Agent Browser Automation

•
2
•

Automates headless and headed browser interactions for web navigation, form filling, testing, and structured data extraction.

Agent Browser equips Claude Code with comprehensive browser automation capabilities powered by a dedicated command-line interface. It allows agents to inspect web pages via accessibility snapshots, interact with interactive elements using dynamic reference handles, fill out complex forms, manage cookies and persistent sessions, and intercept network traffic. Ideal for end-to-end web testing, authenticated data extraction, taking full-page screenshots, and recording video walkthroughs, this skill bridges autonomous agents and modern web applications seamlessly.

Características Principales

01Complete browser action controls including clicks, typing, drag-and-drop, and file uploads
02Network interception, mock routing, proxy configuration, and device emulation
03Interactive element ref snapshotting for precise and robust DOM targeting
04Advanced session handling with state persistence, parallel sessions, and CDP connections
052 GitHub stars
06Visual capture suite supporting full-page screenshots, PDF export, and video recording

Casos de Uso

01Generating automated visual regression captures and interactive video demo recordings
02Automating multi-step form submissions and testing web applications end-to-end
03Extracting structured content and scraping data behind authentication walls