agent-browser

Automate browser navigation, element interaction, and data extraction via CLI.

Updated Jan 16, 2026
One-click install
npx skills add https://github.com/8b-is/8b-is-mp --skill agent-browser-8b-is
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: agent-browser
Source: https://github.com/8b-is/8b-is-mp/tree/main/plugins/compound-engineering/skills/agent-browser
Command: npx skills add https://github.com/8b-is/8b-is-mp --skill agent-browser-8b-is

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

AI agents often struggle with performing repeatable web interactions such as navigation, form filling, and data extraction. This Skill provides a CLI-based solution using Vercel's agent-browser to automate these tasks reliably.

Core Features & Use Cases

  • CLI-based browser automation with ref-based element selectors (@e1, @e2)
  • Navigate, snapshot, and interact with web pages
  • Take screenshots and extract page data for downstream tasks
  • Use Case: Automate a login flow: open login page, snapshot to identify inputs, fill credentials, click Sign in, capture a confirmation screenshot.

Quick Start

Use agent-browser to open a site, snapshot to identify elements, interact with refs, and capture a screenshot.

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 tasks and data collection using a CLI browser?

You can automate web tasks using a CLI browser by navigating pages, taking snapshots to identify elements, interacting via ref-based selectors, and extracting data or screenshots. This approach handles single or multi-page web flows reliably without manual intervention.

What are ref-based selectors and how do they work for web automation?

Ref-based selectors are element identifiers like @e1 or @e2 generated by snapshotting a page. Instead of relying on fragile CSS selectors, you use these refs to click, fill, or extract data from specific page elements during browser automation.

Do I need to install any dependencies before using agent-browser for web automation?

Yes, you need the agent-browser CLI installed. Run npm install -g agent-browser followed by agent-browser install to set up the required browser environment before automating navigation, snapshots, and interactions.

Can I automate a login flow and capture a screenshot using CLI commands?

Yes, you can automate a full login flow by opening the login page, snapshotting to find input refs, filling credentials, clicking sign in, and executing a screenshot command to capture the confirmation state.

What CLI commands are available for browser automation and element interaction?

The available CLI commands include open for navigation, snapshot for capturing page state, click and fill for element interaction, get for data retrieval, and screenshot for capturing visual output across web flows.

Why use CLI browser automation instead of traditional CSS selector-based scripts?

CLI browser automation with ref-based selectors avoids brittle CSS selectors that break on layout changes. By snapshotting pages and interacting with generated refs, your automation scripts remain stable across single or multi-page web flows.