Is my browser data or credentials shared with the AI?

No, all operations run locally on your machine. Only the specific data extracted by the scripts (like page titles or selected DOM elements) is shared with the agent.

Can I run custom JavaScript within the browser context?

Yes, the eval.js script allows you to execute single expressions or IIFEs directly within the browser, returning the results as structured JSON.

How does Claude interact with specific elements on a page?

The skill includes a 'pick' utility that allows the agent to visually select elements and return detailed metadata including IDs, classes, and CSS selectors.

Do I need an MCP server to use this browser skill?

No, this skill uses minimal local scripts to communicate directly with Chrome via the DevTools Protocol, removing the need for an external MCP server setup.

Can I use this with my existing browser login sessions?

Yes, by using the --profile flag during launch, you can use a persistent Chrome profile that preserves your cookies, history, and authentication tokens.

Browser Automation & CDP Controller

Name: Browser Automation & CDP Controller
Author: cexll

bycexll

•

1,960

浏览器自动化

Automates Chrome browsers using the DevTools Protocol to navigate pages, execute JavaScript, and capture visual data without requiring an MCP server.

关于

This skill enables Claude to directly interact with a local Chrome instance via the Chrome DevTools Protocol (CDP), streamlining web-based tasks without the complexity of an external MCP server. It provides a suite of scripts for launching Chrome with persistent profiles, navigating URLs, executing custom JavaScript in the browser context, and capturing screenshots or specific DOM element metadata. It is particularly useful for developers needing to automate web workflows, scrape dynamic content, or perform visual inspections of web applications directly from their local environment while maintaining full control over session data.

主要功能

1,960 GitHub stars
Visual element picking and automated screenshot capture for UI analysis
In-browser JavaScript execution with structured JSON output for agent consumption
Seamless navigation and multi-tab management directly from the terminal
Persistent profile support to maintain session cookies and authenticated states
Direct Chrome DevTools Protocol (CDP) integration for low-latency browser control

使用场景

Generating visual snapshots and inspecting DOM elements for frontend debugging and UI testing
Automated web scraping of dynamic, JavaScript-heavy websites and SPAs
Performing authenticated browser tasks like session-based data extraction and form automation

关于

主要功能

1,960 GitHub stars
Visual element picking and automated screenshot capture for UI analysis
In-browser JavaScript execution with structured JSON output for agent consumption
Seamless navigation and multi-tab management directly from the terminal
Persistent profile support to maintain session cookies and authenticated states
Direct Chrome DevTools Protocol (CDP) integration for low-latency browser control

使用场景

Generating visual snapshots and inspecting DOM elements for frontend debugging and UI testing
Automated web scraping of dynamic, JavaScript-heavy websites and SPAs
Performing authenticated browser tasks like session-based data extraction and form automation