browse-site skill for AI agents using agent-browser CLI

SKILL.md

name	description
browse-site	Browser automation using agent-browser CLI. Use when users ask to navigate websites, fill forms, take screenshots, extract web data, test web apps, or automate browser workflows. Trigger phrases include "go to [url]", "click on", "fill out the form", "take a screenshot", "scrape", "automate", "test the website", "log into", "browse to", "open website", "visit", or any browser interaction request.

Browse Site Skill

Browse and interact with websites using the agent-browser CLI.

Installation

If agent-browser is not installed:

npm install -g agent-browser
agent-browser install  # Download Chromium

Quick Start

agent-browser open https://example.com
agent-browser snapshot -i              # Get interactive elements with refs
agent-browser click @e2                # Click element by ref
agent-browser screenshot tmp/page.png  # Capture screenshot

Workflow

1. Navigate: agent-browser open <url>
2. Snapshot: agent-browser snapshot -i --json returns accessibility tree with refs
3. Interact: Use refs to click, fill, type (e.g., agent-browser click @e2)
4. Re-snapshot: Get new snapshot after page changes
5. Repeat: Continue until task complete

Always use refs from the most recent snapshot. Refs are deterministic and point to exact elements.

Common Commands

Action	Command
Navigate	agent-browser open <url>
Get elements	agent-browser snapshot -i
Click	agent-browser click @e<n>
Type text	agent-browser type @e<n> "text"
Fill field	agent-browser fill @e<n> "text"
Screenshot	agent-browser screenshot [path]
Get text	agent-browser get text @e<n>
Get page title	agent-browser get title
Get current URL	agent-browser get url
Wait	agent-browser wait <selector or ms>

Sessions

Use --session <name> to maintain state across commands:

agent-browser open https://example.com --session mysite
agent-browser click @e3 --session mysite

Snapshot Options

Option	Description
-i, --interactive	Only interactive elements (buttons, links, inputs)
-c, --compact	Remove empty structural elements
-d, --depth <n>	Limit tree depth
-s, --selector <sel>	Scope to CSS selector
--json	Machine-readable output

Combine options: agent-browser snapshot -i -c -d 5 --json

Tips

• Always use --json for parsing snapshot output programmatically
• Use -i (interactive only) to reduce output size
• Use --headed to watch the browser visually
• Refs like @e2 come from the snapshot output and are deterministic
• Close the browser when done: agent-browser close