cmux-browser

Automate cmux browser surfaces to open sites, interact, wait, and extract data.

25.8k|2.2k|Updated Jan 28, 2026
One-click install
npx skills add https://github.com/manaflow-ai/cmux --skill cmux-browser-manaflow-ai
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: cmux-browser
Source: https://github.com/manaflow-ai/cmux/tree/main/skills/cmux-browser
Command: npx skills add https://github.com/manaflow-ai/cmux --skill cmux-browser-manaflow-ai

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

Automates end-user browser surfaces in cmux to open sites, interact with pages, wait for state changes, and extract data from cmux browser surfaces.

Core Features & Use Cases

  • Open or target a browser surface to begin automation.
  • Verify navigation with get url before waiting or snapshotting, then snapshot to refresh element refs.
  • Act with refs (click, fill, type, select, press) and wait for state changes to complete flows.
  • Re-snapshot after DOM/navigation changes to keep refs fresh, and use surface targeting and templates for common patterns.
  • Deep-dive references and templates support advanced scenarios across multiple surfaces.

Quick Start

Open a browser surface, snapshot interactive refs, then perform actions using the exposed cmux browser commands.

Frequently Asked Questions about cmux-browser

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

FAQPage Schema
How do I automate browser interactions and extract data from web pages?

Browser automation executes commands to open sites, interact with elements using refs, wait for state changes, and extract data. You use commands like open, snapshot, fill, and click to execute multi-surface sessions.

What is a browser snapshot and when do I need to refresh element refs?

A browser snapshot captures current interactive element references for targeting actions. You must re-snapshot after any DOM or navigation changes to keep refs fresh and ensure subsequent commands target the correct elements.

How do I handle multi-surface sessions and login flows in web automation?

Multi-surface session automation uses per-surface state management to handle login flows, form submissions, and navigation. You target specific browser surfaces individually and apply templates for common patterns.

Can I verify page navigation before performing actions on web elements?

Yes, you verify navigation by checking the URL before waiting or snapshotting. This ensures the page loaded correctly, allowing you to safely snapshot interactive refs and perform click, fill, select, and press actions.

What are the limitations of using snapshots for browser state management?

Snapshot-based state management requires re-snapshotting after any DOM or navigation changes. If elements shift dynamically without a new snapshot, refs become stale and subsequent clicks or fills may fail to target correctly.