container-mcp

container-mcp

MCP server for Apple container to manage, build, inspect, and diagnose Linux containers on macOS directly from AI agents.

Category
访问服务器

README

container-mcp

npm version License: Apache-2.0 MCP Compatible

MCP server for Apple container — manage, build, inspect, and diagnose Linux containers on macOS directly from AI agents like Claude Desktop, Claude Code, and Cursor.

Built on Apple's container CLI for native Linux container support on Apple Silicon Macs.


Prerequisites

Installation

npm install -g container-mcp

You can run the server directly via npx using the published npm package:

npx -y @arxndev/container-mcp

Or, if you prefer to build from source:

Configuration

For Claude Desktop

Add this to your claude_desktop_config.json file:

{
  "mcpServers": {
    "container-mcp": {
      "command": "npx",
      "args": [
        "-y",
        "@arxndev/container-mcp"
      ]
    }
  }
}

Claude Code

claude mcp add container-mcp -- npx -y container-mcp

Cursor

Add to .cursor/mcp.json in your project:

{
  "mcpServers": {
    "container": {
      "command": "npx",
      "args": ["-y", "container-mcp"]
    }
  }
}

Environment Variables

Variable Default Description
CONTAINER_CLI_PATH container Path to the container CLI binary
CONTAINER_MCP_LOG_LEVEL info Log level: debug, info, warn, error
CONTAINER_MCP_VOLUME_ROOT $HOME Restricts host paths for volume mounts and file copies. Paths outside this root are rejected.
CONTAINER_MCP_CONTEXT_ROOT $HOME Restricts the build context path for build_image. Paths outside this root are rejected.

Tool Reference

🐳 Container Lifecycle (14 tools)

Tool Description Key Parameters
list_containers List all containers all, format
run_container Create & run a new container image, name, ports, env, volumes, detach, command
stop_container Stop running containers names, timeout
start_container Start stopped containers names
delete_container Remove containers names, force
inspect_container Get detailed container info name
exec_in_container Execute command in container name, command, interactive
container_commit Commit container to new image container, image, message
copy_to_container Copy file/dir from host to container hostPath, containerName, containerPath
copy_from_container Copy file/dir from container to host containerName, containerPath, hostPath
wait_container Wait for container stop, get exit code name, timeout
rename_container Rename an existing container container, newName
pause_container Pause processes in containers names
unpause_container Unpause processes in containers names

📦 Image Management (7 tools)

Tool Description Key Parameters
list_images List local images format
pull_image Pull image from registry reference
build_image Build image from Dockerfile context, tag, dockerfile, buildArgs, platform
remove_image Remove images references, force
inspect_image Get detailed image info reference
tag_image Tag an existing local image source, target
push_image Push image to registry image

📋 Logs & Monitoring (3 tools)

Tool Description Key Parameters
get_logs Get container logs name, tail, since
get_container_stats Get CPU/memory/network stats name
get_container_processes List processes in container name

💾 Volumes (3 tools)

Tool Description Key Parameters
list_volumes List all volumes format
create_volume Create a new volume name, size
delete_volume Remove volumes names, force

🌐 Networks (3 tools)

Tool Description Key Parameters
create_network Create a container network name, driver, options
delete_network Remove one or more networks networks
list_networks List all networks format

⚙️ System (2 tools)

Tool Description Key Parameters
system_info Get system info & version —
system_prune Remove all unused resources confirm (must be true)

🤖 AI-Powered Diagnostics (2 tools)

Tool Description Key Parameters
diagnose_container Analyze logs & inspect data for issues name
explain_container Human-readable container summary name

Example Prompts

Here are natural language prompts you can use with Claude or any MCP-compatible AI:

"List all my running containers"

"Run an nginx container named web-server on port 8080"

"Why is my database container crashing?"

"Show me the logs from the api container"

"Pull the latest Ubuntu image and run a container with it"

"What's the memory usage of all running containers?"

"Explain what the postgres container is doing"

"Build a Docker image from the current directory and tag it myapp:latest"

"Stop all containers and clean up unused images"

"Diagnose why the redis container keeps restarting"

Development

Setup

git clone https://github.com/your-org/container-mcp.git
cd container-mcp
npm install

Build

npm run build

Test

npm test
npm run test:coverage

Run locally

npm run dev

Contributing

  1. Fork the repository
  2. Create a feature branch: git checkout -b feature/my-feature
  3. Write code and tests
  4. Ensure tests pass: npm test
  5. Ensure the build works: npm run build
  6. Submit a Pull Request

Code Style

  • TypeScript strict mode
  • ESM modules
  • Prettier + ESLint for formatting
  • Zod schemas for all input validation
  • JSDoc comments on all exported functions

Architecture

container-mcp
├── MCP Server (stdio transport)
│   ├── Container Tools (8) ── container CLI
│   ├── Image Tools (5) ────── container CLI
│   ├── Log Tools (3) ──────── container CLI
│   ├── Volume Tools (3) ───── container CLI
│   ├── Network Tools (1) ──── container CLI
│   ├── System Tools (2) ───── container CLI
│   └── AI Tools (2) ──────── combines multiple CLI calls
│       ├── diagnose_container → logs + inspect → pattern analysis
│       └── explain_container  → inspect + stats + top → summary
└── Utilities
    ├── CLI Executor (execa, safe args)
    ├── Output Parsers (JSON, table)
    └── Logger (structured, stderr)

License

Apache-2.0

推荐服务器

Baidu Map

Baidu Map

百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。

官方
精选
JavaScript
Playwright MCP Server

Playwright MCP Server

一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。

官方
精选
TypeScript
Magic Component Platform (MCP)

Magic Component Platform (MCP)

一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。

官方
精选
本地
TypeScript
Audiense Insights MCP Server

Audiense Insights MCP Server

通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。

官方
精选
本地
TypeScript
VeyraX

VeyraX

一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。

官方
精选
本地
graphlit-mcp-server

graphlit-mcp-server

模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。

官方
精选
TypeScript
Kagi MCP Server

Kagi MCP Server

一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。

官方
精选
Python
e2b-mcp-server

e2b-mcp-server

使用 MCP 通过 e2b 运行代码。

官方
精选
Neon MCP Server

Neon MCP Server

用于与 Neon 管理 API 和数据库交互的 MCP 服务器

官方
精选
Exa MCP Server

Exa MCP Server

模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。

官方
精选