Desktop Touch MCP
未认领Windows computer-use MCP server: drive any desktop app via semantic discover-then-act targeting (entities + leases, not pixel coordinates), with per-action perception guards, a native Rust UIA engine, Chrome CDP, and Key Locker credential autofill for ssh/sudo password prompts. Works with Claude, Cursor, and any MCP client.
01 / 选择客户端
03 / 添加配置
claude mcp add --transport stdio 'desktop-touch-mcp' -- 'npx' '-y' '@harusame64/desktop-touch-mcp'04 / 在客户端中确认
打开客户端的 MCP 设置，确认服务器已连接并列出工具。复制配置不代表连接成功。
更多Files & Storage服务器
浏览完整目录安全概况
已认领和已认证的服务器每周接受一次静态扫描，显示代码能触及的范围（外部服务、环境变量、shell 命令、代理配置目录）以及带有已知公告的依赖。认领此列表即可获得。 安全概况的工作原理
16 个工具中显示 16 个
文档中列出的工具 (16)
内容来自项目文档。服务器握手不会验证每个工具的说明或行为。
browser_eval
Deep inspection via js (scripting), dom (HTML), and appState (SPA data extraction).
click_element
Legacy UIA-based click by name/ID (fallback when entities are unavailable).
desktop_act
Perform actions (click, type, drag, select) on entities via lease validation. Returns semantic diffs — plus an optional roiCapture (changed-region PNG + next-target preview) on visual-only targets.
desktop_discover
Observe the desktop. Returns interactive entities with leases (UIA, CDP, Terminal, Visual SoM).
desktop_state
Lightweight check of focus, active window, cursor, and Auto-Perception attention signal.
excel
Author and run Excel VBA macros via COM. action='runvba' writes a macro into a managed Trusted Location and runs it; action='checkaccessvbom' is a read-only preflight. Runs VBA where formula-only tools cannot. One-time setup: node scripts/enable-access-vbom.mjs.
key_locker
Manage credentials the terminal autofills for you (SSH key passphrases, sudo / login passwords). Secrets are entered once into the locker's own secure dialog and stored encrypted on this machine (Windows DPAPI); they are never shown to the assistant. action='launchconsole' opens an autofill-capable
keyboard
Send keyboard input. Supports background input (WMCHAR) and IME-safe clipboard bypass.
run_macro
Batch up to 50 operations into a single round-trip for maximum efficiency.
screenshot
Multi-mode capture: detail='text' (UIA/OCR), diffMode (P-frame), dotByDot (1:1), and background. Returns a cheap screenshot://by-ref/{id} link to the saved image instead of inlining pixels every time.
scroll
Multi-strategy: raw (notches), toelement, smart (virtual lists), and capture (stitch).
server_status
Diagnostic check for native engine health and feature activation.
terminal
Unified command execution: run (send + wait + read), read (OCR/UIA), and send. run completion modes: quiet, pattern, and exit (waits for the command to finish + returns its exit code — see Terminal command completion).
wait_until
Efficient server-side polling for window, focus, text, or URL state changes.
workspace_launch
Launch apps and auto-detect new HWNDs (supports localized titles).
workspace_snapshot
Instant session orientation: all window thumbnails + UI summaries in one call.