agent-browser

Automate browser interactions using Vercel's agent-browser CLI and ref-based selectors.

23|2|Updated Jan 24, 2026
One-click install
npx skills add https://github.com/marcusrbrown/systematic --skill agent-browser-marcusrbrown
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: agent-browser
Source: https://github.com/marcusrbrown/systematic/tree/main/skills/agent-browser
Command: npx skills add https://github.com/marcusrbrown/systematic --skill agent-browser-marcusrbrown

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill provides a lightweight, CLI-based solution to automate web interactions for AI agents, allowing you to navigate pages, fill forms, capture screenshots, and scrape data without manual browser operations.

Core Features & Use Cases

  • Command-line browser automation: Drive web tasks from chat or automation scripts using Vercel's agent-browser.
  • Ref-based element interactions: Interact with page elements using stable refs (e.g., @e1) derived from accessibility snapshots.
  • Screenshots, scraping, and form handling: Take screenshots for verification, extract data, and fill forms programmatically.
  • Use Case: Example: log into a site, navigate to a report, capture a screenshot, and extract key metrics for daily dashboards.

Quick Start

Use the agent-browser to:

  • Open a URL with agent-browser open https://example.com
  • Snapshot interactive elements with agent-browser snapshot -i
  • Interact with elements using refs like agent-browser click @e1
  • Repeat as needed to automate tasks

Frequently Asked Questions about agent-browser

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

FAQPage Schema
How do I automate browser interactions for AI agents without manual operations?

This CLI-driven web automation tool allows AI agents to navigate pages, fill forms, take screenshots, and scrape data without manual browser operations. It uses Vercel's agent-browser CLI and Bash commands to execute deterministic actions and produce structured outputs.

How do I interact with web page elements programmatically for data scraping?

You interact with web page elements using ref-based selectors derived from accessibility snapshots. By snapshotting interactive elements, you reference them with stable refs like @e1 to perform clicks, form fills, and data extraction deterministically.

Can I use command-line browser automation to log into websites and capture screenshots?

Yes, command-line browser automation supports logging into websites and capturing screenshots. You can open a URL, snapshot interactive elements, input credentials, and capture visual verification screenshots entirely through CLI commands.

Do I need any external dependencies to run CLI-based web automation tasks?

No external dependencies are required beyond the environment itself. The Skill relies on Vercel's agent-browser CLI, Bash commands, and ref-based element selectors to perform web automation tasks without additional package installations.

What's the best way to extract key metrics from a website report for daily dashboards?

The best way to extract metrics is using CLI-driven web automation to log into the site, navigate to the report, and scrape the data. You can capture a screenshot for visual verification and extract key metrics programmatically for your dashboards.

Why use ref-based element selectors instead of CSS selectors for web automation?

Ref-based element selectors provide stable references like @e1 derived from accessibility snapshots, ensuring deterministic actions. Unlike CSS selectors that may break with layout changes, refs offer reliable element interaction for AI agents navigating dynamic web pages.