cybersecurity-mcp-server
A security-first MCP (Model Context Protocol) backend built with Node.js and Express, featuring hardening, rate limiting, structured logging, and Docker support.
README
cybersecurity-mcp-server 🔐
Security-first MCP (Model Context Protocol) backend built with Node.js and Express.
🚀 Overview
This repository contains a secure, production-oriented MCP backend focusing on best practices for hardening, monitoring and safe defaults. It uses Express 5, Helmet for headers, rate limiting middleware, and structured logging (winston + winston-daily-rotate-file).
⚙️ Features
- Express 5 backend
- Security middleware (Helmet, rate limiting)
- Structured logging with
winstonandwinston-daily-rotate-file - Docker-friendly with a
Dockerfile - Simple, modular architecture (controllers, services, middleware)
📁 Project Structure
A short map of core folders and important files so contributors know where to look.
src/
├─ server.js # App entrypoint (starts HTTP server)
├─ app.js # Express app, middleware, routes registration
├─ routes/ # API route definitions (e.g., mcp.routes.js)
├─ controllers/ # Request handlers (mcp.controller.js)
├─ services/ # Business logic & integrations (mcp.service.js)
├─ middleware/ # auth, rateLimit, and other cross-cutting concerns
├─ config/ # env.js, security.js and config helpers
└─ utils/ # logger.js and helper utilities
logs/ # Rotated log files written by winston
Dockerfile # Container image build instructions
package.json # Scripts & dependencies
README.md # This document
Notes:
src/config/env.jscentralizes environment variable loading and defaults.utils/logger.jsuseswinston+winston-daily-rotate-filefor structured, rotated logs.- Keep business logic in
services/and handlers incontrollers/to keep tests and mocking straightforward.
🧑💻 Local Development
Below are platform-specific steps so contributors on Linux/macOS and Windows can get started quickly.
Linux / macOS
- Clone and install:
git clone https://github.com/Krishsakaria26/cybersecurity-mcp-server.git
cd cybersecurity-mcp-server
npm install
- Create a
.envfile (optional but recommended)
# copy example if present
cp .env.example .env || true
# or create manually
cat > .env <<EOF
PORT=3000
NODE_ENV=development
RATE_LIMIT_WINDOW_MS=60000
RATE_LIMIT_MAX=100
LOG_LEVEL=debug
EOF
- Start in development mode (auto-reloads on change)
npm run dev
- Production run
npm start
Tips:
- Use your IDE debugger attached to
src/server.jsfor breakpoints. - Add tests (Jest/Mocha) and a
testnpm script before merging features. - Use
pre-commithooks for linting/formatting consistency.
Windows (PowerShell / CMD / WSL)
For the closest parity with Linux/macOS behavior, use WSL or Git Bash. PowerShell and CMD work, but some shell behaviors differ (path separators, quoting).
- Clone and install (PowerShell / CMD):
git clone https://github.com/Krishsakaria26/cybersecurity-mcp-server.git
cd cybersecurity-mcp-server
npm install
- Create a
.envfile (PowerShell / CMD)
- PowerShell:
New-Item -Path . -Name '.env' -ItemType 'file' -Force
Add-Content -Path .env -Value 'PORT=3000'
Add-Content -Path .env -Value 'NODE_ENV=development'
Add-Content -Path .env -Value 'RATE_LIMIT_WINDOW_MS=60000'
Add-Content -Path .env -Value 'RATE_LIMIT_MAX=100'
Add-Content -Path .env -Value 'LOG_LEVEL=debug'
- CMD:
echo PORT=3000> .env
echo NODE_ENV=development>> .env
echo RATE_LIMIT_WINDOW_MS=60000>> .env
echo RATE_LIMIT_MAX=100>> .env
echo LOG_LEVEL=debug>> .env
- Start in development mode
npm run dev
Notes for Windows:
- If using Windows Defender / Firewall you'll be prompted when the server listens on a port—allow local access.
- If you need POSIX tools or to match Linux behavior (e.g., volume mounts in Docker), prefer using WSL or Git Bash.
🐳 Docker Deployment
Docker commands differ slightly across shells; below are examples for both platforms and tips for Windows users.
Linux / macOS
Build the image:
docker build -t mcp-server:latest .
Run (basic):
docker run -p 3000:3000 --rm --name mcp-server mcp-server:latest
Run with env-file and persistent logs:
docker run -p 3000:3000 --env-file .env -v "$(pwd)/logs:/app/logs" --restart unless-stopped --name mcp-server mcp-server:latest
Windows (PowerShell / Docker Desktop)
Build the image (same):
docker build -t mcp-server:latest .
Run (basic):
docker run -p 3000:3000 --rm --name mcp-server mcp-server:latest
Run with env-file and persistent logs (PowerShell):
docker run -p 3000:3000 --env-file .\.env -v "${PWD}\logs:/app/logs" --restart unless-stopped --name mcp-server mcp-server:latest
If ${PWD} path doesn't mount correctly, provide the full Windows path for the volume mapping (replace C:\path\to\repo):
docker run -p 3000:3000 --env-file .\.env -v "C:\full\path\to\repo\logs:/app/logs" --restart unless-stopped --name mcp-server mcp-server:latest
Notes & recommendations:
- On Windows prefer Docker Desktop with WSL2 integration enabled for better compatibility.
- Use full absolute paths for volume mounts if relative mounts fail in PowerShell/CMD.
- When running containers in production, pass secrets via your orchestrator or Docker secrets (avoid committing
.env).
Optional: Docker Compose (works cross-platform if file paths are adjusted):
version: '3.8'
services:
mcp-server:
build: .
image: mcp-server:latest
ports:
- "3000:3000"
env_file: .env
volumes:
- ./logs:/app/logs
restart: unless-stopped
Health & deployment notes:
- Provide resource limits and environment secrets via your orchestrator (Kubernetes/Swarm).
- Mount
logs/for host access or forward logs to a log aggregator for production. - Consider adding a container healthcheck for orchestrators that rely on it.
🧩 Configuration
Environment settings are loaded from src/config/env.js. Common variables:
PORT- server port (default: 3000)NODE_ENV- environment (development/production)- Rate limit options in
src/config/security.js(window, max requests)
Tip: Create a
.envfile in the project root for local development.
Logging
The app uses winston with winston-daily-rotate-file for rotated logs stored in the logs/ folder. Logs include structured timestamps and levels for easier parsing and aggregation.
🧪 Tests
There are no automated tests included yet. Contributions adding tests (Jest/Mocha) are welcomed.
🤝 Contributing
Contributions are welcome — please open issues or pull requests on the GitHub repository. Follow existing code patterns and keep security considerations in mind.
📄 License
MIT — see the LICENSE file.
👤 Author
Krish Sakaria : Cybersecurity & Backend Engineering
Manish Shah : Backend Engineering
Madhusmita Choudhary : Fullstack Engineering
Aryan Karna : Frontend Developer
推荐服务器
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。
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。
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。
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。
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。
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。
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。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
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。