@harusame64/desktop-touch-mcp 1.4.4 → 1.5.1

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
package/README.ja.md CHANGED
@@ -4,9 +4,13 @@
4
4
 
5
5
  [English](README.md)
6
6
 
7
- > **「座標ルーレット」はもう不要。セマンティック・ワールドグラフと自動認識ガードによる、LLMネイティブなWindows自動化。**
7
+ > **Windows 用 computer-use MCP サーバー。** Claude / Cursor / VS Code Copilot などの MCP クライアントから、あなたの Windows 10/11 デスクトップを「見て」「操作」させられます — スクリーンショット、UI Automation、Chrome CDP、キーボード / マウス、ターミナル。座標ルーレットではない **セマンティックな discover-then-act 設計** と、誤ウィンドウへの入力を未然に防ぐ **action 毎の perception guard** が特徴です。
8
8
 
9
- ClaudeがWindowsを直接「見て」「操作する」ためのMCPサーバー。単なるピクセル指定の座標当てゲームを卒業し、**セマンティック・ワールドグラフ** (`desktop_discover`) による確実なエンティティ特定と、**自動認識(Auto-Perception)ガード**による安全な実行を実現します。28個の最適化されたツール群が、バックグラウンド入力、UIA、Chrome CDP、ターミナル、そしてトークン効率に優れたP-frame差分転送をサポートします。
9
+ ```bash
10
+ npx -y @harusame64/desktop-touch-mcp
11
+ ```
12
+
13
+ 28 ツール、Rust ネイティブエンジン (UIA 2ms)、PowerShell 透過フォールバック、日本語/CJK 完全対応、MIT。上記 1 行を Claude / Cursor / VS Code Copilot の MCP 設定に追加するだけで、Notepad、Excel、Chrome、Windows Terminal、その他あらゆるアプリを Claude が操作できるようになります。
10
14
 
11
15
  > *v0.15: Rust ネイティブエンジンにより**平均 82 倍高速化** — UIA フォーカス取得 2ms、SSE2 SIMD 画像差分 13〜15 倍速。設定不要:エンジンは自動ロード、不在時は PowerShell に透過フォールバック。*
12
16
  > *v0.15.5: **固定リリース検証** — npm ランチャーは対応する GitHub Release tag だけを取得し、Windows runtime zip を検証してから展開します。*
package/README.md CHANGED
@@ -4,9 +4,13 @@
4
4
 
5
5
  [日本語](README.ja.md)
6
6
 
7
- > **Beyond Coordinate Roulette: LLM-native Windows automation with a Semantic World-Graph and Auto-Perception.**
7
+ > **Computer-use MCP server for Windows.** Lets Claude, Cursor, or any MCP client see and operate your Windows 10/11 desktop — screenshots, UI Automation, Chrome CDP, keyboard / mouse, terminal — with **semantic discover-then-act targeting** that avoids pixel-coordinate guessing, and **per-action perception guards** that catch wrong-window typing before it happens.
8
8
 
9
- An MCP server that gives Claude eyes and hands on Windows. It moves beyond pixel-guessing by grounding all interactions in a **Semantic World-Graph** (`desktop_discover`) and verifying every action with **Auto-Perception** guards. Optimized into 28 high-signal tools covering screenshots, background input (WM_CHAR), UIA, Chrome CDP, terminal, and token-efficient P-frame diffing.
9
+ ```bash
10
+ npx -y @harusame64/desktop-touch-mcp
11
+ ```
12
+
13
+ 28 tools, native Rust engine (UIA in 2 ms), zero-config PowerShell fallback, full CJK support, MIT licensed. Add the snippet above to your Claude / Cursor / VS Code Copilot config and Claude can drive Notepad, Excel, Chrome, Windows Terminal, and any other app on your machine.
10
14
 
11
15
  > *v0.15: **82× average speedup** via Rust native engine — UIA focus queries in 2 ms, SSE2-accelerated image diffing at 13–15× native speed. Zero-config: the engine auto-loads when present, with transparent PowerShell fallback.*
12
16
  > *v0.15.5: **Pinned release verification** — the npm launcher now fetches only the matching GitHub Release tag and verifies the Windows runtime zip before extraction.*
package/bin/launcher.js CHANGED
@@ -18,15 +18,15 @@ import path from "node:path";
18
18
  import { Readable } from "node:stream";
19
19
  import { pipeline } from "node:stream/promises";
20
20
 
21
- const PACKAGE_VERSION = "1.4.4";
21
+ const PACKAGE_VERSION = "1.5.1";
22
22
  const RELEASE_TAG = `v${PACKAGE_VERSION}`;
23
23
  const REPO_API_URL = `https://api.github.com/repos/Harusame64/desktop-touch-mcp/releases/tags/${RELEASE_TAG}`;
24
24
  const ASSET_NAME = "desktop-touch-mcp-windows.zip";
25
25
  const RELEASE_METADATA_FILE = ".desktop-touch-release.json";
26
26
  const RELEASE_MANIFEST = {
27
- tagName: "v1.4.4",
27
+ tagName: "v1.5.1",
28
28
  assetName: ASSET_NAME,
29
- sha256: "37c01455936b1b4fcda514976b4e5835094b0e303c41a0f2d60296f3195761c8",
29
+ sha256: "cd5943e7227ae5e1cc1280583f792bb674f095a6a724ab3de955c32a6f15489f",
30
30
  };
31
31
  const CACHE_ROOT = process.env.DESKTOP_TOUCH_MCP_HOME
32
32
  ? path.resolve(process.env.DESKTOP_TOUCH_MCP_HOME)
package/package.json CHANGED
@@ -1,8 +1,34 @@
1
1
  {
2
2
  "name": "@harusame64/desktop-touch-mcp",
3
- "version": "1.4.4",
3
+ "version": "1.5.1",
4
4
  "mcpName": "io.github.Harusame64/desktop-touch-mcp",
5
- "description": "LLM-native Windows computer-use MCP server with 28 tools for screenshots, UIA, mouse/keyboard, Chrome CDP, terminal, SmartScroll, and perception guards",
5
+ "description": "Let Claude, Cursor, or any MCP client see and operate your Windows 10/11 desktop. 28 tools for screenshots, UI Automation, Chrome CDP, keyboard/mouse, terminal, with semantic discover-then-act targeting and per-action perception guards that avoid wrong-window typing and stale-coordinate clicks.",
6
+ "keywords": [
7
+ "mcp",
8
+ "mcp-server",
9
+ "model-context-protocol",
10
+ "claude",
11
+ "claude-desktop",
12
+ "claude-code",
13
+ "anthropic",
14
+ "cursor",
15
+ "vscode-copilot",
16
+ "computer-use",
17
+ "agentic",
18
+ "ai-agent",
19
+ "llm",
20
+ "windows",
21
+ "windows-automation",
22
+ "win32",
23
+ "uia",
24
+ "ui-automation",
25
+ "screenshot",
26
+ "screen-capture",
27
+ "chrome-cdp",
28
+ "chrome-devtools-protocol",
29
+ "terminal-automation",
30
+ "desktop-automation"
31
+ ],
6
32
  "engines": {
7
33
  "node": ">=20.0.0"
8
34
  },