windows-computer-use-skill

<div align="center"> <h1>Windows Computer-Use Skill</h1> <p><strong>A top-level Windows skill with a bundled standalone runtime and MCP server.</strong></p> <p> <a href="https://github.com/wimi321/windows-computer-use-skill">GitHub</a> · <a href="./README.zh-CN.md">简体中文</a> · <a href="./README.ja.md">日本語</a> </p> </div>

This repository is:

• a top-level skill
• a standalone Windows desktop-control runtime
• a computer-use MCP server for agent ecosystems

It is packaged skill-first, not Claude-first, so the same runtime can be adapted for multiple agent products.

The requirement is stricter than "wrap an existing install":

• no dependency on a local Claude app
• no private .node binaries
• no extracted hidden assets
• install the skill, build the server, and use it

This project follows that rule on Windows.

• top-level Windows computer-use skill
• standalone MCP server for screenshots, mouse, keyboard, app launch, window/display mapping, and clipboard
• public dependency chain only: Node.js + Python + pyautogui + mss + Pillow + psutil + pywin32
• first-run runtime bootstrap: the server creates its own virtualenv and installs dependencies automatically
• bundled skill install that copies the full project into ~/.codex/skills/computer-use-windows/project
• extracted TypeScript tool layer wired to a Windows-native Python backend

Implemented in this repository:

• Windows Python helper and runtime bootstrap
• display enumeration and screenshot pipeline
• mouse, keyboard, drag, scroll, and clipboard primitives
• frontmost app, app-under-point, running app, installed app, and window-display lookup paths
• Windows-first skill packaging and bundled project payload
• TypeScript build passing

Still recommended before production use:

• validate on a real Windows host
• exercise UAC, elev