← 返回 Skills 市场
Docker Sandbox
作者
gitgoodordietrying
· GitHub ↗
· v1.0.0
6292
总下载
6
收藏
0
当前安装
1
版本数
在 OpenClaw 中安装
/install docker-sandbox
功能描述
Create and manage Docker sandboxed VM environments for safe agent execution. Use when running untrusted code, exploring packages, or isolating agent workloads. Supports Claude, Codex, Copilot, Gemini, and Kiro agents with network proxy controls.
安全使用建议
Review this skill before installing if you plan to use it for hostile or truly untrusted code. Confirm whether /run/docker.sock is isolated inside the sandbox VM or connected to the host Docker daemon, assume mounted workspaces can be modified by code in the sandbox, use deny-by-default network rules, and avoid docker sandbox reset unless you are prepared to lose sandbox state.
功能分析
Type: OpenClaw Skill
Name: docker-sandbox
Version: 1.0.0
The OpenClaw AgentSkills skill bundle for 'docker-sandbox' is classified as benign. Its stated purpose is to create and manage isolated Docker sandboxed VM environments for safely executing untrusted code or agent workloads. All capabilities described, such as running arbitrary commands, mounting workspaces, and network controls, are explicitly designed to operate within this isolated environment. The documentation emphasizes security features like network policy controls to mitigate risks from code running inside the sandbox, rather than introducing risks to the host system. There is no evidence of prompt injection against the agent, data exfiltration, malicious execution on the host, or obfuscation within the provided files.
能力评估
Credentials
Docker, network proxy configuration, agent execution, background execution, workspace mounts, and Docker-in-Docker capability fit the sandbox-management purpose, but they are powerful enough that the skill should provide stronger scoping and warnings before recommending hostile or destructive workloads.
Install Mechanism
The artifact is a single Markdown skill file with no bundled executable scripts, install hooks, hidden payloads, or dependency installation performed by the skill itself.
Instruction Scope
The commands are user-directed and mostly transparent, but the guidance repeatedly frames the environment as safe for untrusted code while omitting prominent cautions about Docker socket authority, read-write workspace effects, and destructive reset behavior.
Persistence & Privilege
The skill documents persistent sandboxes and reusable templates, plus a reset command that clears all sandbox state; this is disclosed but could cause unexpected data loss without a stronger warning.
Purpose & Capability
The stated purpose is isolated execution for untrusted packages, arbitrary code, and destructive tests, but the instructions also describe a mounted host workspace and an available Docker socket, which are high-impact capabilities that are not clearly bounded in the safety claims.
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 docker-sandbox - 安装完成后，直接呼叫该 Skill 的名称或使用
/docker-sandbox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
v1.0.0
Initial release - Docker Desktop sandbox VM management for safe agent execution with network proxy controls, workspace mounting, and troubleshooting guides.
元数据
常见问题
Docker Sandbox 是什么？
Create and manage Docker sandboxed VM environments for safe agent execution. Use when running untrusted code, exploring packages, or isolating agent workloads. Supports Claude, Codex, Copilot, Gemini, and Kiro agents with network proxy controls. 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 6292 次。
如何安装 Docker Sandbox？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install docker-sandbox」即可一键安装，无需额外配置。
Docker Sandbox 是免费的吗？
是的，Docker Sandbox 完全免费（开源免费），可自由下载、安装和使用。
Docker Sandbox 支持哪些平台？
Docker Sandbox 跨平台运行，可在任意部署了 OpenClaw / Claude Code 的环境中使用（linux, darwin, win32）。
谁开发了 Docker Sandbox？
由 gitgoodordietrying（@gitgoodordietrying）开发并维护，当前版本 v1.0.0。
推荐 Skills