← 返回 Skills 市场
172
总下载
1
收藏
0
当前安装
1
版本数
在 OpenClaw 中安装
/install computer-use-linux
功能描述
Top-level Linux computer-use skill with a bundled standalone runtime that bootstraps itself without any local Claude installation, private native modules, or...
使用说明 (SKILL.md)
Linux Computer-Use Skill
Use this skill when the task needs a portable Linux computer-use skill bundled with its own standalone runtime and MCP server.
What this skill does
- uses the bundled
linux-computer-use-skillproject under the installed skill directory - builds the standalone MCP server
- lets the server auto-bootstrap its Python runtime on first launch
- avoids any dependency on local Claude binaries,
.nodemodules, or extracted app assets - stays explicitly Linux-only because the underlying desktop-control backend is Linux-specific
Default bundled project path
After installation, assume the standalone project lives at:
~/.codex/skills/computer-use-linux/project
If the user installed the skill under a custom CODEX_HOME, use that equivalent path instead.
Build
Always build from the bundled project:
cd ~/.codex/skills/computer-use-linux/project
npm install
npm run build
Run
cd ~/.codex/skills/computer-use-linux/project
node dist/cli.js
The first real run will automatically create .runtime/venv and install the public Python dependencies.
Validation notes
- Version
0.1.1fixes the shared system-key blocklist and tool capability typings so Linux builds use Linux-specific shortcut protections instead of a broken copied platform branch. - This project has been statically validated from macOS for TypeScript build health, bundled project integrity, and Python helper syntax, but still needs end-to-end runtime validation on a real Linux desktop.
Guardrails
- Treat this host as trusted-local only.
- Do not tell the user to search their local Claude install for binaries or hidden assets.
- Be explicit that this runtime is standalone and uses public dependencies only.
- Mention that the current runtime reports
screenshotFiltering: none, so action gating is handled at the MCP layer.
安全使用建议
What to consider before installing:
- Functionality & sensitivity: this skill captures screenshots, accesses clipboard/process info, enumerates installed/running apps, and can synthesize keyboard/mouse events — these are powerful and privacy-sensitive capabilities. Only install if you trust the author and intend to grant those abilities.
- Missing metadata: the skill fails to declare required binaries (Node >=20, npm, Python). Confirm those prerequisites on your system before installing, and request the author update the metadata.
- Network & disk activity: the build/run steps run `npm install` and will create a Python virtualenv and `pip install` packages from PyPI — expect network downloads and files written into your home directory (~/.codex/skills/... and a .runtime/venv). If you require offline or audited installs, review package-lock.json and runtime/requirements.txt and consider installing in a VM or sandbox first.
- Review code & scripts: inspect scripts/install.sh and the runtime (runtime/linux_helper.py and the TypeScript files) yourself or have someone you trust review them. Look specifically for subprocess.exec/execFile calls or any unexpected outbound network calls in the omitted files.
- Run safely: if you want to try it, run the server in an isolated environment (VM, container, or non-production account) and monitor network traffic and filesystem changes. Consider denying autonomous invocation until you’re comfortable with behavior, or run the skill only when explicitly invoked.
- If you need help: ask the author to (1) declare required binaries/env in registry metadata, (2) document exactly what files/paths it writes, and (3) provide a reproducible minimal install/test plan for offline review.
功能分析
Type: OpenClaw Skill
Name: computer-use-linux
Version: 0.1.1
The 'computer-use-linux' skill bundle is a comprehensive implementation of a Linux desktop automation tool using the Model Context Protocol (MCP). It provides capabilities for screenshots, mouse/keyboard control, and clipboard access via a Node.js CLI and a Python helper script (linux_helper.py). The project demonstrates a high degree of security awareness, incorporating robust features such as an application allowlist, a system-level shortcut blocklist (keyBlocklist.ts), and tiered permissions that restrict interaction with sensitive apps like browsers and terminals (deniedApps.ts). The runtime bootstrap process (pythonBridge.ts) and installation script (install.sh) follow standard practices for managing local environments. While the skill possesses high-risk capabilities inherent to desktop control, its behavior is strictly aligned with its stated purpose and includes significant safeguards to prevent abuse.
能力评估
Purpose & Capability
The skill's description says it's a standalone Linux computer-use runtime and MCP server, and the code implements X11 screenshotting, keyboard/mouse control, process/app enumeration, and first-run virtualenv bootstrapping — that matches purpose. However the registry metadata lists no required binaries/env but the SKILL.md and code require Node (>=20), npm, and a Python runtime (to be bootstrapped). The omission of required binaries in metadata is an incoherence developers should fix.
Instruction Scope
SKILL.md explicitly instructs building (npm install, npm run build) and running node dist/cli.js from the bundled project and allows the server to auto-bootstrap a Python venv and pip-install runtime deps. The runtime code enumerates displays, captures screenshots, lists installed/running apps, reads process executable paths (psutil), and can synthesize input via pyautogui — all within the claimed scope. There is no instruction to read unrelated user files or call external endpoints in the SKILL.md, but the runtime will access local process and display state (sensitive).
Install Mechanism
No formal install spec is declared (instruction-only), but SKILL.md requires running npm install and the first-run bootstraps a Python virtualenv and pip-installs requirements from PyPI. That implies network downloads and writing files into the user home (~/.codex/skills/... and .runtime/venv). Using npm/pip is expected for this project, but it is higher risk than an instruction-only, no-download skill because it pulls code from registries at install/run time.
Credentials
The skill does not request credentials or environment secrets. It references optional env flags (CLAUDE_COMPUTER_USE_DEBUG, CODEX_THREAD_ID etc.) which are reasonable for runtime configuration. No disproportionate or unrelated credentials are required.
Persistence & Privilege
The skill is not always-enabled and follows the default model-invocation policy (agent may call the skill autonomously). It writes files under the user's home (config/lock files and the bundled project path) and creates a Python venv on first run — expected for this runtime but means it gains persistent presence and the ability to perform desktop actions on the host. Autonomous invocation combined with desktop-control capabilities increases blast radius; consider whether you want automatic agent-initiated GUI actions.
如何使用
- 确保已安装 OpenClaw（本地或 Docker 部署）
- 在对话框中输入安装命令：
/install computer-use-linux - 安装完成后，直接呼叫该 Skill 的名称或使用
/computer-use-linux触发 - 根据 Skill 的参数说明提供必要输入，即可获得结构化输出
版本历史
v0.1.1
Fix Linux shared platform typing and system-key blocklist logic, sync bundled skill payload, and upgrade multilingual README plus validation notes.
元数据
常见问题
Computer Use Linux 是什么？
Top-level Linux computer-use skill with a bundled standalone runtime that bootstraps itself without any local Claude installation, private native modules, or... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 172 次。
如何安装 Computer Use Linux？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install computer-use-linux」即可一键安装，无需额外配置。
Computer Use Linux 是免费的吗？
是的，Computer Use Linux 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。
Computer Use Linux 支持哪些平台？
Computer Use Linux 跨平台运行，可在任意部署了 OpenClaw / Claude Code 的环境中使用（cross-platform）。
谁开发了 Computer Use Linux？
由 wimi321（@wimi321）开发并维护，当前版本 v0.1.1。
推荐 Skills