computer-use - Control desktop apps without stealing focus
Drive desktop applications across macOS, Windows, and Linux by clicking, typing, scrolling, and dragging in the background.
Tags
Updated: 2026-10-02Capabilities
Typical Inputs
Typical Outputs
What this skill does
- Capture desktop state
- Click interface elements
- Type text
- Scroll content
- Drag interface elements
- Send keyboard shortcuts
- List applications
- Verify action effects
Inputs
- Target application
- Action parameters
- Text input
- Keyboard shortcuts
- Element indices
- Screen coordinates
Outputs
- Screenshots
- Accessibility trees
- Action verdicts
- Desktop state changes
Requirements
- Available computer_use tool
- macOS, Windows, or Linux
- Tool-capable model