agent-browser

Automate browser navigation, form filling, clicking, and data capture via CDP.

Updated May 12, 2026

Search Tags:#automation #cdp #web-testing #multi-session #browser #snapshot #automation-scripts

Ready to run

npx skills add https://github.com/YeomanYe/skills --skill agent-browser-yeomanye

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: agent-browser
Source: https://github.com/YeomanYe/skills/tree/main/agent-browser
Command: npx skills add https://github.com/YeomanYe/skills --skill agent-browser-yeomanye

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Browser automation is time-consuming and error-prone when agents must interact with websites; this Skill centralizes and abstracts those tasks, enabling consistent, checkable automation across pages and apps.

Core Features & Use Cases

Snapshot-based refs for deterministic interactions (click, fill, etc.)
Multi-session isolation to run concurrent automations safely
Batch command execution for efficient workflows
Support for navigation, form handling, data capture, and visual validation across sites
Use Case: automate login, data extraction, testing, and end-to-end tasks in AI agent workflows.

Quick Start

Open a website, run a page snapshot to generate refs, then use those refs to perform automated interactions.

Frequently Asked Questions about agent-browser

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I automate web browser tasks for AI agents?▼

You can automate web browser tasks for AI agents by using a Skill that centralizes website interactions like navigation, form filling, clicking, and data capture through anchored element references and batch command execution.

How does snapshot-based browser automation work for element interactions?▼

Snapshot-based browser automation works by generating a page snapshot to create deterministic element references. You then use these anchored refs to perform reliable interactions like clicking or filling forms across various sites.

Can I run concurrent browser automation workflows safely?▼

Yes, you can run concurrent browser automation workflows safely using multi-session isolation. This feature allows you to manage multiple concurrent automations independently without session conflicts.

Do I need a specific browser to run CDP automation scripts?▼

Yes, you need a Chromium-based browser with CDP access and the agent-browser CLI installed to run CDP automation scripts and execute resilient automation patterns across target sites.

What is the best way to handle login and data extraction in automated testing?▼

The best way to handle login and data extraction in automated testing is to use batch command execution with robust wait and authentication patterns, ensuring resilient automation across many sites.

Why does my web testing automation fail on dynamic page elements?▼

Web testing automation can fail on dynamic elements if not using snapshot-based refs for deterministic interactions. Generating page snapshots before interaction ensures anchored references for resilient automation.