菜单
Automate browser interactions — navigation, clicking, typing, screenshots, JavaScript execution, and cloud browser management — using the browser-use CLI.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
| name | browser-use |
| version | 1.1.0 |
| description | Automate browser interactions — navigation, clicking, typing, screenshots, JavaScript execution, and cloud browser management — using the browser-use CLI. |
| allowed-tools | Browser(browser-use:*), Shell |
Fast, persistent browser automation from the command line via the browser-use CLI.
# macOS / Linux
curl -fsSL https://browser-use.com/cli/install.sh | bash
# Validate installation
browser-use doctor
# Optional: run setup wizard
browser-use setup
openstate to see numbered element indicesbrowser-use open https://example.com
browser-use state
# Shows: [0] input "Name", [1] button "Submit"
browser-use input 0 "John Doe"
browser-use click 1
browser-use --headed open https://example.com # Visible window (default)
browser-use open https://example.com # Headless Chromium
browser-use --profile "Default" open https://gmail.com # Real Chrome, Default profile
browser-use --profile "Profile 1" open https://example.com # Specific Chrome profile
browser-use --connect open https://example.com # Auto-discover running Chrome
browser-use --cdp-url http://localhost:9222 open https://example.com # CDP URL
| Command | Description |
|---|---|
open <url> | Navigate to URL |
back | Go back in history |
scroll down | Scroll down |
scroll up | Scroll up |
scroll down --amount 1000 | Scroll by pixels |
| Command | Description |
|---|---|
state | Get URL, title, and clickable elements |
screenshot [path] | Take screenshot (base64 if no path) |
screenshot --full path.png | Full page screenshot |
| Command | Description |
|---|---|
click <index> | Click element by index |
click <x> <y> | Click at pixel coordinates |
type "text" | Type into focused element |
input <index> "text" | Click element, then type |
keys "Enter" | Send keyboard keys |
keys "Control+a" | Send key combination |
select <index> "value" | Select dropdown option |
upload <index> <path> | Upload file to file input |
hover <index> | Hover over element |
dblclick <index> | Double-click element |
rightclick <index> | Right-click element |
| Command | Description |
|---|---|
get title | Get page title |
get html | Get full page HTML |
get html --selector "h1" | Get HTML of element |
get text <index> | Get text content of element |
get value <index> | Get value of input/textarea |
get attributes <index> | Get all attributes of element |
get bbox <index> | Get bounding box (x, y, width, height) |
| Command | Description |
|---|---|
wait selector "css" | Wait for element to be visible |
wait selector ".loading" --state hidden | Wait for element to disappear |
wait text "Success" | Wait for text to appear |
wait selector "h1" --timeout 5000 | Custom timeout (ms) |
| Command | Description |
|---|---|
eval "js code" | Execute JavaScript |
| Command | Description |
|---|---|
switch <tab> | Switch to tab by index |
close-tab | Close current tab |
close-tab <tab> | Close specific tab |
| Command | Description |
|---|---|
cookies get | Get all cookies |
cookies set <name> <value> | Set a cookie |
cookies clear | Clear all cookies |
cookies export <file> | Export to JSON file |
cookies import <file> | Import from JSON file |
browser-use python "x = 42"
browser-use python "print(x)" # Access variables across calls
browser-use python "print(browser.url)" # Access browser object
browser-use python --vars # Show defined variables
browser-use python --reset # Clear namespace
browser-use python --file script.py # Run Python file
Each --session gets its own daemon, socket, and PID file. The browser stays alive between commands (~50ms latency).
browser-use open https://example.com # uses session 'default'
browser-use --session work open https://example.com
browser-use --session work state
browser-use sessions # list active sessions
browser-use --session work close
browser-use close --all # close all sessions
browser-use cloud login sk-abc123... # save API key
browser-use cloud connect # provision cloud browser
browser-use cloud connect --proxy-country US
browser-use cloud connect --profile-id <id>
browser-use state # works normally with cloud browser
browser-use close # disconnects and stops cloud browser
# REST passthrough
browser-use cloud v2 GET /browsers
browser-use cloud v2 POST /tasks '{"task":"Search for AI news","url":"https://google.com"}'
browser-use cloud v2 poll <task-id>
browser-use cloud v3 --help
npm run dev & # start local server on :3000
browser-use tunnel 3000 # → https://abc.trycloudflare.com
browser-use cloud connect
browser-use open https://abc.trycloudflare.com
browser-use tunnel list
browser-use tunnel stop 3000
| Option | Description |
|---|---|
--headed | Show browser window |
--profile [NAME] | Use real Chrome profile |
--connect | Auto-discover running Chrome via CDP |
--cdp-url <url> | Connect to existing browser via CDP URL |
--session NAME | Target a named session (default: "default") |
--json | Output as JSON |
--mcp | Run as MCP server via stdin/stdout |
browser-use open https://news.ycombinator.com
browser-use eval "Array.from(document.querySelectorAll('.titleline a')).slice(0,5).map(a => a.textContent)"
browser-use open https://example.com/contact
browser-use state
browser-use input 0 "John Doe"
browser-use input 1 "john@example.com"
browser-use click 2
browser-use open https://example.com
browser-use python "
for i in range(5):
browser.scroll('down')
browser.wait(0.5)
browser.screenshot('scrolled.png')
"
Ultra-compressed communication mode. Slash token usage ~75% by speaking like caveman while keeping full technical accuracy. Use when user says "caveman mode", "talk like caveman", "use caveman", "less tokens", "be brief", or invokes /caveman. Also auto-triggers when token efficiency is requested.
Browser automation CLI for AI agents. Use when the user needs to interact with websites, including navigating pages, filling forms, clicking buttons, taking screenshots, extracting data, testing web apps, or automating any browser task. Triggers include requests to "open a website", "fill out a form", "click a button", "take a screenshot", "scrape data from a page", "test this web app", "login to a site", "automate browser actions", or any task requiring programmatic web interaction.
Use when working with Jira using Atlassian's Command Line Interface (acli), to get information about Jira, create and manage tickets and issues.
Create distinctive, production-grade frontend interfaces with high design quality. Use this skill when the user asks to build web components, pages, or applications. Generates creative, polished code that avoids generic AI aesthetics.
Use when working with Github using the Github CLI (gh), to get information about Github, create and manage repositories, issues, and pull requests.
Use when implementing any feature or bugfix, before writing implementation code - write the test first, watch it fail, write minimal code to pass; ensures tests actually verify behavior by requiring failure first