Core

windows_execute, windows_system_info, windows_terminal_*.

Files

windows_file_read, windows_file_write, windows_file_edit, windows_search, windows_drives.

Processes & services

windows_process_list, windows_process_kill, windows_service_list, windows_service_control.

Audio & display

windows_audio_*, windows_display_*, windows_theme_*.

Power & network

windows_power_*, windows_network_*, windows_wifi_*, windows_firewall_*.

System administration

windows_registry_*, windows_env_*, windows_task_scheduler_*, windows_startup_*.

Packages & WSL

windows_winget_*, windows_wsl_*.

Capture & browser

windows_window_capture (single-window, occlusion-tolerant) and windows_chrome_control, a full Chrome automation tool over the DevTools Protocol. It can navigate, click/type/hover/drag, choose <select> options and upload files, screenshot with responsive and mobile viewport emulation, run JavaScript, read the console log and network requests, auto-handle JavaScript dialogs (alert/confirm/prompt), and take an accessibility snapshot that assigns [ref] ids so clicks and typing target elements reliably instead of brittle CSS selectors. It can also export the page as a PDF, screenshot a single element (by ref or selector), and record a short screencast saved as an animated PNG that plays in any modern browser. For testing and debugging it adds emulation (geolocation, timezone, locale, dark mode, network throttling, CPU slowdown), request mocking/blocking, performance metrics and unused CSS/JS coverage, live CSS editing that survives page reloads, computed-style inspection, and a DevTools-style element highlight overlay. Because it lives inside a Windows system MCP it also handles what pure browser tools cannot: real downloads (point Chrome at a known folder and get the finished file path on disk) and driving the native Windows Open/Save dialog. Cookie/storage writes and request interception are opt-in via the chromeRiskyFeatures config key (see SECURITY.md).