windows-computer-use-mcp (sandraschi/windows-computer-use-mcp) is an MCP server listed on the M8ven Trust Index. It scores 53 out of 100, grade D. It declares 15 tools. No publisher has claimed this listing.

Caution

53/100

windows-computer-use-mcp

Item: windows-computer-use-mcp
Rating: 53
Author: M8ven

Windows Computer Use for AI Agents. Both a tool (22 MCP tools for click, type, screenshot, OCR, UI inspection) and an agent (autonomous mission engine, macro recorder, intent-based discovery, event watchers). Built with opencode (DeepSeek V4). Ships as MCP server, web UI, and Tauri desktop app.

→ sandraschi/windows-computer-use-mcp · listed on glama

Caution. Specific findings reduced this grade. They are listed on the page. Grades reflect the full trust pyramid: code, verification depth, and reputation. New projects cap at C until adoption is earned.

How we verified

⚡ Live Monitored: not connected

Verified is a snapshot. Live keeps it current, and builds your track record.

⚡ Connect GitHub → continuous verification on every push why connect →

Who stands behind it

sandraschi

Source: Glama · also listed on github_repo_search

Is this your MCP?

Claim it to get a verified publisher badge, a free copy of our full audit findings, and direct contact for any high-priority issues we find. Or connect your repo for our deepest verification, Live Monitored: read-only, revoke anytime. What we access →

⚡ Connect GitHub for Live Monitored

Install from

Glamanot scored here GitHub

The grade above is for the source repository. Registries can serve a different version, so we mark the ones we were not able to read.

// key findings

⚠️

Tool descriptions don’t match what handlers do

3 tools describe read intent but their handlers mutate — automation_mouse (line 243: subprocess.Popen(); automation_visual (line 181: with open(output_path, "wb") as f:); find_image (line 331: os.unlink(screenshot["file_path"]))

✅

No credential exfiltration, no sensitive file access, no obfuscation

Static analysis found nothing flowing your secrets to unexpected places.

✅

Open source with a license and README

Anyone can audit the code, the license is declared, and the publisher documents what it does.

// tools this server exposes97 tools · 83 behind config

These names and descriptions are the publisher's own, read from the source code. We print them as written. Our assessment is the findings above, not this list.

approve_automation

Approve UI automation actions for a specified duration to prevent repetitive prompts.

automation_safety

Inspect or reset server-side safety limits (rate, kill switch, dry-run). Not a substitute for HITL.

analyze_winapp

Crawl a Windows app and produce UI element map + screenshot portfolio.

start_app_monitoring

Start monitoring sensitive applications.

start_intruder_detection

Start intruder detection.

get_app_monitoring_status

Get the current status of application monitoring.

get_intruder_detection_status

Get the current status of intruder detection.

stop_app_monitoring

Stop application monitoring.

stop_intruder_detection

Stop intruder detection.

get_security_events

Get security events.

enroll_face

Enroll a new face for recognition.

verify_face

Verify a face against known faces.

verify_face_webcam

Verify a face using the webcam.

remove_known_face

Remove a known face by name.

cua_computer_use_screenshotbehind config

Window-scoped screenshot for Claude Code computer-use style grounding. Requires window_handle (HWND). Does not activate the window when windows_computer_use_mcp_DISPATCH=background. Wraps automation_visual(screenshot) with return_base64=true.

get_desktop_statebehind config

Captures the complete hierarchical state of the Windows desktop UI. WHAT IT DOES: This tool traverses the UI tree of all active windows to produce a structured JSON representation of every discoverable element (buttons, panes, edit boxes, etc.). It optionally integrates computer vision (capturing e…

automation_assertbehind config

UI verification for computer-use automation - stability polling, image diff, template/OCR asserts. WHAT IT DOES: Polls screenshots until the UI stabilizes, compares before/after images, and asserts expected visual or text state. Use after keyboard shortcuts or clicks to confirm the action succeeded…

automation_dialogbehind config

Windows file dialog helpers - set path field and confirm. WHAT IT DOES: Fills Save/Open/Export dialog path fields using clipboard paste (default) or typed entry, then confirms with Enter. Prefer this over manual ctrl+a + type sequences. WHEN TO USE: - submit_path: full flow after a Save As / Open …

automation_elementsbehind config

Comprehensive UI element interaction and inspection system. WHAT IT DOES: This tool provides granular control over individual UI elements (buttons, text fields, lists, etc.) within a target window. It handles complex selection logic based on Control IDs, Automation IDs, titles, or relative coordina…

automation_facebehind config

Industrialized local face recognition and biometric enrollment tool. WHAT IT DOES: This tool provides local-only biometric capabilities for registering and recognizing individuals via facial encodings. It can enroll new faces from static images, match unknown faces against a local encrypted databas…

automation_keyboardbehind config

Comprehensive keyboard input and sequence simulation system. WHAT IT DOES: This tool provides a unified interface for simulating keyboard input. It supports sequential typing ('type'), single key presses ('press'), complex multi-key combinations ('hotkey'), and modifier-key holding ('hold'). WHEN …

global_keyloggerbehind config

Session-wide keyboard event capture (Windows low-level hook via pynput). WHAT IT DOES: Records key press (and optionally key release) events into a bounded in-memory ring buffer. Use 'read' to retrieve the most recent events without stopping the listener. WHEN TO USE: - Debugging shortcut or focus…

automation_macrobehind config

Record, replay, and verify UI automation macros. WHAT IT DOES: Captures every tool call during a recording session into a replayable JSONL file. Replay executes the same sequence. Replay-with-verify checks each step's outcome against the recorded result (hash-based screenshot comparison when availa…

automation_missionbehind config

Autonomous multi-step orchestration and complex mission management tool. WHAT IT DOES: This tool elevates automation from individual clicks to high-level objectives. Give it a natural-language goal - it decomposes into steps, executes each with retry + verification, and returns pass/fail with evide…

automation_mousebehind config

Comprehensive mouse control and cursor simulation system. WHAT IT DOES: Moves, clicks, drags, and scrolls using **Win32** ``SetCursorPos`` / ``mouse_event`` (DPI-aware), not PyAutoGUI alone - reliable move/click/drag on scaled monitors. WHEN TO USE: - Use 'position' to get current screen coordinat…

automation_shortcutbehind config

Semantic application shortcuts - send by action name, not raw key codes. WHAT IT DOES: Maps app-specific action names (e.g. vroidstudio/export_vrm → F8) to keyboard sequences, optionally focuses the target HWND (win32 backend), and waits for UI stability after send. WHEN TO USE: - send: execute a …

automation_smartbehind config

Intent-based discovery and interaction across all windows. OPERATIONS: - discover: Scan all visible windows and return a structured inventory of running apps and their controls. Optional query filters by app name or window title. - list_apps: Quick list of identified running applications with win…

automation_systembehind config

Industrialized Windows system utility and synchronization tool. WHAT IT DOES: This tool provides low-level system utilities including application launching, process monitoring, clipboard manipulation, and advanced synchronization primitives (e.g., waiting for windows to appear). It acts as the 'glu…

automation_taskbehind config

Closed-loop task runner - the MVP computer use assistant. WHAT IT DOES: Executes a list of steps (shortcut, dialog, wait_stable, assert_file, focus, screenshot, click, preflight) with automatic retry, refocus recovery, and per-step evidence (screenshot paths). WHEN TO USE: - Prefer over manual too…

automation_visualbehind config

Comprehensive computer vision and OCR tool for Windows UI automation. WHAT IT DOES: This tool provides multimodal capabilities for interacting with the desktop when traditional UI element selectors (UIA/MSAA) are insufficient or unavailable (e.g., in remote desktops, games, or legacy apps). It supp…

automation_watchbehind config

Persistent watcher for event-driven automation triggers. OPERATIONS: - start: Begin monitoring for a condition. Runs in a background thread. Conditions: window_appears, window_closes, text_appears, element_appears. Returns a watcher_id for status/cancel. - status: Check whether a watcher has tr…

automation_windowsbehind config

Comprehensive Windows window management and lifecycle control system. WHAT IT DOES: This tool provides a unified interface for discovering, locating, and controlling windows on the Windows desktop. It handles window state transitions (maximize, minimize, restore), visibility, and spatial positionin…

get_window_statebehind config

Capture UI state for one window (Cua Driver-shaped loop). WHAT IT DOES: Returns a snapshot_id, element_index-addressable elements, and optional set-of-mark screenshot for a single HWND. Use before automation_elements(click, snapshot_id=..., element_index=...). capture_mode: - ax: accessibility tre…

health_checkbehind config

Check if the PyWinAuto MCP server is running and operational.

get_helpbehind config

Get comprehensive help and documentation for all available tools.

get_cursor_positionbehind config

Get current mouse cursor position.

click_at_positionbehind config

Click at specific screen coordinates.

move_mousebehind config

Move mouse to specific coordinates.

scroll_mousebehind config

Scroll mouse wheel at current or specified position.

list_windowsbehind config

List all visible windows on the desktop.

find_window_by_titlebehind config

Find window by title text.

type_textbehind config

Type text at current cursor position.

send_keysbehind config

Send special key combinations.

hover_mousebehind config

Move mouse to coordinates and hover for specified duration.

waitbehind config

Pause execution for specified seconds.

get_system_clipboardbehind config

Get current clipboard content.

set_system_clipboardbehind config

Set clipboard content.

click_elementbehind config

Click on a UI element.

double_click_elementbehind config

Double-click on a UI element.

right_click_elementbehind config

Right-click on a UI element.

hover_elementbehind config

Hover the mouse over a UI element.

get_element_infobehind config

Get detailed information about a UI element.

get_element_textbehind config

Get the text of a UI element.

set_element_textbehind config

Set the text of a UI element.

get_element_rectbehind config

Get the rectangle of a UI element.

is_element_visiblebehind config

Check if a UI element is visible.

is_element_enabledbehind config

Check if a UI element is enabled.

get_all_elementsbehind config

Get all UI elements in a window.

element_existsbehind config

Check if a UI element exists.

wait_for_elementbehind config

Wait for a UI element to appear.

verify_textbehind config

Verify that an element contains the expected text.

add_facebehind config

Add a new face to the recognition system.

recognize_facebehind config

Recognize faces in an image.

list_known_facesbehind config

List all known faces in the system.

delete_facebehind config

Delete a known face from the system.

capture_and_recognizebehind config

Capture an image from the webcam and recognize faces in it.

press_keybehind config

Press a key or combination of keys.

press_hotkeybehind config

Press a combination of keys (hotkey/shortcut).

drag_mousebehind config

Drag the mouse from one point to another.

get_mouse_positionbehind config

Get the current mouse cursor position.

mouse_move_relativebehind config

Move mouse relative to current position.

mouse_move_to_elementbehind config

Move mouse to a UI element.

mouse_hoverbehind config

Hover over an element.

drag_and_dropbehind config

Drag from source to target element.

right_clickbehind config

Right-click on an element or at specific coordinates.

double_clickbehind config

Double-click on an element or at specific coordinates.

mouse_scrollbehind config

Scroll the mouse wheel.

wait_for_windowbehind config

Wait for a window with the specified title to become active.

bring_window_to_foregroundbehind config

Bring a window to the foreground by its title.

get_process_listbehind config

Get a list of running processes.

take_screenshotbehind config

Take a screenshot of the entire screen or a specific window.

extract_textbehind config

Extract text from an image or screen region using OCR.

find_imagebehind config

Find a template image within a screenshot or window.

highlight_elementbehind config

Highlight a UI element with a colored rectangle.

find_image_on_screenbehind config

Find an image on the screen using template matching.

get_text_under_cursorbehind config

Get text at the current cursor position using OCR.

get_ui_treebehind config

Get a hierarchical representation of the UI tree.

maximize_windowbehind config

Maximize a window.

minimize_windowbehind config

Minimize a window.

restore_windowbehind config

Restore a window to its normal state.

set_window_positionbehind config

Set the position and size of a window.

get_active_windowbehind config

Get the currently active window.

close_windowbehind config

Close a window.

get_window_rectbehind config

Get the rectangle of a window.

get_window_titlebehind config

Get the title of a window.

set_window_foregroundbehind config

Brings a window to the foreground and activates it.

get_all_windowsbehind config

Get information about all visible windows.

// environment variables

To run this server yourself, you supply these values. They go in your own MCP client configuration and stay on your machine. The secret label means the value is sensitive, not that the server mishandles it.

configPLAYWRIGHT_BASE_URL

configPYWINAUTO_WEB_PORT

configPYWINAUTO_TAURI

configMCP_PORT

configMCP_HOST

configHOST

configwindows_computer_use_mcp_BYPASS_HITL

configWINDOWS_COMPUTER_USE_MCP_VOICE

configWINDOWS_COMPUTER_USE_MCP_OCR_PROVIDER

configMCP_CENTRAL_DOCS

configWINDOWS_COMPUTER_USE_MCP_BYPASS_HITL

configwindows_computer_use_mcp_AGENT_OVERLAY

configwindows_computer_use_mcp_DEPOT_URL

configDEPOT_URL

configFLEET_ERROR_DB

configSystemDrive

configSYSTEM_ADMIN_MCP_URL

configPYWINAUTO_MCP_API_URL

configSPEECH_MCP_URL

configwindows_computer_use_mcp_CAMERA_MAX_INDEX

configwindows_computer_use_mcp_EVIDENCE_DIRS

configPYWINAUTO_LLM_BASE_URL

Deployment configuration, supplied by whoever hosts the server. Users are not asked for these.

deployPORT

// quality suggestions

Tool annotations

No tools have read-only/destructive annotations

Add readOnlyHint or destructiveHint annotations to every tool so hosts can warn users before invoking.

All four hints declared on every tool

113/113 tools missing one or more hints — approve_automation (missing: readOnlyHint, destructiveHint, idempotentHint, openWorldHint); automation_safety (missing: readOnlyHint, destructiveHint, idempotentHint, openWorldHint); cua_computer_use_screenshot (missing: readOnlyHint, destructiveHint, idempotentHint, openWorldHint), +110 more. OpenAI's directory rejects tools where any of the four hints are missing or non-boolean.

For every tool, set all four hints (readOnlyHint, destructiveHint, idempotentHint, openWorldHint) to explicit true/false values that match the handler’s actual behaviour.

Destructive tools are labelled

6 tools perform destructive updates without destructiveHint — automation_face deletes at line 388 (file_path.unlink()); automation_visual deletes at line 257 (os.remove(temp_path)); capture_and_recognize deletes at line 496 (os.unlink(temp_path))

Add destructiveHint:true to any tool whose handler calls .delete(), .upsert(), .update(), unlink, rm, DELETE, DROP, REPLACE INTO, or any operation that overwrites existing data.

Descriptions match behaviour

Rename the tool, rewrite the description, or move the side-effect into a separate clearly-named tool.

Tool test coverage

38/113 tools referenced in tests (34%)

Write tests that reference each tool by name so every tool has at least one test.

Tool description accuracy

3 tools have description/behavior mismatches: automation_mouse: description implies read-only but handler writes/deletes/executes; automation_visual: description implies read-only but handler writes/deletes/executes; find_image: description implies read-only but handler writes/deletes/executes

Update tool descriptions to accurately reflect all capabilities — especially write, delete, or execute operations.

Claim the listing to review these findings one by one and send us a correction where you disagree, straight to the team. Claiming also means we tell you when the grade moves, and reach you first if we find anything urgent.

// full audit trail

The findings above are the summary. The full trail, every check we ran, each deduction, the network hosts observed and the dependency advisories, goes to verified publishers, along with an alert whenever a new one lands. Verified publishers can also review each finding and dispute it in one click. Publisher corrections have sharpened several of our checks this month, because the maintainer knows the codebase better than any scanner.

// improvement guidance — verified publishers only

We have 6 concrete improvements we can share with the publisher of this MCP. Each comes with specific guidance to raise the trust score.

// embed badge in your README

[![M8ven Verified](https://m8ven.ai/badge/mcp/sandraschi/windows-computer-use-mcp?variant=verified)](https://m8ven.ai/mcp/sandraschi/windows-computer-use-mcp)

Shows verification status without the grade. Want the grade badge instead? Remove ?variant=verified from the URL.

commit: fe73e1f787868d0a17fe17ccdffd0fcef0384fb9

code hash: 47f57eb1ff2876fd7f3e0950c29ef92bce13786b4eb85ea864f9a34e64a66efe

view raw JSON →

Check MCPs from inside your assistant

Tool Check · MCP

Vetting this one by hand? Tool Check is an MCP that scores other MCPs. Add it once and ask Claude, ChatGPT, or any MCP client to grade a server, surface CVEs, check the publisher, and suggest safer alternatives — before you install.

https://m8ven.ai/api/mcp/tool-check

check_toolsearch_toolscompare_toolsrecommend_alternativescheck_publisherreport_concern

How to add it →Free · no account needed · works in any MCP client