anbeime/skill / skills/chrome-automation/chrome-automation
chrome-automation
Flagged网页开发工程师与运营专员在需要自动化现有 Chrome 浏览器时必用!通过 CDP 协议无缝接管当前浏览器,自动配置环境并实时可视化执行任务。完美保留登录状态,一键完成复杂网页自动化流程,让繁琐操作自动化。
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 21.0 KB |
| references/commands.md | markdown | 11.5 KB |
| references/setup-mac.md | markdown | 6.2 KB |
| references/setup-windows.md | markdown | 9.5 KB |
| references/troubleshooting.md | markdown | 10.2 KB |
| scripts/README.md | markdown | 3.0 KB |
| scripts/auto-install-mac.sh | script | 2.5 KB |
| scripts/auto-install-windows.ps1 | script | 3.8 KB |
| scripts/connect_chrome.sh | script | 1.5 KB |
| scripts/install-scripts.md | markdown | 9.2 KB |
| scripts/setup.sh | script | 1.6 KB |
| scripts/start-chrome-mac.sh | script | 3.4 KB |
| scripts/start-chrome-windows.ps1 | script | 5.0 KB |
references/commands.md
View on GitHub# Agent-Browser Command Reference for CDP
Complete command reference for controlling Chrome via CDP (Chrome DevTools Protocol).
## Command Format
### Mac/Linux
```bash
cd ~/Documents/agent-browser
AGENT_BROWSER_HOME=~/Documents/agent-browser \
./bin/agent-browser --cdp 9222 <command> [options]
```
### Windows (PowerShell)
```powershell
Set-Location "$HOME\Documents\agent-browser"
$env:AGENT_BROWSER_HOME="$HOME\Documents\agent-browser"
.\bin\agent-browser.cmd --cdp 9222 <command> [options]
```
For brevity, examples below show only the command portion (works on both platforms).
## Navigation Commands
### open (goto, navigate)
Navigate to a URL.
```bash
--cdp 9222 open https://example.com
--cdp 9222 open google.com # Auto-prepends https://
--cdp 9222 open file:///path/to/file.html
```
**Wait after:** Always wait 2-3 seconds for page load
```bash
--cdp 9222 open https://example.com
sleep 2
```
### back
Go back in history.
```bash
--cdp 9222 back
```
### forward
Go forward in history.
```bash
--cdp 9222 forward
```
### reload
Reload current page.
```bash
--cdp 9222 reload
```
## Page Analysis Commands
### snapshot
Get page structure.
```bash
--cdp 9222 snapshot # Full accessibility tree
--cdp 9222 snapshot -i # Interactive elements only (RECOMMENDED)
--cdp 9222 snapshot -c # Compact output
--cdp 9222 snapshot -d 3 # Limit depth to 3
--cdp 9222 snapshot -s "#main" # Scope to CSS selector
```
**Output format:**
```
- textbox "Email" [ref=e1]
- textbox "Password" [ref=e2]
- button "Submit" [ref=e3]
```
Use refs (@e1, @e2, @e3) in subsequent commands.
### get url
Get current URL.
```bash
--cdp 9222 get url
```
### get title
Get page title.
```bash
--cdp 9222 get title
```
### get text
Get element text content.
```bash
--cdp 9222 get text @e1
```
### get value
Get input field value.
```bash
--cdp 9222 get value @e1
```
### get html
Get element's innerHTML.
```bash
--cdp 9222 get html @e1
```
### get attr
Get element attribute.
```bash
--cdp 9222 get attr @e1 href
--cdp 9222 get attr @e1 class
```
### get count
Count elements matching selector.
```bash
--cdp 9222 get count ".item"
--cdp 9222 get count "button"
```
### get box
Get element bounding box.
```bash
--cdp 9222 get box @e1
```
### get styles
Get computed styles.
```bash
--cdp 9222 get styles @e1
```
## Interaction Commands
### click
Click an element.
```bash
--cdp 9222 click @e1
```
### dblclick
Double-click an element.
```bash
--cdp 9222 dblclick @e1
```
### fill
Clear and type text into input.
```bash
--cdp 9222 fill @e1 "john@example.com"
--cdp 9222 fill @e1 "text with spaces"
```
### type
Type text without clearing (appends).
```bash
--cdp 9222 type @e1 "additional text"
```
### press
Press keyboard key.
```bash
--cdp 9222 press Enter
--cdp 9222 press Tab
--cdp 9222 press Escape
--cdp 9222 press Control+a # Key combination
--cdp 9222 press Command+c # macOS
```
### keydown
Hold key down.
```bash
--cdp 9222 keydown Shift
```
### keyup
Release key.
```bash
--cdp 9222 keyup Shift
```
### hover
Hover over element.
```bash
--cdp 9222 hover @e1
```
### check
Check a checkbox.
```bash
--cdp 9222 check @e1
```
### uncheck
Uncheck a checkbox.
```bash
--cdp 9222 uncheck @e1
```
### select
Select dropdown option(s).
```bash
--cdp 9222 select @e1 "option value"
--cdp 9222 select @e1 "value1" "value2" # Multiple
```
### focus
Focus an element.
```bash
--cdp 9222 focus @e1
```
### scroll
Scroll the page.
```bash
--cdp 9222 scroll down 500 # Scroll down 500px
--cdp 9222 scroll up 300 # Scroll up 300px
--cdp 9222 scroll down # Default: 300px down
```
### scrollintoview (scrollinto)
Scroll element into view.
```bash
--cdp 9222 scrollintoview @e1
```
### drag
Drag and drop.
```bash
--cdp 9222 drag @e1 @e2 # Drag e1 to e2
```
### upload
Upload file to file input.
```bash
--cdp 9222 upload @e1 /path/to/file.pdf
```
## Wait Commands
### wait
Wait for conditions.
```bash
--cdp 9222 wait @e1 # Wait for element
--cdp 9222 wait 2000 # Wait 2000ms
--cdp 9222 wait --text "Success" # Wait for text
--cdp 9222 wait -t "Success" # Short form
--cdp 9222 wait --url "**/dashboard" # Wait for URL pattern
--cdp 9222 wait -u "**/dashboard" # Short form
--cdp 9222 wait --load networkidle # Wait for network idle
--cdp 9222 wait -l networkidle # Short form
--cdp 9222 wait --fn "window.ready" # Wait for JS condition
--cdp 9222 wait -f "window.ready" # Short form
```
## State Check Commands
### is visible
Check if element is visible.
```bash
--cdp 9222 is visible @e1
```
### is enabled
Check if element is enabled.
```bash
--cdp 9222 is enabled @e1
```
### is checked
Check if checkbox is checked.
```bash
--cdp 9222 is checked @e1
```
## Capture Commands
### screenshot
Take screenshot.
```bash
--cdp 9222 screenshot # Output to stdout
--cdp 9222 screenshot ~/Desktop/page.png # Save to file
--cdp 9222 screenshot --full # Full page screenshot
--cdp 9222 screenshot -f # Short form
```
### pdf
Save page as PDF.
```bash
--cdp 9222 pdf output.pdf
```
## JavaScript Command
### eval
Execute JavaScript in page context.
```bash
--cdp 9222 eval "document.title"
--cdp 9222 eval "window.scrollTo(0, 500)"
--cdp 9222 eval "document.querySelector('.btn').click()"
# Multi-line with variables
--cdp 9222 eval "var el = document.querySelector('input'); el.value = 'text';"
```
**Note:** Use proper JavaScript syntax. Avoid optional chaining (?.), use standard querySelector instead.
## Find Commands (Semantic Locators)
Alternative to refs when you know specific text/attributes.
### By role
```bash
--cdp 9222 find role button click --name "Submit"
--cdp 9222 find role textbox fill --name "Email" "user@test.com"
```
### By text
```bash
--cdp 9222 find text "Sign In" click
--cdp 9222 find text "Sign In" click --exact # Exact match only
```
### By label
```bash
--cdp 9222 find label "Email" fill "user@test.com"
```
### By placeholder
```bash
--cdp 9222 find placeholder "Search" type "query"
```
### By alt text
```bash
--cdp 9222 find alt "Logo" click
```
### By title
```bash
--cdp 9222 find title "Close" click
```
### By test ID
```bash
--cdp 9222 find testid "submit-btn" click
```
### By position
```bash
--cdp 9222 find first ".item" click # First match
--cdp 9222 find last ".item" click # Last match
--cdp 9222 find nth 2 "a" hover # 2nd match (0-indexed)
```
## Tab Commands
### tab
List or switch tabs.
```bash
--cdp 9222 tab # List all tabs
--cdp 9222 tab 2 # Switch to tab index 2
--cdp 9222 tab close # Close current tab
--cdp 9222 tab close 2 # Close tab index 2
--cdp 9222 tab new # New tab
--cdp 9222 tab new https://example.com # New tab with URL
```
## Frame Commands
### frame
Switch to iframe or back to main.
```bash
--cdp 9222 frame "#iframe-id" # Switch to iframe
--cdp 9222 frame main # Back to main frame
```
## Dialog Commands
### dialog
Handle JavaScript dialogs (alert, confirm, prompt).
```bash
--cdp 9222 dialog accept # Accept dialog
--cdp 9222 dialog accept "text" # Accept with input text
--cdp 9222 dialog dismiss # Dismiss dialog
```
## Mouse Commands
### mouse move
Move mouse to coordinates.
```bash
--cdp 9222 mouse move 100 200
```
### mouse down
Press mouse button.
```bash
--cdp 9222 mouse down left
--cdp 9222 mouse down right
```
### mouse up
Release mouse button.
```bash
--cdp 9222 mouse up left
```
### mouse wheel
Scroll with mouse wheel.
```bash
--cdp 9222 mouse wheel 100 # Scroll down 100px
--cdp 9222 mouse wheel -100 # Scroll up 100px
```
## Debug Commands
### console
View console messages.
```bash
--cdp 9222 console # Show all messages
--cdp 9222 console --clear # Clear console
```
### errors
View page errors.
```bash
--cdp 9222 errors # Show errors
--cdp 9222 errors --clear # Clear errors
```
### highlight
Highlight element visually.
```bash
--cdp 9222 highlight @e1
```
## Cookie & Storage Commands
### cookies
Manage cookies.
```bash
--cdp 9222 cookies # Get all cookies
--cdp 9222 cookies set name value # Set cookie
--cdp 9222 cookies clear # Clear all cookies
```
### storage
Manage localStorage.
```bash
--cdp 9222 storage local # Get all localStorage
--cdp 9222 storage local key # Get specific key
--cdp 9222 storage local set key value # Set value
--cdp 9222 storage local clear # Clear all
```
## Browser Settings
### set viewport
Set viewport size.
```bash
--cdp 9222 set viewport 1920 1080
```
### set device
Emulate device.
```bash
--cdp 9222 set device "iPhone 14"
--cdp 9222 set device "iPad Pro"
```
### set geo (geolocation)
Set geolocation.
```bash
--cdp 9222 set geo 37.7749 -122.4194
```
### set offline
Toggle offline mode.
```bash
--cdp 9222 set offline on
--cdp 9222 set offline off
```
### set credentials (auth)
Set HTTP basic auth.
```bash
--cdp 9222 set credentials username password
```
### set media
Set media preferences.
```bash
--cdp 9222 set media dark # Dark color scheme
--cdp 9222 set media light # Light color scheme
--cdp 9222 set media light reduced-motion # Multiple preferences
```
## Common Workflows
### Form submission workflow
**Mac/Linux:**
```bash
--cdp 9222 open https://example.com/form
sleep 2
--cdp 9222 snapshot -i
# Output: textbox "Email" [ref=e1], textbox "Password" [ref=e2], button "Submit" [ref=e3]
--cdp 9222 fill @e1 "user@example.com"
--cdp 9222 fill @e2 "password123"
--cdp 9222 click @e3
sleep 2
--cdp 9222 screenshot result.png
```
**Windows:**
```powershell
.\bin\agent-browser.cmd --cdp 9222 open https://example.com/form
Start-Sleep -Seconds 2
.\bin\agent-browser.cmd --cdp 9222 snapshot -i
.\bin\agent-browser.cmd --cdp 9222 fill @e1 "user@example.com"
.\bin\agent-browser.cmd --cdp 9222 fill @e2 "password123"
.\bin\agent-browser.cmd --cdp 9222 click @e3
Start-Sleep -Seconds 2
.\bin\agent-browser.cmd --cdp 9222 screenshot result.png
```
### Search workflow
**Mac/Linux:**
```bash
--cdp 9222 open google.com
sleep 2
--cdp 9222 snapshot -i
# Output: textbox "Search" [ref=e1]
--cdp 9222 fill @e1 "search query"
--cdp 9222 press Enter
sleep 2
--cdp 9222 screenshot results.png
```
**Windows:**
```powershell
.\bin\agent-browser.cmd --cdp 9222 open google.com
Start-Sleep -Seconds 2
.\bin\agent-browser.cmd --cdp 9222 snapshot -i
.\bin\agent-browser.cmd --cdp 9222 fill @e1 "search query"
.\bin\agent-browser.cmd --cdp 9222 press Enter
Start-Sleep -Seconds 2
.\bin\agent-browser.cmd --cdp 9222 screenshot results.png
```
### Multi-page workflow
```bash
--cdp 9222 open https://site.com
sleep 2 # Windows: Start-Sleep -Seconds 2
--cdp 9222 snapshot -i
--cdp 9222 click @e1 # Click link
sleep 2 # Windows: Start-Sleep -Seconds 2
--cdp 9222 snapshot -i # Get new refs after navigation
--cdp 9222 fill @e2 "text"
```
## Best Practices
1. **Always wait after navigation**
```bash
--cdp 9222 open url
sleep 2 # Wait for load
```
2. **Re-snapshot after DOM changes**
```bash
--cdp 9222 click @e1
sleep 1
--cdp 9222 snapshot -i # Get fresh refs
```
3. **Use screenshots for feedback**
```bash
--cdp 9222 screenshot ~/Desktop/state.png
# Then read and show to user
```
4. **Check URL after navigation**
```bash
--cdp 9222 click @e1
sleep 2
--cdp 9222 get url # Verify navigation
```
5. **Use -i flag for snapshots**
```bash
--cdp 9222 snapshot -i # Faster, shows only interactive elements
```
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| slack-gif-creatoranthropics/skillsKnowledge and utilities for creating animated GIFs optimized for Slack. Provides constraints, validation tools, and animation concepts. Use when users request animated GIFs for Slack like "make me a GIF of X doing Y for Slack." | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |
| ponytailDietrichGebert/ponytailLazy senior dev mode for any coding task: the smallest change that fully solves it, and a reply a busy human understands in one read. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-debtDietrichGebert/ponytailHarvest every shortcut: comment into one debt ledger, so deferrals get tracked instead of forgotten. One-shot report. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-debtDietrichGebert/ponytailList every `shortcut:` comment (and older `ponytail:` ones) as a debt ledger. One-shot report, changes nothing. Use for "ponytail debt", "what did ponytail defer", "list the shortcuts", /ponytail-debt. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |