containerize-mcp-server
关于
This skill helps developers containerize R-based MCP servers using Docker, enabling portable deployment without local R installations. It covers mcptools integration, transport configuration (stdio/HTTP), and connecting Claude Code to the container. Use it for reproducible environments, distributing servers to other developers, or running MCP services alongside other containerized applications.
快速安装
Claude Code
推荐npx skills add pjt222/agent-almanac -a claude-code/plugin add https://github.com/pjt222/agent-almanacgit clone https://github.com/pjt222/agent-almanac.git ~/.claude/skills/containerize-mcp-server在 Claude Code 中复制并粘贴此命令以安装该技能
技能文档
Containerize MCP Server
Package R MCP server into Docker container → portable deployment.
Use When
- Deploying R MCP server w/o local R install req'd
- Creating reproducible MCP server env
- Running MCP servers alongside other containerized services
- Distributing MCP server to other devs
In
- Required: R MCP server impl (mcptools-based or custom)
- Required: Docker installed + running
- Optional: Additional R pkgs server needs
- Optional: Transport mode (stdio or HTTP)
Do
Step 1: Create Dockerfile for MCP Server
FROM rocker/r-ver:4.5.0
# Install system dependencies
RUN apt-get update && apt-get install -y \
libcurl4-openssl-dev \
libssl-dev \
libxml2-dev \
libgit2-dev \
libssh2-1-dev \
git \
curl \
&& rm -rf /var/lib/apt/lists/*
# Install R packages
RUN R -e "install.packages(c( \
'remotes', \
'ellmer' \
), repos='https://cloud.r-project.org/')"
# Install mcptools
RUN R -e "remotes::install_github('posit-dev/mcptools')"
# Set working directory
WORKDIR /workspace
# Expose MCP server ports
EXPOSE 3000 3001 3002
# Environment variables
ENV R_LIBS_USER=/workspace/renv/library
ENV RENV_PATHS_CACHE=/workspace/renv/cache
# Default: start MCP server
CMD ["R", "-e", "mcptools::mcp_server()"]
→ Dockerfile exists in project root w/ rocker/r-ver base image, sys deps, mcptools install, MCP server as default cmd.
If err: Valid. base image tag matches R ver. remotes::install_github fails → check git + libgit2-dev in sys deps layer.
Step 2: Create docker-compose.yml
version: '3.8'
services:
mcp-server:
build:
context: .
dockerfile: Dockerfile
container_name: r-mcp-server
image: r-mcp-server:latest
volumes:
- /path/to/projects:/workspace
- renv-cache:/workspace/renv/cache
stdin_open: true
tty: true
network_mode: "host"
environment:
- TERM=xterm-256color
- R_LIBS_USER=/workspace/renv/library
restart: unless-stopped
volumes:
renv-cache:
driver: local
Using network_mode: "host" ensures MCP server ports accessible on localhost.
→ docker-compose.yml in project root w/ MCP server service, vol mounts for project files + renv cache, stdin_open/tty enabled for stdio transport.
If err: Vol paths invalid → adjust /path/to/projects to actual project dir. Windows/WSL → use /mnt/c/... or /mnt/d/... paths.
Step 3: Build + Start
docker compose build
docker compose up -d
→ Container starts w/ MCP server running.
If err: Check logs w/ docker compose logs mcp-server. Common issues:
- Missing R pkgs: Add to Dockerfile RUN install step
- Port already in use: Change exposed port or stop conflicting service
Step 4: Connect Claude Code to Container
Stdio transport (container must stay running w/ stdin):
claude mcp add r-mcp-docker stdio "docker" "exec" "-i" "r-mcp-server" "R" "-e" "mcptools::mcp_server()"
HTTP transport (if MCP server supports):
{
"mcpServers": {
"r-mcp-docker": {
"type": "http",
"url": "http://localhost:3000/mcp"
}
}
}
→ Claude Code's MCP config includes r-mcp-docker server entry, claude mcp list shows new server.
If err: Stdio transport → ensure container name matches (r-mcp-server) + container running w/ docker ps. HTTP transport → valid. port exposed + reachable w/ curl http://localhost:3000/mcp.
Step 5: Verify Connection
# Check container is running
docker ps | grep mcp-server
# Test R session inside container
docker exec -it r-mcp-server R -e "sessionInfo()"
# Verify mcptools is available
docker exec -it r-mcp-server R -e "library(mcptools)"
→ docker ps shows r-mcp-server container running, sessionInfo() returns expected R ver, library(mcptools) loads w/o err.
If err: Container not running → check docker compose logs mcp-server for startup errs. mcptools fails to load → rebuild image to ensure pkg installed correct.
Step 6: Add Custom MCP Tools
Add project-specific MCP tools → mount R scripts:
volumes:
- ./mcp-tools:/mcp-tools
Load in CMD:
CMD ["R", "-e", "source('/mcp-tools/custom_tools.R'); mcptools::mcp_server()"]
→ Custom R scripts accessible inside container at /mcp-tools/, MCP server loads them on startup alongside default tools.
If err: Valid. vol mount path correct w/ docker exec -it r-mcp-server ls /mcp-tools/. Scripts fail to source → check missing pkg deps in custom tools.
Check
- Container builds w/o errs
- MCP server starts inside container
- Claude Code can connect to containerized server
- MCP tools respond correct to reqs
- Container restarts clean
- Vol mounts allow access to project files
Traps
- stdin/tty req's: MCP stdio transport requires
stdin_open: true+tty: true - Network isolation: Default Docker networking may prevent localhost access. Use
network_mode: "host"or expose specific ports. - Pkg vers: Pin mcptools to specific commit for reproducibility
- Large image size: mcptools + deps can be large. Consider multi-stage builds for prod.
- Windows Docker paths: Running Docker Desktop on Windows w/ WSL → path mapping differs
→
create-r-dockerfile- base Dockerfile patterns for Rsetup-docker-compose- compose config detailsconfigure-mcp-server- MCP server config w/o Dockertroubleshoot-mcp-connection- debugging MCP connectivity issues
GitHub 仓库
相关推荐技能
content-collections
元Content Collections 是一个 TypeScript 优先的构建工具,可将本地 Markdown/MDX 文件转换为类型安全的数据集合。它专为构建博客、文档站和内容密集型 Vite+React 应用而设计,提供基于 Zod 的自动模式验证。该工具涵盖从 Vite 插件配置、MDX 编译到生产环境部署的完整工作流。
polymarket
元这个Claude Skill为开发者提供完整的Polymarket预测市场开发支持,涵盖API调用、交易执行和市场数据分析。关键特性包括实时WebSocket数据流,可监控实时交易、订单和市场动态。开发者可用它构建预测市场应用、实施交易策略并集成实时市场预测功能。
creating-opencode-plugins
元该Skill帮助开发者创建OpenCode插件,用于接入命令、文件、LSP等25+种事件。它提供了插件结构、事件API规范和JavaScript/TypeScript实现模式,适合需要拦截操作、扩展功能或自定义事件处理的场景。开发者可通过它快速构建响应式模块来增强OpenCode AI助手的能力。
sglang
元SGLang是一个专为LLM设计的高性能推理框架,特别适用于需要结构化输出的场景。它通过RadixAttention前缀缓存技术,在处理JSON、正则表达式、工具调用等具有重复前缀的复杂工作流时,能实现极速生成。如果你正在构建智能体或多轮对话系统,并追求远超vLLM的推理性能,SGLang是理想选择。
