← Back to Skills Marketplace
135
Downloads
0
Stars
0
Active Installs
1
Versions
Install in OpenClaw
/install computer-use-windows
Description
Top-level Windows computer-use skill with a bundled standalone runtime that bootstraps itself without any local Claude installation, private native modules,...
README (SKILL.md)
Windows Computer-Use Skill
Use this skill when the task needs a portable Windows computer-use skill bundled with its own standalone runtime and MCP server.
What this skill does
- uses the bundled
windows-computer-use-skillproject under the installed skill directory - builds the standalone MCP server
- lets the server auto-bootstrap its Python runtime on first launch
- avoids any dependency on local Claude binaries,
.nodemodules, or extracted app assets - stays explicitly Windows-only because the underlying desktop-control backend is Windows-specific
Default bundled project path
After installation, assume the standalone project lives at:
~/.codex/skills/computer-use-windows/project
If the user installed the skill under a custom CODEX_HOME, use that equivalent path instead.
Build
Always build from the bundled project:
cd ~/.codex/skills/computer-use-windows/project
npm install
npm run build
Run
cd ~/.codex/skills/computer-use-windows/project
node dist/cli.js
The first real run will automatically create .runtime/venv and install the public Python dependencies.
Validation notes
- Version
0.1.1fixes the shared system-key blocklist logic so Windows runtime builds use the correct Windows shortcut denylist instead of a broken cross-platform branch. - This project has been statically validated from macOS for TypeScript build health, bundled project integrity, and Python helper syntax, but still needs end-to-end runtime validation on a real Windows machine.
Guardrails
- Treat this host as trusted-local only.
- Do not tell the user to search their local Claude install for binaries or hidden assets.
- Be explicit that this runtime is standalone and uses public dependencies only.
- Mention that the current runtime reports
screenshotFiltering: none, so action gating is handled at the MCP layer.
Usage Guidance
This package is internally consistent with its stated purpose, but it grants powerful local privileges (screenshots, clipboard access, mouse/keyboard, registry and process enumeration). Before installing or running it: 1) only run on a Windows machine you control or in an isolated VM; 2) review runtime/windows_helper.py and the TypeScript server code yourself (they perform registry reads, base64-encode screenshots, and expose JSON over stdio/MCP); 3) be aware npm install and first-run pip installs will fetch public packages from npm/PyPI; 4) ensure your agent's MCP gating/policy prevents automatic exfiltration of screenshots/clipboard or requires explicit user confirmation for sensitive actions; 5) prefer running first in a sandboxed/test environment and verify the upstream source (GitHub/ClawHub link) and signatures if you must deploy to production.
Capability Analysis
Type: OpenClaw Skill
Name: computer-use-windows
Version: 0.1.1
This skill implements a full-featured remote desktop control suite for Windows, providing the AI agent with capabilities to take screenshots, read/write the clipboard, and simulate mouse/keyboard input via 'project/runtime/windows_helper.py'. It includes a self-bootstrapping mechanism in 'project/src/computer-use/pythonBridge.ts' that automatically creates a Python virtual environment and installs dependencies from the internet. While the project contains extensive safety logic and application denylists (e.g., in 'project/src/vendor/computer-use-mcp/deniedApps.ts') to prevent interaction with sensitive software like banking or crypto apps, the high-risk nature of providing an LLM with unrestricted desktop access and the automated execution of remote installers make it inherently risky.
Capability Assessment
Purpose & Capability
The name/description (Windows computer-use skill with bundled standalone runtime) matches the included files: a Node TypeScript tool layer and a Python Windows helper that performs screenshots, window enumeration, input control (pyautogui), registry reads, and app/process inspection. There are no declared env vars or unrelated credentials required.
Instruction Scope
SKILL.md instructs building the bundled project (npm install, npm run build) and running node dist/cli.js; the runtime auto-creates a Python virtualenv and installs requirements. The actions the code performs (screenshots, clipboard access, mouse/keyboard control, registry reads) are within the stated purpose. Note: these capabilities allow capture of screen/clipboard and remote agent-driven input — SKILL.md does not (and cannot) prevent an agent from capturing and transmitting sensitive local data once allowed.
Install Mechanism
No install spec is declared for the platform; SKILL.md expects you to run npm install and npm build and the runtime will pip-install standard public Python packages from PyPI (mss, pyautogui, Pillow, psutil, pywin32). All installs are from well-known package registries (npm/PyPI); there are no obscure download URLs or remote extract steps in the manifest. This is normal but means network fetches will occur at build/first-run.
Credentials
The skill declares no required env vars or credentials. Example MCP config uses non-sensitive flags (debug, coordinate mode). The bundle does create local config/lock files under the user's home directory and reads the Windows registry for app info — behavior expected for app/window discovery and coherent with the skill's purpose.
Persistence & Privilege
The skill is not 'always' enabled and does not request system-wide config changes. It runs as a user-level MCP server and writes a venv and a local lock file under the user's profile. However, because it can autonomously be invoked by an agent (default platform behavior) and it controls the desktop and can capture screenshots/clipboard, this is a high-privilege capability — run only where you trust the agent and policies that gate screenshot/exfiltration.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install computer-use-windows - After installation, invoke the skill by name or use
/computer-use-windows - Provide required inputs per the skill's parameter spec and get structured output
Version History
v0.1.1
Fix Windows system-key blocklist branching, sync bundled skill payload, and upgrade multilingual README plus validation notes.
Metadata
Frequently Asked Questions
What is Computer Use Windows?
Top-level Windows computer-use skill with a bundled standalone runtime that bootstraps itself without any local Claude installation, private native modules,... It is an AI Agent Skill for Claude Code / OpenClaw, with 135 downloads so far.
How do I install Computer Use Windows?
Run "/install computer-use-windows" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is Computer Use Windows free?
Yes, Computer Use Windows is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does Computer Use Windows support?
Computer Use Windows is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created Computer Use Windows?
It is built and maintained by wimi321 (@wimi321); the current version is v0.1.1.
More Skills