Browser Automation

•
13k
•
소유권 확인됨
관리

Automates web navigation and interaction using AI-optimized accessibility snapshots to significantly reduce token usage.

The Browser Automation skill enables Claude agents to perform complex web-based tasks by utilizing a specialized agent-browser that translates web pages into lightweight accessibility trees. By replacing bulky DOM structures with compact element references, it reduces context window consumption by up to 93%, allowing agents to navigate, click, fill forms, and scrape data with high precision and lower costs. It is ideal for developers building autonomous workflows that require interacting with real-world websites, managing multiple browser sessions, and capturing visual feedback via screenshots.

주요 기능

01Interactive-only filtering to focus the AI on actionable UI elements
02Multi-session support for parallel agent workflows and state persistence
03AI-optimized accessibility snapshots with 93% context reduction
04Simplified element reference system (@e1, @e2) for deterministic interactions
05Advanced wait conditions for dynamic content and network idle states
0612,774 GitHub stars

사용 사례

01Automated end-to-end testing and visual verification of web applications
02Automating multi-step authentication and dashboard navigation flows
03Efficient web scraping and data extraction from complex SPAs