Chrome CDP Controller — Chrome CDP 控制器

Name: Chrome CDP Controller — Chrome CDP 控制器
Author: xiaoxiaoxi

xiaoxiaoxi

Chrome CDP Controller — Chrome CDP 控制器

v1.0.0

Chrome CDP 控制器工具。

0· 74·0 当前·0 累计

by @hanxiaolin (xiaoxiaoxi)·MIT-0

开发工具浏览器自动化自动化

下载技能包

License

MIT-0

最后更新

2026/3/30

安全扫描

VirusTotal

可疑

查看报告

OpenClaw

安全

high confidence

The skill is internally consistent with its stated purpose (controlling a local Chrome instance via CDP with Puppeteer); nothing in the files or instructions requests unrelated credentials or installs unexpected remote code.

评估建议

This skill appears to be what it says: a Puppeteer-based controller for a local Chrome instance. Before installing or running it, consider these precautions: - Only connect it to browsers you control. Running Chrome with remote debugging exposes a WebSocket that this tool will use; ensure the debugging port is bound to localhost and not reachable by others. - Audit any commands/JSON files you pass to the script. Commands can (a) upload local files you specify, (b) read and execute JavaScript fil...

详细分析 ▾

✓ 用途与能力

Name/description match the implementation: the package depends on puppeteer-core, the SKILL.md explains connecting to a local CDP WebSocket, and scripts/cdp_controller.js implements navigation, interaction, evaluation, uploads, screenshots and network interception as described. No extraneous credentials, binaries, or unrelated services are required.

ℹ 指令范围

Instructions stay within browser automation scope, but include functionality that can access local files (upload() verifies filesystem paths; the 'evaluate' command supports executing JS from a file by reading it from disk) and capture network responses (start_intercept -> in-memory storage retrievable via get_intercepted). Those capabilities are expected for this tool but can be abused to exfiltrate sensitive data if untrusted command files or commands are executed. SKILL.md also instructs starting Chrome with remote debugging on localhost (expected).

✓ 安装机制

No registry install spec in the metadata; the project is instruction-only but includes a package.json and expects the user to run `npm install` which will fetch puppeteer-core from the npm registry. Dependencies resolve to normal npm packages and no ad-hoc download URLs or archive extraction from untrusted hosts are present.

✓ 凭证需求

The skill requests no environment variables or external credentials. It connects to a user-provided local CDP WebSocket URL only. Note: the code can read local file paths supplied as command parameters (for upload and evaluate-from-file), so sensitive files on disk could be accessed if commands include them.

✓ 持久化与权限

Skill is not always-enabled, does not request persistent platform-level privileges, and does not modify other skills or global agent configuration. It is run on-demand and requires an explicit WebSocket URL to connect to the browser.

⚠ scripts/cdp_controller.js:167

Dynamic code execution detected.

⚠ scripts/cdp_controller.js:278

File read combined with network send (possible exfiltration).

安全有层次，运行前请审查代码。

License

MIT-0

可自由使用、修改和再分发，无需署名。

查看条款 ↗

运行时依赖

无特殊依赖

版本

latestv1.0.02026/3/30

Initial release: control a running Chrome via CDP using Puppeteer. Includes navigate/click/fill/evaluate/screenshot/network intercept.

● 可疑

安装命令点击复制

官方npx clawhub@latest install chrome-cdp-controller

镜像加速npx clawhub@latest install chrome-cdp-controller --registry https://cn.clawhub-mirror.com

技能文档

Control and automate a Chrome browser that's already running with CDP enabled, using Puppeteer.

键 Features

Non-intrusive: Connects 到 existing Chrome, opens 新的标签页, operates, 然后 closes 仅标签页
browser stays 打开: Never closes existing tabs 或 browser
Clean automation: 每个 task runs 在...中 own 标签页, 哪个 automatically closed 当...时已完成

1. Chrome 带有 CDP 已启用

Chrome must be running with remote debugging enabled. If not already started:

# macOS /Applications/Google\ Chrome.app/Contents/MacOS/Google\ Chrome --remote-debugging-port=9222 & # Windows "C:\Program Files\Google\Chrome\Application\chrome.exe" --remote-debugging-port=9222

# Linux google-chrome --remote-debugging-port=9222 &

2. 获取 WebSocket URL

Visit http://localhost:9222/json/version to get the webSocketDebuggerUrl, for example:

ws://127.0.0.1:9222/devtools/browser/FFA7276F8E8E51645BD2AC9BE6B79607

Or use this one-liner:

curl -s http://localhost:9222/json/version | grep -o '"webSocketDebuggerUrl":"[^"]"' | cut -d'"' -f4

3. Install Dependencies

cd chrome-cdp-controller
npm install

This installs puppeteer-core which is lightweight (doesn't download Chromium).

Usage

Command-Based Execution

Create a JSON file with commands:

commands.json:

[
  {"type": "navigate", "url": "https://www.baidu.com"},
  {"type": "wait", "seconds": 2},
  {"type": "screenshot", "path": "/tmp/screenshot.png"},
  {"type": "evaluate", "script": "document.title"}
]

Execute:

node scripts/cdp_controller.js --ws "ws://127.0.0.1:9222/devtools/browser/..." --commands commands.json

节点.js API

const { CDPController } = require('./scripts/cdp_controller.js');(async () => {
  const controller = new CDPController('ws://127.0.0.1:9222/devtools/browser/...');
  await controller.connect();
  
  // Navigate
  await controller.navigate('https://www.taobao.com');
  
  // Fill form
  await controller.fill('#q', 'iPhone');
  await controller.press('Enter');
  await controller.wait(3);
  
  // Extract data
  const result = await controller.evaluate(
    Array.from(document.querySelectorAll('.item'))
      .slice(0, 10)
      .map(item => ({
        title: item.querySelector('.title')?.textContent.trim(),
        price: item.querySelector('.price')?.textContent.trim()
      }))
  );
  console.log(result.result);
  
  // Intercept network
  await controller.startIntercept('api');
  // ... perform actions ...
  const responses = controller.getInterceptedResponses();
  
  await controller.close();
})();

可用 Commands

navigate - Go 到 URL

- url: Target URL - waitUntil: "加载", "domcontentloaded", 或 "networkidle2" (默认)
Interaction

click - Click 元素

- selector: CSS selector - 超时: 超时在...中 milliseconds (默认: 5000)
fill - Fill 表单字段 (selects 所有, 然后 types)

- selector: CSS selector - text: Text 到 fill - 超时: 超时在...中 milliseconds (默认: 5000)
类型 - 类型 text character 由 character

- selector: CSS selector - text: Text 到类型 - 延迟: 延迟之间 characters 在...中 ms (默认: 50) - 超时: 超时在...中 milliseconds (默认: 5000)
press - Press 键

- 键: 键 name (Enter, 标签页, 转义, etc.)
Data Extraction

get_text - 获取 text content 的 single 元素

- selector: CSS selector
get_all_text - 获取 text content 的所有 matching elements

- selector: CSS selector
evaluate - Execute JavaScript 和 return 结果

- script: JavaScript code
Network Interception

start_intercept - 开始 intercepting network responses

- url_pattern: URL pattern 到 match (substring match, e.g., "api")
get_intercepted - 获取所有 intercepted responses
clear_intercepted - 清除 intercepted responses 列表
Utilities

screenshot - Take screenshot

- path: 输出 file path - fullPage: Capture 满 page (默认: 假)
wait_for_selector - Wait 对于元素到 appear

- selector: CSS selector - 超时: 超时在...中 milliseconds (默认: 5000)
wait - Sleep 对于持续时间

- seconds: 数字的 seconds 到 wait
Common Workflows
示例 1: 搜索 Taobao 对于 iPhone Prices
taobao-搜索.json:

[
  {"type": "navigate", "url": "https://www.taobao.com"},
  {"type": "wait", "seconds": 2},
  {"type": "fill", "selector": "#q", "text": "iPhone"},
  {"type": "press", "key": "Enter"},
  {"type": "wait", "seconds": 5},
  {"type": "evaluate", "script": "Array.from(document.querySelectorAll('.item, [class=\"Item\"]')).slice(0, 10).map(item => ({ title: item.querySelector('.title, [class=\"title\"]')?.textContent.trim().substring(0, 80), price: item.querySelector('.price, [class=\"price\"]')?.textContent.trim() }))"}
]

Run:

WS_URL=$(curl -s http://localhost:9222/json/version | grep -o '"webSocketDebuggerUrl":"[^"]*"' | cut -d'"' -f4)
node scripts/cdp_controller.js --ws "$WS_URL" --commands taobao-search.json

Note: Selectors 可能 vary. 使用 browser DevTools (F12) 到 inspect elements.

示例 2: Ask ChatGPT Question

chatgpt-ask.json:

[
  {"type": "navigate", "url": "https://chat.openai.com"},
  {"type": "wait_for_selector", "selector": "textarea", "timeout": 10000},
  {"type": "type", "selector": "textarea", "text": "What is artificial intelligence?"},
  {"type": "press", "key": "Enter"},
  {"type": "wait", "seconds": 10},
  {"type": "get_text", "selector": "[data-message-author-role='assistant']:last-of-type"}
]

示例 3: Intercept API Responses

intercept-api.json:

[
  {"type": "start_intercept", "url_pattern": "graphql"},
  {"type": "navigate", "url": "https://example.com"},
  {"type": "wait", "seconds": 3},
  {"type": "get_intercepted"}
]

For more examples, see references/examples.md.

Workflow

When automating browser tasks:

获取 WebSocket URL:

- 如果您已经有 (e.g., ws://127.0.0.1:9222/devtools/browser/...), 使用 directly - 否则, 获取从 http://localhost:9222/json/version

Determine selectors:

- 对于 known sites (Taobao, ChatGPT, etc.), check references/examples.md - 对于 unknown sites, navigate 第一个和使用 evaluate 带有 document.querySelector 到 test selectors - 使用 browser DevTools (F12) 到 inspect elements

Build command sequence:

- 开始带有 navigate - 添加 wait 或 wait_for_selector 之间 steps - 使用 fill 对于表单 inputs, click 对于 buttons - 使用 evaluate 对于 complex data extraction - 添加 start_intercept 之前导航如果 monitoring network traffic

Execute:

- 保存 commands 到 JSON file - Run: 节点 scripts/cdp_controller.js --ws "" --commands - 解析 JSON 输出

Handle failures:

- 如果 selectors 失败, inspect 带有 DevTools - 如果 timing issues occur, increase wait 乘以或使用 wait_for_selector - 如果连接 fails, 验证 Chrome running 带有 CDP 已启用

Tips

Wait 乘以: 使用 wait_for_selector 代替的 fixed wait 当...时 possible
Selectors: Prefer ID (#id) > 类 (.类) > 属性 ([attr='值'])
Network interception: 开始 intercepting 之前 navigating
JavaScript: 使用 evaluate 对于 complex data extraction
Screenshots: Useful 对于 debugging

Troubleshooting

连接失败

Ensure Chrome running 带有 --remote-debugging-port=9222
验证 WebSocket URL 正确
Check Chrome version compatibility 带有 puppeteer-core

元素不 Found

验证 selector 带有 DevTools (F12)
添加 wait_for_selector 之前 interaction
Check 如果元素在...中 iframe

模块不 Found

cd chrome-cdp-controller
npm install

Control and automate a Chrome browser that's already running with CDP enabled, using Puppeteer.

Key Features

Non-intrusive: Connects to your existing Chrome, opens a new tab, operates, then closes only that tab
Your browser stays open: Never closes your existing tabs or browser
Clean automation: Each task runs in its own tab, which is automatically closed when done

1. Chrome with CDP Enabled

Chrome must be running with remote debugging enabled. If not already started:

# macOS /Applications/Google\ Chrome.app/Contents/MacOS/Google\ Chrome --remote-debugging-port=9222 & # Windows "C:\Program Files\Google\Chrome\Application\chrome.exe" --remote-debugging-port=9222

# Linux google-chrome --remote-debugging-port=9222 &

2. Get WebSocket URL

Visit http://localhost:9222/json/version to get the webSocketDebuggerUrl, for example:

ws://127.0.0.1:9222/devtools/browser/FFA7276F8E8E51645BD2AC9BE6B79607

Or use this one-liner:

curl -s http://localhost:9222/json/version | grep -o '"webSocketDebuggerUrl":"[^"]"' | cut -d'"' -f4

3. Install Dependencies

cd chrome-cdp-controller
npm install

This installs puppeteer-core which is lightweight (doesn't download Chromium).

Usage

Command-Based Execution

Create a JSON file with commands:

commands.json:

[
  {"type": "navigate", "url": "https://www.baidu.com"},
  {"type": "wait", "seconds": 2},
  {"type": "screenshot", "path": "/tmp/screenshot.png"},
  {"type": "evaluate", "script": "document.title"}
]

Execute:

node scripts/cdp_controller.js --ws "ws://127.0.0.1:9222/devtools/browser/..." --commands commands.json

Node.js API

const { CDPController } = require('./scripts/cdp_controller.js');(async () => {
  const controller = new CDPController('ws://127.0.0.1:9222/devtools/browser/...');
  await controller.connect();
  
  // Navigate
  await controller.navigate('https://www.taobao.com');
  
  // Fill form
  await controller.fill('#q', 'iPhone');
  await controller.press('Enter');
  await controller.wait(3);
  
  // Extract data
  const result = await controller.evaluate(
    Array.from(document.querySelectorAll('.item'))
      .slice(0, 10)
      .map(item => ({
        title: item.querySelector('.title')?.textContent.trim(),
        price: item.querySelector('.price')?.textContent.trim()
      }))
  );
  console.log(result.result);
  
  // Intercept network
  await controller.startIntercept('api');
  // ... perform actions ...
  const responses = controller.getInterceptedResponses();
  
  await controller.close();
})();

Available Commands

Navigation

navigate - Go to URL

- url: Target URL - waitUntil: "load", "domcontentloaded", or "networkidle2" (default)
Interaction

click - Click an element

- selector: CSS selector - timeout: Timeout in milliseconds (default: 5000)
fill - Fill a form field (selects all, then types)

- selector: CSS selector - text: Text to fill - timeout: Timeout in milliseconds (default: 5000)
type - Type text character by character

- selector: CSS selector - text: Text to type - delay: Delay between characters in ms (default: 50) - timeout: Timeout in milliseconds (default: 5000)
press - Press a key

- key: Key name (Enter, Tab, Escape, etc.)
Data Extraction

get_text - Get text content of a single element

- selector: CSS selector
get_all_text - Get text content of all matching elements

- selector: CSS selector
evaluate - Execute JavaScript and return result

- script: JavaScript code
Network Interception

start_intercept - Start intercepting network responses

- url_pattern: URL pattern to match (substring match, e.g., "api")
get_intercepted - Get all intercepted responses
clear_intercepted - Clear intercepted responses list
Utilities

screenshot - Take a screenshot

- path: Output file path - fullPage: Capture full page (default: false)
wait_for_selector - Wait for element to appear

- selector: CSS selector - timeout: Timeout in milliseconds (default: 5000)
wait - Sleep for a duration

- seconds: Number of seconds to wait
Common Workflows
Example 1: Search Taobao for iPhone Prices
taobao-search.json:

[
  {"type": "navigate", "url": "https://www.taobao.com"},
  {"type": "wait", "seconds": 2},
  {"type": "fill", "selector": "#q", "text": "iPhone"},
  {"type": "press", "key": "Enter"},
  {"type": "wait", "seconds": 5},
  {"type": "evaluate", "script": "Array.from(document.querySelectorAll('.item, [class=\"Item\"]')).slice(0, 10).map(item => ({ title: item.querySelector('.title, [class=\"title\"]')?.textContent.trim().substring(0, 80), price: item.querySelector('.price, [class=\"price\"]')?.textContent.trim() }))"}
]

Run:

WS_URL=$(curl -s http://localhost:9222/json/version | grep -o '"webSocketDebuggerUrl":"[^"]*"' | cut -d'"' -f4)
node scripts/cdp_controller.js --ws "$WS_URL" --commands taobao-search.json

Note: Selectors may vary. Use browser DevTools (F12) to inspect elements.

Example 2: Ask ChatGPT a Question

chatgpt-ask.json:

[
  {"type": "navigate", "url": "https://chat.openai.com"},
  {"type": "wait_for_selector", "selector": "textarea", "timeout": 10000},
  {"type": "type", "selector": "textarea", "text": "What is artificial intelligence?"},
  {"type": "press", "key": "Enter"},
  {"type": "wait", "seconds": 10},
  {"type": "get_text", "selector": "[data-message-author-role='assistant']:last-of-type"}
]

Example 3: Intercept API Responses

intercept-api.json:

[
  {"type": "start_intercept", "url_pattern": "graphql"},
  {"type": "navigate", "url": "https://example.com"},
  {"type": "wait", "seconds": 3},
  {"type": "get_intercepted"}
]

For more examples, see references/examples.md.

Workflow

When automating browser tasks:

Get WebSocket URL:

- If you already have it (e.g., ws://127.0.0.1:9222/devtools/browser/...), use it directly - Otherwise, fetch from http://localhost:9222/json/version

Determine selectors:

- For known sites (Taobao, ChatGPT, etc.), check references/examples.md - For unknown sites, navigate first and use evaluate with document.querySelector to test selectors - Use browser DevTools (F12) to inspect elements

Build command sequence:

- Start with navigate - Add wait or wait_for_selector between steps - Use fill for form inputs, click for buttons - Use evaluate for complex data extraction - Add start_intercept before navigation if monitoring network traffic

Execute:

- Save commands to JSON file - Run: node scripts/cdp_controller.js --ws "" --commands - Parse JSON output

Handle failures:

- If selectors fail, inspect with DevTools - If timing issues occur, increase wait times or use wait_for_selector - If connection fails, verify Chrome is running with CDP enabled

Tips

Wait times: Use wait_for_selector instead of fixed wait when possible
Selectors: Prefer ID (#id) > class (.class) > attribute ([attr='value'])
Network interception: Start intercepting BEFORE navigating
JavaScript: Use evaluate for complex data extraction
Screenshots: Useful for debugging

Troubleshooting

Connection Failed

Ensure Chrome is running with --remote-debugging-port=9222
Verify WebSocket URL is correct
Check Chrome version compatibility with puppeteer-core

Element Not Found

Verify selector with DevTools (F12)
Add wait_for_selector before interaction
Check if element is in an iframe

Module Not Found

cd chrome-cdp-controller
npm install

数据来源：ClawHub ↗ · 中文优化：龙虾技能库

OpenClaw 技能定制 / 插件定制 / 私有工作流定制

免费技能或插件可能存在安全风险，如需更匹配、更安全的方案，建议联系付费定制

了解定制服务

License

运行时依赖

版本

安装命令 点击复制

技能文档

键 Features

1. Chrome 带有 CDP 已启用

2. 获取 WebSocket URL

3. Install Dependencies

Usage

Command-Based Execution

节点.js API

可用 Commands

导航

Interaction

Data Extraction

Network Interception

Utilities

Common Workflows

示例 1: 搜索 Taobao 对于 iPhone Prices

示例 2: Ask ChatGPT Question

示例 3: Intercept API Responses

Workflow

Tips

Troubleshooting

连接 失败

元素 不 Found

模块 不 Found

Key Features

1. Chrome with CDP Enabled

2. Get WebSocket URL

3. Install Dependencies

Usage

Command-Based Execution

Node.js API

Available Commands

Navigation

Interaction

Data Extraction

Network Interception

Utilities

Common Workflows

Example 1: Search Taobao for iPhone Prices

Example 2: Ask ChatGPT a Question

Example 3: Intercept API Responses

Workflow

Tips

Troubleshooting

Connection Failed

Element Not Found

Module Not Found

安装命令点击复制

连接失败

元素不 Found

模块不 Found