Can I use this for websites requiring login?

Yes, the skill supports saving and loading authentication states (cookies and localStorage) to maintain persistent sessions across different automated runs.

Can I record sessions for debugging?

Yes, Agent Browser supports taking full-page screenshots, saving pages as PDFs, and even recording video sessions in .webm format for visual debugging.

How do I handle complex JavaScript execution?

You can use the eval command with the --stdin flag or base64 encoding to execute complex scripts while bypassing shell interpretation and escaping issues.

How does Agent Browser handle dynamic content?

It uses a snapshot-ref system that assigns unique identifiers like @e1 to interactive elements. Users are encouraged to re-snapshot after any DOM changes to ensure refs remain valid.

Does it support mobile browser automation?

Yes, it includes specialized support for the iOS Simulator, allowing you to automate Safari on mobile devices using the same core workflow as desktop.

Agent Browser Automation

Name: Agent Browser Automation
Author: JantonioFC

byJantonioFC

•

浏览器自动化

Automates web browser interactions through a specialized CLI for AI agents to navigate, scrape, and test websites.

Agent Browser is a robust CLI tool designed for AI agents to perform complex browser tasks with high precision. It features a unique element referencing system that allows agents to interact with buttons, forms, and dynamic content without needing to manage complex CSS selectors manually. Supporting session persistence, authentication state management, and even iOS mobile simulation, it is an essential skill for web scraping, automated testing, and multi-step web workflows that require real browser rendering.

主要功能

01Mobile Safari automation via iOS Simulator integration

02Snapshot-based element referencing (@refs) for reliable AI interaction

03Comprehensive interaction suite including click, fill, select, and scroll

04Session persistence and state management for authenticated workflows

052 GitHub stars

06Advanced data extraction with JSON output and screenshot capabilities

使用场景

01Web scraping and data extraction from dynamic, JavaScript-heavy sites

02Automated UI and cross-platform testing for web applications

03Automated web form submission and account registration workflows

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add jantoniofc/skillsbank agent-browser

For use in Claude.ai and ChatGPT

主要功能

01Mobile Safari automation via iOS Simulator integration

02Snapshot-based element referencing (@refs) for reliable AI interaction

03Comprehensive interaction suite including click, fill, select, and scroll

04Session persistence and state management for authenticated workflows

052 GitHub stars

06Advanced data extraction with JSON output and screenshot capabilities

使用场景

01Web scraping and data extraction from dynamic, JavaScript-heavy sites

02Automated UI and cross-platform testing for web applications

03Automated web form submission and account registration workflows

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add jantoniofc/skillsbank agent-browser

For use in Claude.ai and ChatGPT