Can I use this for authenticated sessions?

Yes, it supports saving and loading browser states, including cookies and local storage, to maintain authentication across different sessions.

What is the agent-browser skill?

It is a specialized capability for Claude Code that allows the AI to control a web browser, perform actions like clicking and typing, and extract data directly from the web.

Can it record videos or take screenshots?

Absolutely, it includes built-in commands for taking full-page screenshots, generating PDFs, and recording high-quality videos of web interactions.

How does it identify elements on a page?

It uses a snapshot command that generates unique references (like @e1) for interactive elements based on the accessibility tree, ensuring precise and reliable interaction.

Does it support headless and headed modes?

Yes, you can run it in the default headless mode for efficiency or use the --headed flag to visually see the browser interactions during debugging.

Agent Browser Automation

Name: Agent Browser Automation
Author: Jelidia

byJelidia

0•

ブラウザ自動化

Automates web browser interactions for testing, data extraction, and workflow automation.

Agent Browser Automation is a powerful skill for Claude Code that enables sophisticated interaction with web pages through a CLI-driven browser engine. It provides a unique 'snapshot and ref' system that allows Claude to accurately identify and interact with elements using accessibility-tree-based references. Beyond simple navigation, it supports complex workflows including form filling, multi-session management, network request interception, and media capture via screenshots, PDFs, and video recording. It is an essential tool for developers building automated tests, web scrapers, or AI-driven web agents.

主な機能

01Device emulation for testing mobile viewports, geolocation, and media queries

02Interactive snapshot system using element references (@refs) for precise control

03Advanced media capture for screenshots, PDFs, and full-session video recording

040 GitHub stars

05Network-level control with proxy support, request routing, and cookie management

06Complete browser lifecycle management including navigation, tabs, and frames

ユースケース

01Programmatic web scraping and structured data extraction from complex SPAs

02Automating repetitive web-based administrative tasks and form submissions

03Automated end-to-end (E2E) web application testing and bug reproduction

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add jelidia/entretien-prestige agent-browser

For use in Claude.ai and ChatGPT

Download Skill