agent-browser

Drive a headless browser via inference.sh for web navigation and element interaction.

Updated Apr 25, 2026
One-click install
npx skills add https://github.com/hendripermana/permoney --skill agent-browser-hendripermana
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: agent-browser
Source: https://github.com/hendripermana/permoney/tree/main/.agents/skills/agent-browser
Command: npx skills add https://github.com/hendripermana/permoney --skill agent-browser-hendripermana

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

Browser automation for AI agents to browse pages, interact with elements using compact @e refs, take screenshots, and record video.

Core Features & Use Cases

  • Open pages, manage sessions, and interact with elements via the @e refs system.
  • Take screenshots and record video for debugging and documentation.
  • Execute custom JavaScript and perform common tasks like form filling, clicking, and drag-and-drop.
  • Use for research, testing, and automated data extraction across dynamic web apps.

Quick Start

Open https://example.com and return the current URL, title, and a compact element snapshot using compact @eRefs.

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 browsing tasks for AI agents?

Automate web browsing for AI agents by driving a headless browser via Playwright under inference.sh. This allows agents to perform web navigation, element interaction, and data capture through defined open, snapshot, interact, screenshot, execute, and close workflows.

Can I use Playwright to extract data from dynamic web apps with AI?

Yes, you can extract data from dynamic web apps by using Playwright under inference.sh. The Skill executes custom JavaScript and captures compact element snapshots, enabling reliable automated data extraction across interactive pages.

What is the @e refs system for browser automation?

The @e refs system provides compact element references for browser automation, allowing AI agents to reliably identify and interact with web page elements. It supports tasks like form filling, clicking, and drag-and-drop during automated sessions.

How do I take screenshots and record video during web scraping for debugging?

Take screenshots and record video during web scraping by using the screenshot and optional video capture features. These functions document the automated browser sessions for debugging and verification of dynamic web app interactions.

Does browser automation for AI agents require a specific environment setup?

Browser automation requires a root SKILL.md file with YAML frontmatter specifying name and description, and optionally includes scripts, references, and assets directories to extend functionality for deterministic tasks driven by Playwright.

What are the limitations of using Playwright under inference.sh for web navigation?

Playwright under inference.sh is designed for deterministic web navigation and element interaction tasks. Limitations arise when handling highly non-deterministic, unpredictable web behaviors that fall outside standard open, snapshot, interact, and execute workflows.