sleepinginsummer/agent-browser-cli

使用 agent-browser-cli 进行浏览器感知与控制。适用于标签页扫描/切换、页面 JS 执行、Cookie、CDP、contentSettings、截图、文件上传、下拉框点击、tmwd_cdp_bridge 初始化和 Web 工具排障

What it solves

It provides a way for AI Agents to perceive and control a real Chrome browser session. Unlike headless automation tools like Selenium or Playwright, this tool allows Agents to operate within a user's actual browser, meaning they can reuse existing login states, cookies, and active sessions without needing to re-authenticate.

How it works

The system consists of a Rust-based CLI and a Chrome extension (tmwd_cdp_bridge). The extension connects the real browser to the CLI via a WebSocket bridge. This setup enables the CLI to send commands to the browser to execute JavaScript, interact with DOM elements, and utilize the Chrome DevTools Protocol (CDP) for low-level control.

Who it’s for

  • AI Agent Developers: Who need their agents to perform tasks on websites that require authentication or complex session management.
  • QA Engineers: For automating tests in a real user environment.
  • Frontend Developers: For debugging page states, styles, and DOM structures using a CLI.
  • Data Collectors: For scraping data from pages that require a logged-in state.

Highlights

  • Session Reuse: Maintains login states and cookies from the actual user session.
  • Comprehensive Control: Supports tab scanning, JS execution, CDP calls, screenshots, and file uploads.
  • Agent-Ready: Includes a specific SKILL.md to help AI agents understand how to use the tool.
  • Cross-Platform: Works on Windows (including WSL), macOS, and Linux.
  • Page Simplification: Includes scripts to simplify HTML and identify list structures to help agents process pages more efficiently.

Related

  • Project
  • Project
  • Project
  • Project
  • Project