What problem does it solve? Many tasks involve native app windows, OS dialogs, or browser chrome that expose no API and no DOM, leaving automation tools blind. This Skill drives the real desktop directly with mouse, keyboard, and screenshots so those interfaces become operable. ## Core Features & Use Cases - Desktop control via one MCP tool: Perform screenshots, cursor moves, clicks, drags, scrolling, key presses, and typing through a single deferred tool. - OS and browser chrome automation: Handle file pickers, permission prompts, System Settings, the menu bar, address bar, extension popups, and autofill dropdowns. - Visual verification: Capture real screenshots to confirm what rendered, what has focus, or which window is in front. - Use Case: After a Chrome restart, chrome-devtools calls hang behind the native "Allow remote debugging?" dialog; take a screenshot, click Allow at its coordinates, confirm it is gone, and re-run the blocked call. ## Quick Start Ask the assistant to take a screenshot of the screen and click the Allow button on the remote debugging dialog blocking the browser session.
