App Screenshot & UI Control Skill
Capture live, pixel-exact desktop screenshots of the running LibrAgent application window across Windows and Linux, with optional mouse click and UI interaction automation.
Features
- Window Focus & Screenshot: Finds and focuses LibrAgent app window, then captures exact window geometry.
- Relative Coordinate Mouse Click: Automatically clicks buttons, menus, or cards relative to top-left of the app window before capturing (
--click X Y). - Cross-Platform: Zero extra native dependencies (uses Python
ctypes.windll.user32on Windows,Xlib/xwininfoon Linux).
Usage Examples
1. Capture App Screenshot
python src-tauri/bundled_skills/app-screenshot/scripts/capture_app.py docs/user/assets/screenshots/getting-started/new-session.png
2. Click Relative Coordinate & Capture
Click inside the app window at relative (X, Y) (e.g. X=710, Y=410 for Assistant card) and capture result:
python src-tauri/bundled_skills/app-screenshot/scripts/capture_app.py docs/user/assets/screenshots/getting-started/new-session.png --click 710 410