Sponsored by APIMart - AI API, 99.9% SLA. Your AI, Always On.Most models offer cost savings of 20–60%; nanobanana pro api costs only 0.03.

AI Ad Library

Openclaw Skills

Windows Control

Windows Control for Openclaw

A comprehensive Windows desktop automation skill for controlling mouse, keyboard, and UI elements programmatically.

spliff7777

v1.0.0

Feb 2, 2026

Install

Information

Skill.md

Install & Download

1. ClawHub CLI

The fastest way to install a skill directly from the registry.


                  npx clawhub@latest install windows-control

2. Manual Installation

Copy the skill folder to one of these locations

Global

~/.openclaw/skills/

Workspace

<project>/skills/

Priority: Workspace > Local > Bundled

3. Prompt Installation

Copy this prompt to OpenClaw to install it automatically.


                  Help me install windows-control using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).

Prefer to download?

Get the raw skill files in a ZIP archive.

What is Windows Control?

The Windows Control skill provides an exhaustive suite of tools for full desktop automation on Windows environments. By leveraging Openclaw Skills, developers can interact with any application just like a human user—controlling the mouse, typing text, managing windows, and extracting UI data without relying solely on slow OCR methods. It bridges the gap between AI agents and native Windows software, enabling complex cross-application workflows and robust UI testing.

This skill is designed to be both fast and accurate, utilizing Windows UI Automation to read internal element structures directly. This allows the agent to "see" buttons, links, and text fields as data rather than just pixels, significantly improving reliability over traditional coordinate-based automation.

Windows Control Use Cases

Automating repetitive data entry across legacy Windows applications.
Performing end-to-end UI testing for desktop software and web browsers.
Extracting structured data from non-API enabled Windows programs.
Managing system-level tasks like handling file dialogs, alerts, and window positioning.
Creating complex multi-step workflows involving screenshots and visual analysis.

How Windows Control Works

The skill identifies the target window or screen region using built-in discovery scripts like list_windows.py.
It extracts structural UI information (buttons, links, text) via Windows UI Automation for high precision using read_ui_elements.py.
If visual confirmation is needed, a base64 screenshot is captured for analysis using screenshot.py.
The agent executes precise mouse movements, clicks, or keystrokes to interact with specific elements based on coordinates or element names.
It handles popups and modal dialogs dynamically through the handle_dialog.py module.
The workflow concludes by verifying the updated state of the UI to ensure the action was successful.

Windows Control Setup

To begin using these Openclaw Skills, ensure your environment meets the Python 3.11+ requirement and has the necessary libraries installed:

# Install core dependencies
pip install pyautogui pillow

# All scripts are located in the skills directory
cd skills/windows-control/scripts/

# Verify installation by taking a screenshot
py screenshot.py > test.b64

Windows Control Data Schema & Taxonomy

The Windows Control skill organizes interaction data and screen captures through a series of specialized Python scripts. Below is the primary data output schema:

Script	Output Format	Description
screenshot.py	Base64 PNG	Captures the full screen for visual processing.
read_window.py	String/Plain Text	Extracts text content from a specific window handle using UI Automation.
read_ui_elements.py	JSON/List	Maps interactive elements like buttons and links with their coordinates.
handle_dialog.py	JSON/Status	Reports on and interacts with modal dialogs like 'Save As' or 'Alert'.
get_active_window.py	String	Returns the title of the currently focused window.

Windows Control Advanced Features

Click by Name: Interact with UI elements (Save, OK, Submit) without needing coordinate mapping through click_element.py.
Intelligent Dialog Handling: Automatically detect, read, and dismiss system alerts or save-as dialogs.
Browser Content Extraction: Deep inspection of web pages within Chrome or Edge via native UI automation for faster scraping.
Fail-safe Protection: Built-in safety move-to-corner mouse abort to prevent runaway automation scripts.
Structural UI Reading: High-speed text extraction that bypasses the need for resource-heavy OCR by reading the window's underlying accessibility tree.