Can I maintain a login state across different commands?

Yes, by using the --session flag, you can maintain a persistent browser state, including cookies and local storage, across multiple interactions.

How does Agent Browser identify page elements?

Agent Browser uses an accessibility tree snapshot to assign unique references (like @e1, @e2) to interactive elements. This allows the AI to accurately target buttons, inputs, and links without complex CSS selectors.

Can Agent Browser handle dynamic content?

Absolutely. It uses a real headless browser engine that executes JavaScript, allowing it to interact with Single Page Applications (SPAs) and dynamic web content effectively.

Does this skill require external dependencies?

Yes, it requires the agent-browser CLI to be installed globally via npm, along with the necessary browser binaries which can be installed using the 'agent-browser install' command.

Agent Browser

Name: Agent Browser
Author: linuxlewis

bylinuxlewis

•

ブラウザ自動化

Automates web browsing and interaction tasks using the agent-browser CLI to navigate websites, scrape content, and capture visual snapshots.

Agent Browser provides Claude with a powerful interface to interact with the live web using a headless browser. It allows the agent to navigate URLs, interact with elements via an accessibility-first reference system, and manage persistent sessions for complex, multi-step workflows. This skill is essential for tasks requiring real-time web data, automated form filling, visual verification, or scraping content from dynamic JavaScript-heavy websites.

主な機能

01Persistent session management to maintain login states and cookies

02Accessibility-driven element interaction using unique @ref handles

03Comprehensive capture options including full-page screenshots and PDFs

04Direct CLI integration for fast, headless web automation

05Interactive accessibility tree snapshots for precise element targeting

061 GitHub stars

ユースケース

01End-to-end UI testing and verification of web applications

02Automated web research and data collection from live websites

03Automating repetitive web-based workflows like form submissions or data entry

主な機能

01Persistent session management to maintain login states and cookies

02Accessibility-driven element interaction using unique @ref handles

03Comprehensive capture options including full-page screenshots and PDFs

04Direct CLI integration for fast, headless web automation

05Interactive accessibility tree snapshots for precise element targeting

061 GitHub stars

ユースケース

01End-to-end UI testing and verification of web applications

02Automated web research and data collection from live websites

03Automating repetitive web-based workflows like form submissions or data entry