Agent Browser Automation
Automates full-featured browser interactions for web testing, data extraction, and form filling directly within the Claude Code environment.
Agent Browser is a powerful skill for Claude Code that enables AI agents to navigate the web, interact with complex DOM elements, and execute automated workflows. It utilizes a robust reference-based system to target elements discovered via snapshots, allowing for reliable clicking, typing, and form submission. With built-in support for authentication state management, parallel sessions, and visual captures, it serves as an essential tool for developers performing end-to-end testing, visual regression, or complex data collection across dynamic web applications.
主要功能
01High-resolution visual capture with full-page and element-specific screenshots
020 GitHub stars
03Snapshot-based interactive element targeting using unique @refs
04Comprehensive interaction suite including click, fill, hover, and scroll
05Support for parallel browser sessions and semantic locator-based searching
06Authentication state persistence for managing secure, logged-in sessions
使用场景
01Extracting structured data from dynamic, JavaScript-heavy websites
02Automating end-to-end testing and QA workflows for web applications
03Automating repetitive web-based administrative tasks and form submissions