How does Agent Browser identify elements for interaction?

The skill uses a 'snapshot' command that analyzes the DOM and assigns unique reference tags (like @e1 or @e2) to interactive elements, allowing Claude to target buttons and inputs reliably.

Can I use this skill to automate sites that require a login?

Yes, you can save and load browser states (including cookies and local storage) using the 'state save' and 'state load' commands to maintain authenticated sessions.

Is it possible to mock API responses during browser sessions?

Yes, the 'network route' command allows you to intercept requests and provide mock JSON bodies or status codes to simulate different backend scenarios.

Does Agent Browser support recording interactions?

Absolutely. It includes commands to start and stop video recordings, which is ideal for creating demos or debugging failed automation sequences.

Can it handle dynamic content and network delays?

Yes, it features advanced 'wait' commands that can trigger based on specific text, URL patterns, JS conditions, or network idle states.

Agent Browser

Name: Agent Browser
Author: horuz-ai

byhoruz-ai

•

Automatización de Navegadores

Automates full-featured web browser interactions, testing, and data extraction through Claude Code.

Acerca de

The Agent Browser skill empowers Claude Code to operate a web browser with high precision, enabling it to perform tasks like automated UI testing, form filling, and web scraping. It utilizes a sophisticated snapshot system that identifies interactive elements via unique references, allowing for reliable navigation, session management, and visual debugging through screenshots and video recordings. This skill is essential for developers who need to interact with web applications, handle complex authentication flows, or extract structured data from modern, JavaScript-heavy websites directly from their terminal.

Características Principales

Interactive element snapshots with unique @refs for precise targeting
1 GitHub stars
Network interception for mocking API responses and blocking requests
Advanced session persistence with browser state saving and loading
Full-featured interaction suite including click, fill, scroll, and drag-and-drop
Visual debugging tools for screenshots, PDF generation, and video recording

Casos de Uso

Automating repetitive web-based workflows like form submissions and data entry
Automating end-to-end (E2E) testing and visual regression checks
Extracting structured data from complex or authenticated web pages

Acerca de

Características Principales

Interactive element snapshots with unique @refs for precise targeting
1 GitHub stars
Network interception for mocking API responses and blocking requests
Advanced session persistence with browser state saving and loading
Full-featured interaction suite including click, fill, scroll, and drag-and-drop
Visual debugging tools for screenshots, PDF generation, and video recording

Casos de Uso

Automating repetitive web-based workflows like form submissions and data entry
Automating end-to-end (E2E) testing and visual regression checks
Extracting structured data from complex or authenticated web pages