What problem does it solve? Taking screenshots from an AI agent or terminal requires different OS-specific commands, permission handling, and output conventions on every platform, which leads to repeated trial and error. ## Core Features & Use Cases - Cross-platform capture: One Python helper covers macOS and Linux, with a PowerShell helper for Windows, supporting full screen, pixel regions, active windows, and specific window IDs. - macOS app and window targeting: Capture all windows of a named app, filter by window title, list window IDs, and handle multi-display output with per-display files. - Permission preflight: A bundled script checks and requests macOS Screen Recording permission once, reducing repeated system prompts in sandboxed runs. - Use Case: Ask the agent to look at a running desktop app and compare it against a Figma design; it captures the app windows to the temp directory and inspects each image. ## Quick Start Take a screenshot of the currently active window and save it to the temp folder so you can look at it.