← 返回 Skills 市场
635
总下载
0
收藏
2
当前安装
1
版本数
在 OpenClaw 中安装
/install browser-automation-ctrl
功能描述
通过 Selenium 自动化控制浏览器，支持网页打开、元素操作、标签页管理、截图、JS 执行及代理设置等功能。
使用说明 (SKILL.md)
Browser Control Skill
浏览器控制工具 - 通过 Selenium 实现浏览器自动化控制。
功能列表 (35+ 命令)
基础操作
| 命令 | 说明 |
|---|---|
open \x3Curl> |
打开网页 |
screenshot |
网页截图 |
screenshotb64 |
网页截图(base64) |
fullscreen |
全页面截图 |
js \x3Cscript> |
执行 JavaScript |
cookies |
获取 Cookies |
setcookie \x3Cname> \x3Cvalue> |
设置 Cookie |
clearcookies |
清除 Cookies |
source |
获取页面源码 |
title |
获取标题 |
url |
获取当前 URL |
元素操作
| 命令 | 说明 |
|---|---|
find \x3Cselector> |
查找元素 |
click \x3Cselector> |
点击元素 |
fill \x3Cselector> \x3Ctext> |
填写输入框 |
submit \x3Cselector> |
提交表单 |
hover \x3Cselector> |
悬停 |
attr \x3Cselector> \x3Cname> |
获取元素属性 |
导航操作
| 命令 | 说明 |
|---|---|
back |
后退 |
forward |
前进 |
refresh |
刷新 |
wait \x3Cseconds> |
等待 |
scroll [x] [y] |
滚动页面 |
scrollto \x3Cselector> |
滚动到元素 |
waitelem \x3Cselector> |
等待元素出现 |
标签页操作
| 命令 | 说明 |
|---|---|
newtab \x3Curl> |
新建标签页 |
switchtab \x3Cindex> |
切换标签页 |
closetab |
关闭标签页 |
tabs |
获取所有标签页 |
高级功能
| 命令 | 说明 |
|---|---|
links |
获取所有链接 |
images |
获取所有图片 |
size |
获取页面尺寸 |
ua \x3Cuser-agent> |
设置 User-Agent |
proxy \x3Caddress> |
设置代理 |
依赖安装
pip install selenium
需要 Chrome 浏览器和 ChromeDriver。
使用示例
# 打开网页并截图
python scripts/browser_ctrl.py open https://www.baidu.com
python scripts/browser_ctrl.py screenshot
# 查找元素并填写
python scripts/browser_ctrl.py find "#kw"
python scripts/browser_ctrl.py fill "#kw" "搜索内容"
python scripts/browser_ctrl.py click "#su"
# 获取页面所有链接
python scripts/browser_ctrl.py links
# 设置代理
python scripts/browser_ctrl.py proxy "http://127.0.0.1:8080"
# 设置 User-Agent
python scripts/browser_ctrl.py ua "Mozilla/5.0..."
触发关键词
- "打开网页"、"浏览器"
- "网页截图"
- "执行 JS"
- "点击"、"填写"
- "前进"、"后退"
- "获取链接"、"获取图片"
- "设置代理"
注意事项
- 需要安装 Chrome 浏览器
- 需要 chromedriver 与 Chrome 版本匹配
- 无头模式运行（不显示浏览器窗口）
- 截图保存在 ~/Pictures/OpenClaw/
- 支持 CSS 选择器和 XPath
安全使用建议
This package appears to be a straightforward Selenium browser automation helper. Before installing: (1) ensure you want a tool that can open arbitrary URLs and save screenshots to ~/Pictures/OpenClaw (it will create that directory); (2) be aware it requires a matching Chrome + chromedriver and 'pip install selenium'; (3) avoid giving it untrusted URLs or sensitive credentials — automated browsing can load malicious pages or capture sensitive on‑screen content in screenshots; (4) if you allow autonomous agent invocation, the agent could run browser actions without further prompts, so restrict usage or keep invocation manual if you want tighter control.
功能分析
Type: OpenClaw Skill
Name: browser-automation-ctrl
Version: 1.0.0
The skill bundle provides powerful browser automation capabilities via Selenium, including arbitrary JavaScript execution (`js` command), session cookie extraction (`cookies`), and proxy configuration (`proxy`). While these features are aligned with the stated purpose in `SKILL.md`, they represent a high-risk capability set that could be easily leveraged for data exfiltration or session hijacking if the agent is manipulated. The script `scripts/browser_ctrl.py` includes anti-detection measures (CDP commands to hide the webdriver attribute) and allows the agent to capture and return full page sources and cookies to the caller, which are significant security risks in an agentic environment.
能力评估
Purpose & Capability
Name, description, SKILL.md and the included Python script all describe Selenium-based browser automation (open pages, click, fill, screenshot, proxy, UA). There are no unrelated dependencies, credentials, or config paths requested.
Instruction Scope
Runtime instructions and code focus on browser automation. The SKILL.md tells users to install selenium and run the script; the script implements the documented commands. It does create screenshots under ~/Pictures/OpenClaw and requires Chrome/Chromedriver, which is consistent with its purpose.
Install Mechanism
No install spec is provided (instruction-only), and SKILL.md recommends 'pip install selenium' and installing Chrome/Chromedriver. No remote downloads or obscure install URLs are present in the package.
Credentials
The skill requests no environment variables, no credentials, and no config paths. This is proportional to a local browser automation utility.
Persistence & Privilege
always is false and the skill is user-invocable. It does not request persistent system-wide changes or modify other skills’ configs. Note: disable-model-invocation is false (agent may call it autonomously), which is the platform default.
如何使用
- 确保已安装 OpenClaw（本地或 Docker 部署）
- 在对话框中输入安装命令：
/install browser-automation-ctrl - 安装完成后，直接呼叫该 Skill 的名称或使用
/browser-automation-ctrl触发 - 根据 Skill 的参数说明提供必要输入，即可获得结构化输出
版本历史
v1.0.0
- Major update: Rebuilt skill to provide browser control via Selenium automation instead of remote VNC/Google OAuth access.
- Simplified implementation to a single Python script (browser_ctrl.py); removed all shell script management and OAuth setup.
- Added 35+ browser automation commands for navigation, element interaction, screenshots, cookies, tabs, advanced features, and more.
- Updated documentation to list all available commands, usage examples, setup (Python, Selenium, chromedriver), and trigger phrases.
- No longer relies on external tunnel or user’s Google authentication; actions are run directly via local headless Chrome.
元数据
常见问题
Selenium Browser Control 是什么？
通过 Selenium 自动化控制浏览器，支持网页打开、元素操作、标签页管理、截图、JS 执行及代理设置等功能。 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 635 次。
如何安装 Selenium Browser Control？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install browser-automation-ctrl」即可一键安装，无需额外配置。
Selenium Browser Control 是免费的吗？
是的，Selenium Browser Control 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。
Selenium Browser Control 支持哪些平台？
Selenium Browser Control 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 Selenium Browser Control？
由 楊小虎（@tr0812）开发并维护，当前版本 v1.0.0。
推荐 Skills