshot
Captures a visual screenshot of the page or element.
**Group:** Inspection & Debugging
**Optional Parameters:**
- `--format STRING` - Image format: png or jpeg (default: png)
- `--quality NUMBER` - JPEG quality 0-100 (default: 90)
- `--uid STRING` - Element UID to screenshot (if omitted, screenshots viewport)
- `--fullPage BOOLEAN` - Capture full scrollable page (default: false)
- `--filePath STRING` - Path to save image file
**Examples:**
```bash
# Full page PNG
node scripts/take_screenshot.js --fullPage true --filePath page.png
# Viewport only
node scripts/take_screenshot.js --filePath viewport.png
# Specific element
node scripts/take_screenshot.js --uid element_abc123 --filePath element.png
# Compressed JPEG
node scripts/take_screenshot.js --format jpeg --quality 80 --filePath page.jpg
```
**Related Tools:** take_snapshot
---
## take_snapshot
Captures text-based page structure with element UIDs.
**Group:** Inspection & Debugging
**Optional Parameters:**
- `--verbose BOOLEAN` - Include more details (default: false)
- `--filePath STRING` - Path to save snapshot file
**Examples:**
```bash
# Console output
node scripts/take_snapshot.js
# Verbose mode
node scripts/take_snapshot.js --verbose true
# Save to file
node scripts/take_snapshot.js --filePath snapshot.txt
# Both
node scripts/take_snapshot.js --verbose true --filePath detailed_snapshot.txt
```
**Output Format:**
```
Page: https://example.com
Title: Example Domain
button "Submit" [uid: button_submit_abc123]
input "Email" [uid: input_email_def456]
link "About" [uid: link_about_ghi789]
```
**Important:** Always use UIDs from the most recent snapshot. UIDs regenerate on each snapshot.
**Related Tools:** click, fill, hover, drag, take_screenshot
---
## upload_file
Uploads a file through a file input element.
**Group:** Element Interaction
**Required Parameters:**
- `--uid UID` - File input element UID
- `--filePath STRING` - Absolute path to file to upload
**Examples:**
```bash
# Upload document
node scripts/upload_file.js --uid input_file_abc --filePath /Users/username/documents/resume.pdf
# Upload image
node scripts/upload_file.js --uid input_avatar_xyz --filePath /Users/username/pictures/profile.jpg
```
**Notes:**
- Use absolute paths (not relative)
- Use forward slashes in paths
- Verify file exists before upload
**Related Tools:** click, fill, wait_for
---
## wait_for
Waits for specific text to appear on the page.
**Group:** Performance Analysis
**Required Parameters:**
- `--text STRING` - Text to wait for (exact match)
**Optional Parameters:**
- `--timeout NUMBER` - Maximum wait time in milliseconds (default: 30000)
**Examples:**
```bash
# Wait up to 30 seconds (default)
node scripts/wait_for.js --text "Loading complete"
# Wait up to 10 seconds
node scripts/wait_for.js --text "Welcome" --timeout 10000
# Wait for error message
node scripts/wait_for.js --text "Error occurred" --timeout 5000
```
**Notes:**
- Text match is exact (case-sensitive)
- Times out if text doesn't appear within timeout period
- Use after navigation, clicks, or any action that triggers loading
**Related Tools:** click, navigate_page, take_snapshot

View File

@@ -0,0 +1,743 @@
# Troubleshooting Guide
Common issues and solutions for chrome-devtools skill.
## Table of Contents
- [Connection Issues](#connection-issues)
- [Page Management Problems](#page-management-problems)
- [Element Interaction Issues](#element-interaction-issues)
- [Snapshot and Screenshot Problems](#snapshot-and-screenshot-problems)
- [Console and Network Debugging](#console-and-network-debugging)
- [Performance Analysis Issues](#performance-analysis-issues)
- [General Best Practices](#general-best-practices)
---
## Connection Issues
### Problem: "Cannot connect to mcp2rest"
**Symptoms:**
- Scripts fail with connection error
- "ECONNREFUSED" error on port 28888
**Solutions:**
1. **Check if mcp2rest is running:**
```bash
curl http://localhost:28888/health
```
Expected: `{"status": "healthy"}`
2. **Start mcp2rest if not running:**
```bash
mcp2rest start
```
3. **Verify chrome-devtools server is loaded:**
```bash
curl http://localhost:28888/servers
```
Should list chrome-devtools with "connected" status.
4. **Reload server if disconnected:**
```bash
mcp2rest reload chrome-devtools
```
### Problem: "Server not found"
**Symptoms:**
- Server not listed in mcp2rest
- Tools return "server not configured" error
**Solutions:**
1. **Check installed servers:**
```bash
mcp2rest list
```
2. **Add chrome-devtools server:**
```bash
mcp2rest add chrome-devtools
```
3. **Restart mcp2rest:**
```bash
mcp2rest restart
```
---
## Page Management Problems
### Problem: "Cannot close last page"
**Symptoms:**
- close_page fails with error about last page
**Solution:**
Always keep at least one page open. Open a new page before closing the last one:
```bash
# Check current pages
node scripts/list_pages.js
# Open new page first
node scripts/new_page.js --url https://example.com
# Now close the old page
node scripts/close_page.js --pageIdx 1
```
### Problem: "Page index out of range"
**Symptoms:**
- select_page or close_page fails
- "Invalid page index" error
**Solutions:**
1. **Always list pages first:**
```bash
node scripts/list_pages.js
```
2. **Use indices from the current list:**
Page indices shift when pages are closed. Always get fresh indices.
3. **Close pages in reverse order:**
```bash
# If closing multiple pages, start from highest index
node scripts/close_page.js --pageIdx 3
node scripts/close_page.js --pageIdx 2
node scripts/close_page.js --pageIdx 1
```
### Problem: "Navigation timeout"
**Symptoms:**
- new_page or navigate_page times out
- Page doesn't load within timeout period
**Solutions:**
1. **Increase timeout:**
```bash
node scripts/new_page.js --url https://slow-site.com --timeout 60000
```
2. **Check network connectivity:**
Verify the URL loads in a regular browser.
3. **Use emulation to debug slow loading:**
```bash
# Test without throttling first
node scripts/emulate.js --cpuThrottlingRate 1
node scripts/new_page.js --url https://example.com
```
### Problem: "Wrong page context"
**Symptoms:**
- Commands execute on wrong page
- Unexpected elements in snapshot
**Solutions:**
1. **Check selected page:**
```bash
node scripts/list_pages.js
# Look for "(selected)" indicator
```
2. **Explicitly select target page:**
```bash
node scripts/select_page.js --pageIdx 0
```
3. **Pattern: Always select before interaction:**
```bash
node scripts/list_pages.js
node scripts/select_page.js --pageIdx 1
node scripts/take_snapshot.js
```
---
## Element Interaction Issues
### Problem: "Element UID not found"
**Symptoms:**
- click, fill, or hover fails
- "UID does not exist" error
**Solutions:**
1. **Take fresh snapshot:**
UIDs regenerate on each snapshot. Never reuse old UIDs.
```bash
# Always take new snapshot before interaction
node scripts/take_snapshot.js
# Use UIDs from this output
node scripts/click.js --uid <fresh-uid>
```
2. **Verify element exists:**
Check the snapshot output to confirm the element is present.
3. **Wait for page to load:**
```bash
node scripts/wait_for.js --text "Page loaded" --timeout 10000
node scripts/take_snapshot.js
```
### Problem: "Click doesn't trigger action"
**Symptoms:**
- Element is clicked but nothing happens
- No error, but expected behavior doesn't occur
**Solutions:**
1. **Wait for page to fully load:**
```bash
node scripts/wait_for.js --text "Ready" --timeout 5000
node scripts/click.js --uid button_abc
```
2. **Try hover first:**
Some elements require hover to become clickable.
```bash
node scripts/hover.js --uid element_abc
node scripts/click.js --uid element_abc
```
3. **Check if element is visible:**
Hidden or off-screen elements may not be clickable. Use take_snapshot with verbose to check visibility.
4. **Use double-click if needed:**
```bash
node scripts/click.js --uid element_abc --dblClick true
```
### Problem: "Fill doesn't work on input"
**Symptoms:**
- Text not entered into field
- Select dropdown not changing
**Solutions:**
1. **Verify element type:**
Use take_snapshot to confirm it's an input/textarea/select.
2. **Click to focus first:**
```bash
node scripts/click.js --uid input_abc
node scripts/fill.js --uid input_abc --value "text"
```
3. **For custom components, use press_key:**
```bash
node scripts/click.js --uid custom_input
node scripts/press_key.js --key "H"
node scripts/press_key.js --key "i"
```
4. **For dropdowns, use exact option value:**
Check the HTML to find the correct option value.
### Problem: "Drag and drop doesn't work"
**Symptoms:**
- Drag operation completes but element doesn't move
- "Invalid UID" error
**Solutions:**
1. **Verify both UIDs are current:**
```bash
node scripts/take_snapshot.js
# Use UIDs from this output
node scripts/drag.js --from-uid <fresh-from> --to-uid <fresh-to>
```
2. **Add wait between operations:**
```bash
node scripts/drag.js --from-uid item1 --to-uid zone1
node scripts/wait_for.js --text "Moved" --timeout 3000
```
3. **Check if dropzone accepts the element:**
Some drag-and-drop UIs have restrictions on what can be dropped where.
### Problem: "File upload fails"
**Symptoms:**
- upload_file fails
- File path error
**Solutions:**
1. **Use absolute paths:**
```bash
# Good
node scripts/upload_file.js --uid input_file --filePath /Users/username/file.pdf
# Bad
# node scripts/upload_file.js --uid input_file --filePath ./file.pdf
```
2. **Use forward slashes:**
```bash
# Good
/Users/username/documents/file.pdf
# Bad (Windows-style)
# C:\Users\username\documents\file.pdf
```
3. **Verify file exists:**
```bash
ls -la /Users/username/documents/file.pdf
```
4. **Check file input UID:**
Take fresh snapshot and verify the element is a file input.
---
## Snapshot and Screenshot Problems
### Problem: "Snapshot shows different UIDs each time"
**Symptoms:**
- UIDs change on each snapshot
- Previously working UIDs stop working
**Solution:**
**This is expected behavior.** UIDs are dynamically generated on each snapshot. Always use the most recent snapshot:
```bash
# Pattern: Always snapshot before interaction
node scripts/take_snapshot.js
# Immediately use UIDs from output
node scripts/click.js --uid <current-uid>
```
### Problem: "Screenshot is blank or black"
**Symptoms:**
- Screenshot file is created but shows blank/black image
**Solutions:**
1. **Wait for page to render:**
```bash
node scripts/wait_for.js --text "Content" --timeout 10000
node scripts/take_screenshot.js --filePath page.png
```
2. **Try full page screenshot:**
```bash
node scripts/take_screenshot.js --fullPage true --filePath page.png
```
3. **Verify element is visible:**
For element screenshots, ensure the element isn't hidden or off-screen.
4. **Check page actually loaded:**
```bash
node scripts/take_snapshot.js
# Verify page content appears in snapshot
```
### Problem: "Screenshot file not saving"
**Symptoms:**
- No error but file not created
- "File not found" when opening
**Solutions:**
1. **Use absolute path:**
```bash
node scripts/take_screenshot.js --filePath /Users/username/screenshots/page.png
```
2. **Ensure directory exists:**
```bash
mkdir -p /Users/username/screenshots
node scripts/take_screenshot.js --filePath /Users/username/screenshots/page.png
```
3. **Check file permissions:**
Ensure you have write access to the target directory.
---
## Console and Network Debugging
### Problem: "Console messages not showing up"
**Symptoms:**
- list_console_messages returns empty or incomplete results
**Solutions:**
1. **Messages accumulate since navigation:**
Console messages are captured from the current page navigation onwards.
2. **Reload if needed:**
```bash
node scripts/navigate_page.js --type reload
# Wait for page load
node scripts/wait_for.js --text "Ready" --timeout 10000
# Now check console
node scripts/list_console_messages.js
```
3. **Include preserved messages:**
```bash
node scripts/list_console_messages.js --includePreservedMessages true
```
4. **Ensure page has errors to show:**
Use take_snapshot or evaluate_script to verify page loaded correctly.
### Problem: "Network requests list is empty"
**Symptoms:**
- list_network_requests returns no results
- Expected API calls not showing
**Solutions:**
1. **Requests are captured from navigation onwards:**
```bash
# Navigate first to capture requests
node scripts/navigate_page.js --url https://example.com
# Wait for page to load
node scripts/wait_for.js --text "Loaded" --timeout 10000
# Now check requests
node scripts/list_network_requests.js
```
2. **Filter by resource type:**
```bash
# Maybe filtering too strictly
node scripts/list_network_requests.js --resourceTypes fetch,xhr,document
```
3. **Include preserved requests:**
```bash
node scripts/list_network_requests.js --includePreservedRequests true
```
### Problem: "Cannot find console message or network request ID"
**Symptoms:**
- get_console_message or get_network_request fails
- "Invalid ID" error
**Solutions:**
1. **List first to get IDs:**
```bash
# For console
node scripts/list_console_messages.js
# Copy message ID from output
node scripts/get_console_message.js --msgid <id-from-list>
# For network
node scripts/list_network_requests.js
# Copy request ID from output
node scripts/get_network_request.js --reqid <id-from-list>
```
2. **IDs are session-specific:**
IDs don't persist across page navigations. Get fresh IDs after navigation.
---
## Performance Analysis Issues
### Problem: "evaluate_script returns null or error"
**Symptoms:**
- Script execution fails
- Returns null unexpectedly
- "Evaluation failed" error
**Solutions:**
1. **Return JSON-serializable values only:**
```bash
# Good: primitives, arrays, plain objects
node scripts/evaluate_script.js --function "() => document.title"
node scripts/evaluate_script.js --function "() => [1, 2, 3]"
node scripts/evaluate_script.js --function "() => ({key: 'value'})"
# Bad: DOM nodes, functions, circular references
# node scripts/evaluate_script.js --function "() => document.querySelector('div')"
```
2. **Extract values from DOM elements:**
```bash
# Instead of returning DOM nodes, extract their properties
node scripts/evaluate_script.js --function "() => Array.from(document.querySelectorAll('a')).map(a => a.href)"
```
3. **Check for JavaScript errors:**
```bash
node scripts/list_console_messages.js --types error
```
4. **Verify function syntax:**
```bash
# Arrow function
node scripts/evaluate_script.js --function "() => document.title"
# Function with parameters
node scripts/evaluate_script.js --function "(x, y) => x + y" --args '[5, 3]'
```
### Problem: "wait_for times out"
**Symptoms:**
- wait_for exceeds timeout
- Text never appears
**Solutions:**
1. **Increase timeout:**
```bash
node scripts/wait_for.js --text "Loaded" --timeout 30000
```
2. **Verify text actually appears:**
```bash
# Check with snapshot
node scripts/take_snapshot.js
# Look for the expected text
```
3. **Text match is exact:**
Check spelling, capitalization, and spacing.
```bash
# Case-sensitive match
# "Welcome" ≠ "welcome" ≠ "Welcome!"
```
4. **Wait for network first:**
```bash
# If text appears after API call, wait longer
node scripts/wait_for.js --text "Data loaded" --timeout 20000
```
### Problem: "Dialog not handled"
**Symptoms:**
- handle_dialog fails
- "No dialog present" error
**Solutions:**
1. **Dialog must be open first:**
```bash
# Trigger dialog
node scripts/click.js --uid button_delete
# Immediately handle it
node scripts/handle_dialog.js --action accept
```
2. **Some "dialogs" are custom HTML:**
If it's not a real browser dialog (alert/confirm/prompt), use click instead:
```bash
# For custom modal dialogs
node scripts/take_snapshot.js
node scripts/click.js --uid button_modal_ok
```
3. **Timing matters:**
Handle dialog immediately after triggering action.
### Problem: "Performance trace shows unexpected results"
**Symptoms:**
- CWV scores don't match expectations
- Missing performance data
**Solutions:**
1. **Ensure page fully loads:**
```bash
node scripts/performance_start_trace.js --reload true --autoStop false
node scripts/wait_for.js --text "Loaded" --timeout 20000
node scripts/performance_stop_trace.js
```
2. **Use reload for consistent measurement:**
```bash
# For initial page load measurement
node scripts/performance_start_trace.js --reload true --autoStop false
```
3. **Clear browser state:**
Navigation with cache clear:
```bash
node scripts/navigate_page.js --url https://example.com --ignoreCache true
```
4. **Disable emulation if active:**
```bash
node scripts/emulate.js --cpuThrottlingRate 1
```
### Problem: "Emulation not working"
**Symptoms:**
- Page loads at normal speed despite emulation
- No visible effect from throttling
**Solutions:**
1. **Emulation persists for session:**
Reset to normal first, then apply:
```bash
# Reset
node scripts/emulate.js --cpuThrottlingRate 1
# Then apply desired emulation
node scripts/emulate.js --cpuThrottlingRate 4
```
2. **Network conditions must be valid JSON:**
```bash
# All three properties required
node scripts/emulate.js --networkConditions '{"downloadThroughput":50000,"uploadThroughput":50000,"latency":2000}'
```
3. **Verify effect with performance trace:**
```bash
node scripts/emulate.js --cpuThrottlingRate 4
node scripts/performance_start_trace.js --reload true --autoStop false
node scripts/wait_for.js --text "Loaded" --timeout 60000
node scripts/performance_stop_trace.js
# Check timing metrics in output
```
---
## General Best Practices
### Always Snapshot Before Interaction
UIDs are dynamic. Take a fresh snapshot before every interaction:
```bash
# Good pattern
node scripts/take_snapshot.js
node scripts/click.js --uid <current-uid>
# Bad pattern (will fail)
# node scripts/take_snapshot.js
# ... do other things ...
# node scripts/click.js --uid <old-uid> # UID expired!
```
### Wait After Actions
Allow time for UI updates:
```bash
node scripts/click.js --uid button_submit
node scripts/wait_for.js --text "Success" --timeout 5000
node scripts/take_snapshot.js
```
### Use Absolute Paths
Always use absolute paths for files:
```bash
# Good
/Users/username/documents/file.pdf
/home/user/screenshots/page.png
# Bad
./file.pdf
../screenshots/page.png
```
### Check Connection First
Before running workflows, verify mcp2rest is healthy:
```bash
curl http://localhost:28888/health
```
### List Before Selecting
Always list pages before selecting or closing:
```bash
node scripts/list_pages.js
node scripts/select_page.js --pageIdx 1
```
### Handle Errors Gracefully
Check console and network for debugging:
```bash
# After error occurs
node scripts/list_console_messages.js --types error
node scripts/list_network_requests.js
node scripts/take_screenshot.js --filePath error_state.png
```
### Use Verbose for Debugging
Get more details when troubleshooting:
```bash
node scripts/take_snapshot.js --verbose true
```
### Save Outputs for Analysis
Redirect output to files for later review:
```bash
node scripts/list_console_messages.js > console_log.txt
node scripts/list_network_requests.js > network_log.txt
node scripts/take_snapshot.js --filePath snapshot.txt
```
---
## Getting Help
If you encounter issues not covered here:
1. **Check tool documentation:** @reference/all-tools.md
2. **Review workflows:** @workflows/[tool-group].md
3. **Test with minimal example:** Isolate the problem with simplest reproduction
4. **Check mcp2rest status:** Ensure server is connected and healthy
5. **Restart mcp2rest:** Sometimes a fresh start helps: `mcp2rest restart`