skills.rest
adamfeldman
adamfeldmanCOMMUNITY
Explore all skills by this author →
Website𝕏@adamfeldmanGitHub

agent-browser

Automate web browser interactions via CLI using accessibility snapshots and ref-based locators.

Updated Aug 23, 2026
Search Tags:#AI agent#browser automation#CLI#web scraping#form filling#headless browser
One-click install
npx skills add https://github.com/adamfeldman/compound-workflows --skill agent-browser-adamfeldman
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: agent-browser
Source: https://github.com/adamfeldman/compound-workflows/tree/main/plugins/compound-workflows/skills/agent-browser
Command: npx skills add https://github.com/adamfeldman/compound-workflows --skill agent-browser-adamfeldman

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This Skill automates web browser interactions for AI agents, enabling them to navigate websites, fill forms, scrape data, and take screenshots programmatically.

Core Features & Use Cases

  • Headless Browser Automation: Control a browser instance via CLI commands.
  • AI-Friendly Interaction: Uses accessibility snapshots and ref-based locators (@e1, @e2) for reliable element targeting.
  • Use Case: An AI agent needs to sign up for a new service. It uses agent-browser to open the signup page, snapshot the form elements, fill in the required fields using refs, and submit the form.

Quick Start

Use the agent-browser skill to open the URL https://example.com and then take an interactive snapshot.

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 interactions for an AI agent?▼

You can automate web browser interactions for an AI agent using a CLI tool that handles navigation, form filling, data scraping, and screenshotting. It employs accessibility snapshots and ref-based element selection for reliable targeting.

How does browser automation locate elements using accessibility snapshots?▼

Browser automation locates elements by taking interactive accessibility snapshots of the web page, then using semantic ref-based locators like @e1 or @e2 to reliably target and interact with specific elements.

Can I run parallel browser sessions for web scraping tasks?▼

Yes, you can run parallel browser sessions for web scraping and other automation tasks. This allows an AI agent to handle multiple web pages or browser interactions concurrently.

What is the best way to fill out web forms programmatically using a headless browser?▼

The best way to fill out web forms programmatically is using a CLI-driven headless browser. It snapshots the form elements via accessibility data and targets fields using ref-based locators for accurate data entry and submission.

Do I need any external dependencies to use this browser automation CLI?▼

No, you do not need to install any external dependencies to use this browser automation CLI. The Skill operates independently without requiring additional packages or libraries.

Are there limitations to using ref-based locators for web automation?▼

Ref-based locators rely on accessibility snapshots, meaning web automation may fail if a website's elements lack proper accessibility attributes or if the page structure changes dynamically between the snapshot and the interaction.

Related Skills

Explore More Software Engineering Skills→
AnEntrypointAnEntrypoint
20

agent-browser

Automate browser tasks via command-line interface for web navigation and data extraction.

Official
Advanced
paul-hosking-cdw-compaul-hosking-cdw-com

identify-page-structure

Identify section boundaries and content sequences from scraped HTML, screenshots, and metadata.

Community
Intermediate
hamzaPixlhamzaPixl
2

agent-browser

Automates browser UI verification and interaction via agent-browser CLI for AI agent workflows.

Community
Intermediate
harshaldhadukharshaldhaduk
1

browser

Automate web browser tasks using accessibility tree snapshots with element references.

Community
Intermediate