Web Browse

0•

Enables Claude to navigate, interact with, and extract data from websites using a headless browser with advanced automation capabilities.

The Web Browse skill provides Claude with a robust headless browser interface for complex web automation tasks. It handles high-level actions such as navigating authentication walls, interacting with single-page applications (SPAs), and executing UI macros like multi-page pagination or form filling by label. With built-in detection for content blocking and support for both headless and headed sessions for CAPTCHA solving, it allows Claude to perform comprehensive web research, site testing, and automated data collection while maintaining a structured understanding of page layouts via accessibility trees.

主要功能

01Accessibility tree snapshots for accurate, AI-friendly understanding of page structures.
02Headed checkpoint mode for manual user interaction, CAPTCHA resolution, and secure logins.
03Built-in content-blocking detection with automatic recovery workflows for headless-resistant sites.
04Advanced headless browser control with automatic authentication wall detection and headed fallback.
050 GitHub stars
06Comprehensive UI macros for complex actions like pagination, modal dismissal, and date picking.

使用场景

01Automated web research and structured data extraction from complex, interactive websites.
02Navigating and interacting with internal tools or SaaS platforms requiring authentication or file uploads.
03End-to-end testing of web applications, including form submissions and dynamic UI interactions.