Agent Browser

This is a browser automation tool that lets Claude interact with web pages through a ref-based system. You open a URL, take a snapshot to get interactive elements labeled as @e1, @e2, etc., then use those refs to click buttons, fill forms, and navigate around. It's built for the request-website command to submit indexing requests, but works for any form filling or page interaction task. The workflow is straightforward: snapshot after each navigation to get fresh refs, interact using those refs, then always close the browser when done. If you need Claude to fill out forms or click through multi-step web flows without writing Playwright scripts, this handles it through simple bash commands.

skills.sh installs: 136
GitHub stars: 1.6K
last updated: Aug 21, 2026

View on GitHub

Install

$ npx -y skills add actionbook/actionbook --skill agent-browser --agent claude-code

Installs into .claude/skills of the current project.

Files

Browser Automation with agent-browser

Quick start

agent-browser open <url>        # Navigate to page
agent-browser snapshot -i       # Get interactive elements with refs
agent-browser click @e1         # Click element by ref
agent-browser fill @e2 "text"   # Fill input by ref
agent-browser close             # Close browser

Core workflow

Navigate: agent-browser open <url>
Snapshot: agent-browser snapshot -i (returns elements with refs like @e1, @e2)
Interact using refs from the snapshot
Re-snapshot after navigation or significant DOM changes
Always close: agent-browser close

Commands

Navigation

agent-browser open <url>      # Navigate to URL
agent-browser close           # Close browser (ALWAYS do this)

Snapshot (page analysis)

agent-browser snapshot        # Full accessibility tree
agent-browser snapshot -i     # Interactive elements only (recommended)

SKILL.md on GitHub

Details

Repository: github.com/actionbook/actionbook
Categories: AI & Agent Building, Automation & Workflows, Design & UI/UX
First seen: Jun 3, 2026

More from actionbook/actionbook

All 7