功能描述

Automates browser interactions for web testing, form filling, screenshots, and data extraction. Use when the user needs to navigate websites, interact with w...

使用说明 (SKILL.md)

Browser Automation with browser-use CLI

Name: Browser Use 2.0.0 Local
Author: wings229

The browser-use command provides fast, persistent browser automation. A background daemon keeps the browser open across commands, giving ~50ms latency per call.

Prerequisites

browser-use doctor    # Verify installation

For setup details, see https://github.com/browser-use/browser-use/blob/main/browser_use/skill_cli/README.md

Core Workflow

Navigate: browser-use open \x3Curl> — starts browser if needed
Inspect: browser-use state — returns clickable elements with indices
Interact: use indices from state (browser-use click 5, browser-use input 3 "text")
Verify: browser-use state or browser-use screenshot to confirm
Repeat: browser stays open between commands
Cleanup: browser-use close when done

Browser Modes

browser-use open \x3Curl>                         # Default: headless Chromium
browser-use --headed open \x3Curl>                # Visible window
browser-use --profile "Default" open \x3Curl>      # Real Chrome with Default profile (existing logins/cookies)
browser-use --profile "Profile 1" open \x3Curl>   # Real Chrome with named profile
browser-use --connect open \x3Curl>               # Auto-discover running Chrome via CDP
browser-use --cdp-url ws://localhost:9222/... open \x3Curl>  # Connect via CDP URL

--connect, --cdp-url, and --profile are mutually exclusive.

Commands

# Navigation
browser-use open \x3Curl>                    # Navigate to URL
browser-use back                          # Go back in history
browser-use scroll down                   # Scroll down (--amount N for pixels)
browser-use scroll up                     # Scroll up
browser-use switch \x3Ctab>                  # Switch to tab by index
browser-use close-tab [tab]              # Close tab (current if no index)

# Page State — always run state first to get element indices
browser-use state                         # URL, title, clickable elements with indices
browser-use screenshot [path.png]         # Screenshot (base64 if no path, --full for full page)

# Interactions — use indices from state
browser-use click \x3Cindex>                 # Click element by index
browser-use click \x3Cx> \x3Cy>                 # Click at pixel coordinates
browser-use type "text"                   # Type into focused element
browser-use input \x3Cindex> "text"          # Click element, then type
browser-use keys "Enter"                  # Send keyboard keys (also "Control+a", etc.)
browser-use select \x3Cindex> "option"       # Select dropdown option
browser-use upload \x3Cindex> \x3Cpath>         # Upload file to file input
browser-use hover \x3Cindex>                 # Hover over element
browser-use dblclick \x3Cindex>              # Double-click element
browser-use rightclick \x3Cindex>            # Right-click element

# Data Extraction
browser-use eval "js code"                # Execute JavaScript, return result
browser-use get title                     # Page title
browser-use get html [--selector "h1"]    # Page HTML (or scoped to selector)
browser-use get text \x3Cindex>              # Element text content
browser-use get value \x3Cindex>             # Input/textarea value
browser-use get attributes \x3Cindex>        # Element attributes
browser-use get bbox \x3Cindex>              # Bounding box (x, y, width, height)

# Wait
browser-use wait selector "css"           # Wait for element (--state visible|hidden|attached|detached, --timeout ms)
browser-use wait text "text"              # Wait for text to appear

# Cookies
browser-use cookies get [--url \x3Curl>]     # Get cookies (optionally filtered)
browser-use cookies set \x3Cname> \x3Cvalue>    # Set cookie (--domain, --secure, --http-only, --same-site, --expires)
browser-use cookies clear [--url \x3Curl>]   # Clear cookies
browser-use cookies export \x3Cfile>         # Export to JSON
browser-use cookies import \x3Cfile>         # Import from JSON

# Python — persistent session with browser access
browser-use python "code"                 # Execute Python (variables persist across calls)
browser-use python --file script.py       # Run file
browser-use python --vars                 # Show defined variables
browser-use python --reset                # Clear namespace

# Session
browser-use close                         # Close browser and stop daemon
browser-use sessions                      # List active sessions
browser-use close --all                   # Close all sessions

The Python browser object provides: browser.url, browser.title, browser.html, browser.goto(url), browser.back(), browser.click(index), browser.type(text), browser.input(index, text), browser.keys(keys), browser.upload(index, path), browser.screenshot(path), browser.scroll(direction, amount), browser.wait(seconds).

Cloud API

browser-use cloud connect                 # Provision cloud browser and connect
browser-use cloud connect --timeout 120 --proxy-country US  # With options
browser-use cloud login \x3Capi-key>         # Save API key (or set BROWSER_USE_API_KEY)
browser-use cloud logout                  # Remove API key
browser-use cloud v2 GET /browsers        # REST passthrough (v2 or v3)
browser-use cloud v2 POST /tasks '{"task":"...","url":"..."}'
browser-use cloud v2 poll \x3Ctask-id>       # Poll task until done
browser-use cloud v2 --help               # Show API endpoints

cloud connect provisions a cloud browser, connects via CDP, and prints a live URL. browser-use close disconnects AND stops the cloud browser.

Tunnels

browser-use tunnel \x3Cport>                 # Start Cloudflare tunnel (idempotent)
browser-use tunnel list                   # Show active tunnels
browser-use tunnel stop \x3Cport>            # Stop tunnel
browser-use tunnel stop --all             # Stop all tunnels

Profile Management

browser-use profile list                  # List detected browsers and profiles
browser-use profile sync --all            # Sync profiles to cloud
browser-use profile update                # Download/update profile-use binary

Command Chaining

Commands can be chained with &&. The browser persists via the daemon, so chaining is safe and efficient.

browser-use open https://example.com && browser-use state
browser-use input 5 "[email protected]" && browser-use input 6 "password" && browser-use click 7

Chain when you don't need intermediate output. Run separately when you need to parse state to discover indices first.

Common Workflows

Authenticated Browsing

When a task requires an authenticated site (Gmail, GitHub, internal tools), use Chrome profiles:

browser-use profile list                           # Check available profiles
# Ask the user which profile to use, then:
browser-use --profile "Default" open https://github.com  # Already logged in

Connecting to Existing Chrome

browser-use --connect open https://example.com     # Auto-discovers Chrome's CDP endpoint

Requires Chrome with remote debugging enabled. Falls back to probing ports 9222/9229.

Exposing Local Dev Servers

browser-use tunnel 3000                            # → https://abc.trycloudflare.com
browser-use open https://abc.trycloudflare.com     # Browse the tunnel

Global Options

Option	Description
`--headed`	Show browser window
`--profile [NAME]`	Use real Chrome (bare `--profile` uses "Default")
`--connect`	Auto-discover running Chrome via CDP
`--cdp-url \x3Curl>`	Connect via CDP URL (`http://` or `ws://`)
`--session NAME`	Target a named session (default: "default")
`--json`	Output as JSON
`--mcp`	Run as MCP server via stdin/stdout

Tips

Always run state first to see available elements and their indices
Use --headed for debugging to see what the browser is doing
Sessions persist — browser stays open between commands
CLI aliases: bu, browser, and browseruse all work

Troubleshooting

Browser won't start? browser-use close then browser-use --headed open \x3Curl>
Element not found? browser-use scroll down then browser-use state
Run diagnostics: browser-use doctor

Cleanup

browser-use close                         # Close browser session
browser-use tunnel stop --all             # Stop tunnels (if any)

安全使用建议

This skill's documentation expects a third-party 'browser-use' CLI and describes sensitive actions (using real Chrome profiles, exporting cookies, starting tunnels, running arbitrary Python), yet the package metadata doesn't declare the binary, env vars, or install steps. Before installing or using: 1) Verify the origin — find the repository/homepage and inspect the 'browser-use' binary/source; do not run unknown binaries. 2) Avoid using --profile Default or importing/exporting cookies unless you understand the privacy risks. 3) Do not set or expose BROWSER_USE_API_KEY unless you trust the cloud provider; prefer ephemeral/test credentials. 4) Consider running this in an isolated VM/container to limit access to your real Chrome profile and files. 5) Ask the publisher for a clear install spec, a checksum for the CLI, and justification for required capabilities (profile access, tunnel, cloud). If the publisher cannot supply provenance, treat the skill as untrusted.

功能分析

Type: OpenClaw Skill Name: browser-use-2-0-0-local Version: 1.0.0 The skill bundle provides a comprehensive CLI wrapper for the 'browser-use' automation tool, granting an AI agent high-risk capabilities including arbitrary Python execution (browser-use python), JavaScript evaluation (browser-use eval), and cookie exfiltration (browser-use cookies export). It explicitly instructs the agent on how to use the '--profile' flag to access the user's primary browser sessions (e.g., 'Default' profile) and how to expose local network ports to the internet using Cloudflare tunnels (browser-use tunnel). While these features are documented as part of the tool's functionality, the combination of local code execution, access to sensitive browser secrets, and network tunneling represents a significant security risk.

能力评估

⚠ Purpose & Capability

The skill claims to automate browsers (reasonable), but the metadata declares no required binaries, no install spec, and no required env vars even though the instructions repeatedly call a 'browser-use' CLI and reference an API key (BROWSER_USE_API_KEY). At minimum the skill should declare that the 'browser-use' binary is required and disclose the optional cloud API key. The omission is incoherent.

⚠ Instruction Scope

SKILL.md instructs the agent to: use real Chrome profiles (--profile Default), export/import cookies to JSON, start Cloudflare tunnels, provision cloud browsers, and execute persistent Python code (browser-use python --file / python REPL). These runtime actions can access sensitive local data (cookies, logged-in sessions, filesystem via script execution) and create external network endpoints. The instructions grant broad discretion (arbitrary JS eval, arbitrary Python files) beyond simple form-filling or screenshots.

ℹ Install Mechanism

There is no install spec (instruction-only), which is lower-risk in itself, but it means the skill implicitly depends on a preinstalled 'browser-use' binary. That implicit dependency should be declared; otherwise agents/users may run commands that don't exist or assume a particular third-party binary is present without verifying provenance.

⚠ Credentials

The documentation references saving an API key (BROWSER_USE_API_KEY) for cloud operations and also exposing Chrome profiles and cookies. The skill metadata lists no required env vars or config paths. Requesting access to profile data, cookies, filesystem paths (for uploads and python scripts), or an API key is sensitive; these are not declared or justified in the metadata, which is disproportionate and inconsistent.

ℹ Persistence & Privilege

always is false and the skill is user-invocable (normal). However, the SKILL.md describes commands that persist state (a background daemon, persistent Python sessions, saved cloud API keys, tunnels). Combined with the unanswered capability/information gaps, these persistent behaviors increase risk if the binary is present and the agent is allowed to run commands autonomously.

版本历史

v1.0.0

Initial release of browser-use-2-0-0-local - Automates browser tasks: navigation, form filling, screenshots, and data extraction. - Provides fast, persistent browser control using a background daemon for low-latency automation. - Supports multiple browser modes including headless Chromium, headed Chrome with real user profiles, and remote debugging connections. - Includes commands for web interactions, state inspection, file uploads, cookie management, and Python scripting within the same browser session. - Adds cloud browser provisioning, session handling, and easy tunneling for local development. - Workflow documentation and usage examples provided in SKILL.md.

元数据

Slug browser-use-2-0-0-local

版本 1.0.0

许可证 MIT-0

累计安装 0

当前安装数 0

历史版本数 1

常见问题

Browser Use 2.0.0 Local 是什么？

Automates browser interactions for web testing, form filling, screenshots, and data extraction. Use when the user needs to navigate websites, interact with w... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 163 次。

如何安装 Browser Use 2.0.0 Local？

在 OpenClaw 或 Claude Code 对话框中运行命令「/install browser-use-2-0-0-local」即可一键安装，无需额外配置。

Browser Use 2.0.0 Local 是免费的吗？

是的，Browser Use 2.0.0 Local 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。

Browser Use 2.0.0 Local 支持哪些平台？

Browser Use 2.0.0 Local 跨平台运行，可在任意部署了 OpenClaw / Claude Code 的环境中使用（cross-platform）。

谁开发了 Browser Use 2.0.0 Local？

由 Wings229（@wings229）开发并维护，当前版本 v1.0.0。

Browser Use 2.0.0 Local