A Model Context Protocol (MCP) server for interacting with Microsoft 365 and Office services through the Graph API
Nothing to run from here
Windows is installed from its publisher's own source and answers where it runs, so this marketplace is not in the path of a single call. There is no address here to send one to, and a panel that pretended otherwise would be showing you an answer we made up. Install it and call it from your own client — the Installation tab has the entry for each one.
93/100
trust score
2k
downloads a day
local
runs on your machine
46k
downloads in 30d
What it does
An MCP Server for computer-use in Windows OS
Quickstart
# 1 — add it to Claude Code
claude mcp add --scope user cursortouch-windows-mcp '--' uvx windows-mcp
# 2 — ask your agent for something these tools do
# App, Registry, Clipboard and 18 more
Windows costs nothing on mcprush: there is no plan to choose here, no cap to set and nothing mcprush can bill you for.
Publisher
Installation
Where are you running it?
Every route below installs the same thing and ends at the same approval screen. This one runs on your machine: your client starts Windows as a process under your own user, with your files and your network, so the tool surface below is what it can do to your computer rather than to a server somewhere else. Its scan result is on the Security tab — read the surface before you approve it.
Claude Codedeclared by this listing
This is a public server: you run it yourself and this marketplace is not in the path. You need uv (docs.astral.sh/uv): uvx comes with it. Claude Code registers it in one command; --scope user makes it available in every project. The '--' is in quotes so that PowerShell passes it on.
claude mcp add --scope user cursortouch-windows-mcp '--' uvx windows-mcpReconnect, or start a new session, and the tools appear in the model’s tool list.
Free on mcprush, and it runs from its publisher’s own package: this marketplace is not in the path, so there is no key of ours to hold and no card to add.
What it writes
One config entry your client uses to start the process locally. A local server runs with your file system and your network, which is why it is priced without metering.
What you approve
21 tools, with what each one reads, writes and reaches shown before you agree — the same list on every route above. Read the tool surface.
Tool surface
What the model actually sees. Descriptions are diffed on every release — see version history.
Open/start/launch applications and manage windows. Keywords: open, start, launch, program, application, window, foreground, focus, resize. Four modes: 'launch' (opens an application by Start Menu name), 'launch_executable' (strictly launches one executable path with separated argv and optional…
| Parameter | Type | Required | Notes |
|---|---|---|---|
| mode | string | no | — |
| name | string | no | — |
| window_loc | array | no | — |
| window_size | array | no | — |
| executable | string | no | — |
| args | array/string | no | — |
| cwd | string | no | — |
Read and write the Windows Registry. Keywords: regedit, registry key, HKEY, HKCU, HKLM, Windows settings, registry value. Use mode="get" to read a value, mode="set" to create/update a value, mode="delete" to remove a value or key, mode="list" to list values and sub-keys under a path. Paths use…
| Parameter | Type | Required | Notes |
|---|---|---|---|
| mode | string | yes | — |
| path | string | yes | — |
| name | string | no | — |
| value | string | no | — |
| type | no | — |
Copy/paste clipboard operations. Keywords: copy, paste, cut, clipboard, text transfer. Use mode="get" to read current clipboard content, mode="set" to set clipboard text.
| Parameter | Type | Required | Notes |
|---|---|---|---|
| mode | string | yes | — |
| text | string | no | — |
Get desktop control owner, enabled state, configured idle timeout and remaining user idle wait in seconds. Settings are read-only and changes require a server restart.
No parameter schema on file.
Read active display layout and DPI metadata. Reports display index, device name, monitor/work-area bounds, resolution, orientation, primary flag, effective DPI, and scale.
No parameter schema on file.
Manages file system operations with eight modes: 'read' (read text file contents with optional line offset/limit), 'write' (create or overwrite a file, set append=True to append), 'copy' (copy file or directory to destination), 'move' (move or rename file/directory), 'delete' (delete file or…
| Parameter | Type | Required | Notes |
|---|---|---|---|
| mode | string | yes | — |
| path | string | yes | — |
| destination | string | no | — |
| content | string | no | — |
| pattern | string | no | — |
| recursive | boolean/string | no | — |
| append | boolean/string | no | — |
| overwrite | boolean/string | no | — |
| offset | integer | no | — |
| limit | integer | no | — |
| encoding | string | no | — |
| show_hidden | boolean/string | no | — |
Performs mouse clicks at specified coordinates [x, y] or passing a UI element's label/id. Supports button types: 'left' for selection/activation, 'right' for context menus, 'middle'. Supports clicks: 0=hover only (no click), 1=single click (select/focus), 2=double click (open/activate). Provide…
| Parameter | Type | Required | Notes |
|---|---|---|---|
| loc | array/string | no | — |
| label | integer | no | — |
| button | string | no | — |
| clicks | integer | no | — |
Scrolls at coordinates [x, y], a UI element's label/id, or current mouse position if loc=None. Type: vertical (default) or horizontal. Direction: up/down for vertical, left/right for horizontal. wheel_times controls amount (1 wheel ≈ 3-5 lines). Use for navigating long content, lists, and web pages.
| Parameter | Type | Required | Notes |
|---|---|---|---|
| loc | array/string | no | — |
| label | integer | no | — |
| type | string | no | — |
| direction | string | no | — |
| wheel_times | integer | no | — |
Moves mouse cursor to coordinates [x, y] or passing a UI element's label/id. Set drag=True to perform a drag-and-drop operation from the current mouse position to the target coordinates, or provide from_loc=[x, y] to make the drag explicit-start and atomic in one tool call. Optional duration…
| Parameter | Type | Required | Notes |
|---|---|---|---|
| loc | array/string | no | — |
| label | integer | no | — |
| drag | boolean/string | no | — |
| from_loc | array/string | no | — |
| duration | number/integer/string | no | — |
Executes keyboard shortcuts using key combinations separated by +. Examples: "ctrl+c" (copy), "ctrl+v" (paste), "alt+tab" (switch apps), "win+r" (Run dialog), "win" (Start menu), "ctrl+shift+esc" (Task Manager). Use for quick actions and system commands.
| Parameter | Type | Required | Notes |
|---|---|---|---|
| shortcut | string | yes | — |
Pauses execution for specified duration in seconds. Use when waiting for: applications to launch/load, UI animations to complete, page content to render, dialogs to appear, or between rapid actions. Helps ensure UI is ready before next interaction.
| Parameter | Type | Required | Notes |
|---|---|---|---|
| duration | integer | yes | — |
Waits until a UI condition is satisfied, polling the Windows accessibility tree inside the tool to avoid repeated Snapshot calls. Conditions: text_exists, active_window, element_exists, element_enabled, focused_element. Provide text and/or window_name depending on the condition. Set use_dom=True…
| Parameter | Type | Required | Notes |
|---|---|---|---|
| condition | string | yes | — |
| text | string | no | — |
| window_name | string | no | — |
| timeout | number | no | — |
| interval | number | no | — |
| use_dom | boolean/string | no | — |
Selects multiple items such as files, folders, or checkboxes if press_ctrl=True, or performs multiple clicks if False. Pass locs (list of coordinates) or labels (list of UI element labels/ids).
| Parameter | Type | Required | Notes |
|---|---|---|---|
| locs | array/string | no | — |
| labels | array/string | no | — |
| press_ctrl | boolean/string | no | — |
Enters text into multiple input fields at specified coordinates locs=[[x,y,text], ...] or using labels=[[label,text], ...]. Provide either locs or labels.
| Parameter | Type | Required | Notes |
|---|---|---|---|
| locs | array/string | no | — |
| labels | array/string | no | — |
Sends a Windows toast notification with a title and message.
| Parameter | Type | Required | Notes |
|---|---|---|---|
| title | string | yes | The title/heading of the toast notification. |
| message | string | yes | The body text of the toast notification displayed below the title. |
| app_id | string | yes | The valid Application User Model ID of the toast notification. Required to display the notification in a specific app. |
List and kill running processes. Keywords: task manager, running tasks, kill, terminate, stop process, PID, CPU, memory usage. Use mode="list" to list running processes with filtering and sorting options. Use mode="kill" to terminate processes by PID or name.
| Parameter | Type | Required | Notes |
|---|---|---|---|
| mode | string | yes | — |
| name | string | no | — |
| pid | integer | no | — |
| sort_by | string | no | — |
| limit | integer | no | — |
| force | boolean/string | no | — |
Fetch/scrape web page content from a URL. Keywords: scrape, fetch, browse, web, URL, extract, download, read webpage. By default (use_dom=False), performs a lightweight HTTP request to the URL and returns a clean LLM-processed summary of the page to avoid context bloat. Provide query to focus…
| Parameter | Type | Required | Notes |
|---|---|---|---|
| url | string | yes | — |
| query | string | no | — |
| use_dom | boolean/string | no | — |
| use_sampling | boolean/string | no | — |
Shell/command execution. Keywords: shell, run, execute, cmd, terminal, command line, script. A comprehensive system tool for executing any PowerShell commands. Use it to navigate the file system, manage files and processes, and execute system-level operations. Capable of accessing web content…
| Parameter | Type | Required | Notes |
|---|---|---|---|
| command | string | yes | — |
| timeout | integer | no | — |
Take a screenshot and inspect the screen. Keywords: screenshot, screen capture, see screen, observe, look, inspect, UI elements, what's on screen. Captures complete desktop state including: system language, focused/opened windows, interactive elements (buttons, text fields, links, menus with…
| Parameter | Type | Required | Notes |
|---|---|---|---|
| use_vision | boolean/string | no | — |
| use_dom | boolean/string | no | — |
| use_annotation | boolean/string | no | — |
| use_ui_tree | boolean/string | no | — |
| width_reference_line | integer | no | — |
| height_reference_line | integer | no | — |
| display | array | no | — |
| region | array/string | no | — |
Captures a fast screenshot-first desktop snapshot with cursor position, desktop/window summaries, and an image. This path skips UI tree extraction for speed. Use Snapshot when you need interactive element ids, scrollable regions, or browser DOM extraction. Set display=[0] or display=[0,1] using…
| Parameter | Type | Required | Notes |
|---|---|---|---|
| use_annotation | boolean/string | no | — |
| width_reference_line | integer | no | — |
| height_reference_line | integer | no | — |
| display | array | no | — |
| region | array/string | no | — |
Types text at specified coordinates [x, y] or passing a UI element's label/id. Set clear=True to clear existing text first, False to append. Set press_enter=True to submit after typing. Set caret_position to 'start' (beginning), 'end' (end), or 'idle' (default). Provide either loc or label. clear…
No parameter schema on file.
Free
$0 forever
- Every tool, no call limit
- No card, and mcprush charges nothing for it
- Source published under MIT
- Runs on your machine — nothing of it reaches our gateway
- Nothing to cap, because mcprush bills nothing
What counts against your monthly calls
| Tool | Unit | Calls used | Out of the allowance |
|---|
Nothing here is billable: mcprush charges nothing to install Windows or to call it, at any volume.
Call allowance · this installenforced at the gateway
Monthly allowance, in calls
—
—
Alert at 80% · then the tool returns a normal MCP error
Used this month: —
ATrust score 93 / 100
Release 0.8.8 · scanned 14 Oct 2026
Two independent axes, because powerful and malicious are different questions. The grade is threat only. The capability level is blast radius, and it is never a penalty on the grade — it is priced as one subtract-only term in the score, where you can see it.
Threat score100
Client score93
Findings7 threat · 2 capability
Capabilityhigh
Coveragesource
Verificationrepo
Surface digestdd32e99012f68fbb3ce67ad4570aa39900a54ca142e7ca26842d8b078da6bf98
Score vectorthreat 100 → client 93
| Term | Level | What it prices | Points |
|---|---|---|---|
| capability-exposure | high | capability blast radius (high) — client exposure if the model is manipulated | −6 |
| verification-discount | repo | publisher verification (public source) — no provenance, but the source is public and inspectable | −1 |
| coverage-honesty | source | inspection depth (source) — how much of the target the scan could see | −0 |
score = clamp(round(threatScore − E_cap − E_ver − E_cov), 0, 100). No model, no opinion, no network call at scoring time: the same methodology version against the same bytes always produces this number, and every line above is a field of the report rather than a summary of it.
What the scan could actually read
A grade is only as meaningful as its coverage, so the scanner publishes its own depth before it publishes its result.
Tool surfaceread
Implementation sourceread
Package metadataread
Live transportnot available
Tools were statically extracted from the published source (21 recovered), not enumerated from a running server. Tool-poisoning, Unicode-smuggling, capability and toxic-flow analysis ran on this inferred surface, but a mis-parsed registration could be missed or mis-attributed, so tool-derived findings are capped below “confirmed”. To grade the real runtime surface, scan the running server: --command "npx -y <package>".
Capability — what it could do if the model were manipulated
Tags derived from each tool’s schema and the implementation, not from what the tool calls itself. high is the level these add up to.
| Tool | Capability tags | Why the tag was assigned |
|---|---|---|
| App | state-write | state-write: declares readOnlyHint: false, destructiveHint: true |
| Clipboard | state-write | state-write: declares readOnlyHint: false |
| ControlStatus | no tags | |
| DisplayInventory | no tags | |
| FileSystem | state-write | state-write: declares readOnlyHint: false, destructiveHint: true |
| Click | state-write | state-write: declares readOnlyHint: false, destructiveHint: true |
| Type | state-write | state-write: declares readOnlyHint: false, destructiveHint: true |
| Scroll | state-write | state-write: declares readOnlyHint: false |
| Move | state-write | state-write: declares readOnlyHint: false, destructiveHint: true |
| Shortcut | state-write | state-write: declares readOnlyHint: false, destructiveHint: true |
| Wait | no tags | |
| WaitFor | no tags | |
| MultiSelect | state-write | state-write: declares readOnlyHint: false, destructiveHint: true |
| MultiEdit | state-write | state-write: declares readOnlyHint: false, destructiveHint: true |
| Notification | state-write | state-write: declares readOnlyHint: false, destructiveHint: true |
| Process | state-write | state-write: declares readOnlyHint: false, destructiveHint: true |
| Registry | state-write | state-write: declares readOnlyHint: false, destructiveHint: true |
| Scrape | untrusted-inputexternal-sink | untrusted-input: keyword "fetch", declares openWorldHint: true (interacts with entities outside the server); external-sink: keyword "http_request" |
| PowerShell | code-execexternal-sink | code-exec: keyword "powershell" in tool name; external-sink: declares readOnlyHint: false, destructiveHint: true, openWorldHint: true |
| Snapshot | no tags | |
| Screenshot | no tags |
Toxic-flow graph
The lethal trifecta, checked as a graph rather than as a checklist: untrusted input, a sensitive source and an external sink have to meet before there is a path worth worrying about.
flow-1
Untrusted input can drive an external action even though no sensitive source is exposed.
Supply chain and provenance
Integrityfirst-seen
Surface digestdd32e99012f68fbb
Scanned2026-10-14
This is the first scan of this surface here, so there is nothing yet to compare it against.
Every result on this tab comes from one deterministic pass over the published package — offline, rule by rule, and auditable line by line above. Same methodology version, same bytes, same score.
Release history
Always latest — the install command names no version, so it installs the newest release when it runs; the current release on file is 0.8.8. Pin a release below and the command asks for exactly that one.
0.8.8
10 Oct 2026
latestNo release note was published with this version.
0.8.7
30 Sep 2026
No release note was published with this version.
0.8.6
26 Sep 2026
No release note was published with this version.
0.8.5
1 Aug 2026
No release note was published with this version.
0.8.1
undated
- The version this listing was on when mcprush began following its releases. The registry was not asked when it shipped, so this row carries no date.
—
no reviews yet
Who can review
One review per account, from a signed-in account that does not publish this listing. Nothing else is asked — you can say what you think before you install it. Publishers can reply once per review, and a review can be edited or deleted by whoever wrote it.
Sign in to write one
Any signed-in account that does not publish Windows can review it, once — before installing it or after. A review from an account that has installed it is marked as one.
No reviews yet
Nobody has reviewed this listing. The rating on the card is the mean of the reviews written here and nothing else, so there is no rating until somebody writes the first — which takes a signed-in account that does not publish it, and nothing else.