gsd-browser

Automate Chrome/Chromium browser navigation and interaction via CDP.

38|6|Updated May 29, 2026
One-click install
npx skills add https://github.com/open-gsd/gsd-browser --skill gsd-browser-open-gsd
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: gsd-browser
Source: https://github.com/open-gsd/gsd-browser/tree/main
Command: npx skills add https://github.com/open-gsd/gsd-browser --skill gsd-browser-open-gsd

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires chromiumoxide, rustenium, tokio, serde, clap, chromiumoxide, rustenium, futures, nix, which, thiserror, tracing, toml, tracing-subscriber, image, aes-gcm, argon2, rand, reqwest, and includes scripts (resource) and references (resource) and assets (resource) components.

What problem does it solve?

gsd-browser solves the problem of manual browser automation, allowing AI agents, developers, and CI pipelines to interact with web pages programmatically without the need for a full browser test framework.

Core Features & Use Cases

  • Full Chrome/Chromium Control: Navigate, interact, and control Chrome/Chromium with a persistent background daemon.
  • Advanced Commands: Over 90+ top-level commands for navigation, interaction, authenticated live viewing, annotations, recording bundles, snapshots, assertions, structured extraction, network control, visual diffing, tracing, and stateful auth flows.
  • MCP Server: Provides a first-class, powerful browser automation platform for AI agents with over 50+ tools and resources.
  • Use Case: Use gsd-browser to automatically navigate to a website, fill out a form, capture a screenshot, and export the data as a structured JSON file.

Quick Start

Run the following command to start the MCP server and point your MCP client at it:

gsd-browser mcp

Frequently Asked Questions about gsd-browser

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

FAQPage Schema
How do I automate browser control for an AI agent using Chrome?

Browser control for AI agents is automated using Chrome/Chromium via CDP. The Skill provides an MCP server with 50+ tools and resources, enabling agents to navigate, interact, and extract web data programmatically.

Can I use this for automated web testing in a CI/CD pipeline?

Yes, automated web testing in CI/CD pipelines is a core use case. It offers 90+ commands for navigation, assertions, visual diffing, and network control, allowing pipelines to validate web applications without a full test framework.

What is the best way to extract structured data from web pages programmatically?

Structured data extraction from web pages is handled through advanced commands that capture snapshots and export data as structured JSON. This allows developers and AI agents to process extracted information directly.

Does browser automation via CDP support authenticated sessions and stateful auth flows?

Browser automation via CDP supports stateful auth flows and authenticated live viewing. This enables persistent background daemon sessions to maintain login states across multiple navigation and interaction commands.

How do I start the MCP server for AI-driven browser automation?

To start the MCP server for AI-driven browser automation, run the command `gsd-browser mcp`. Point your MCP client at this server to access over 50 tools for web interaction and data extraction.

What are the limitations of using a background daemon for browser automation?

Using a background daemon for browser automation requires specific dependencies like chromiumoxide and tokio. While it provides deterministic execution, users must manage the daemon process and ensure Chrome/Chromium is properly configured for CDP communication.