Learn it. Build it. Ship it for others.
MCP security checklists and guides for agents
This repository collects practical security guidance for MCP usage, from local server deployment to OAuth, logging, and threat review. The guides link out to checklists and indexes that help you assess risks and choose safer MCP setups.
Builders who use MCP servers with AI agents and want security checklists before they connect them to real systems.
You can review MCP setups with concrete checklists instead of relying on ad hoc security judgment.
What it does
MCP deployment guidance
Covers running local MCP servers in Docker, exposing localhost MCPs to the internet, and running STDIO MCPs on remote servers.
Authentication and identity checklists
Includes troubleshooting OAuth in MCP and an MCP identity management checklist.
Logging and observability guidance
Provides a checklist for logging, auditing, and observability around MCP usage.
Threat screening resources
Includes a cybersecurity threat list, a reported vulnerabilities index, and a checklist for evaluating MCP servers for threats and risks.
Agent-building guidance
Adds checklists for building AI agents and improving MCP tool selection.
README
MCP-Checklists
We're a team of security and AI enthusiasts building MCP Manager, a comprehensive MCP security solution for businesses of all sizes.
In this repository we will publish a range of checklists, indexes, lessons learned and helpful utilities to help you adopt and use AI agents and MCP servers securely - without losing pace in the AI race.
⭐Star this repo to stay up to date and avoid missing that guide you know you'll need in the future!
We welcome contributions and suggestions - here's the instructions for contributing.
📚 Table of Contents
- MCP Deployment Infrastructure
- Authentication and Authorization
- Logging, Auditing, and Observability
- Security Threats and Mitigation
- AI Agent Building, Optimization, & Security
🔌 MCP Deployment Infrastructure
- Using Docker to Run Local MCP Servers Securely
- How To Expose LocalHost MCPs To The Internet
- How To Run STDIO MCPs On Remote Servers
🐳 Using Docker to Run Local MCP Servers Securely
Installing and running MCP servers locally is equivalent to installing and running any other software on your computer. Locally running MCP servers have unlimited access to all your files, creating risks of data exfiltration, token theft, virus infection and propagation, or data encryption attacks (Ransomware).
Docker is a containerization solution that is free, open source, and widely supported across all major operating systems.
Why You Should Use Docker to Containerize Local MCP Servers
Running MCP servers inside Docker containers allows you to run them in a sandboxed environment that you have complete control over. You decide which files and folders to expose to the container, can define rules for HTTP and WebSocket traffic, and selectively expose environment variables instead of unintentionally leaking secrets.
Using Docker to containerize your MCP servers reduces security risks and gives you more control over what data and capabilities the server has access to. It's not a complete bulletproof solution, however, because if you're not careful, you can still give Docker containers running locally unfettered access to your VPN / private networks.
Guide and Docker Files
We understand that learning complex technologies like Docker can be intimidating, but we've made our best effort to provide you with examples, documentation, and helpful scripts to get you started running MCPs securely.
Use How to Run Local MCP Servers Securely to learn about our helpful scripts and Dockerfiles that will get you started running local MCP Servers securely in no time.
🔐 Authentication and Authorization
Checklists
📝 Logging, Auditing, and Observability
Checklists
🛡️ Threats and Mitigation
Checklists
Index Lists
🤖 AI Agent Building and Security
Checklists/Guides
Files in the repo
- .vscode
- infrastructure
- .env.example
- .gitattributes
- .gitignore
- CLAUDE.md
- CODE_OF_CONDUCT.md
- CONTRIBUTING.md
- Dockerfile
- LICENSE
- mcp-security-threat-list.md
- package.json
- pnpm-lock.yaml
- pnpm-workspace.yaml
- README.md
Discussion (0)
Ask about usage, or say what you built with itSign in to join the discussion.
No comments yet. Be the first to say what this is good for.
More tutorials & guides
Bash is all you need - A nano claude code–like 「agent harness」, built from 0 to 1

End-to-end, code-first tutorials for building production-grade GenAI agents. From prototype to enterprise deployment.
This open-source curriculum introduces the fundamentals of Model Context Protocol (MCP) through real-world, cross-language examples in .NET, Java, TypeScript, JavaScript, Rust and Python. Designed for developers, it focuses on practical techniques for building modular, scalable, and secure AI workflows from session setup to service orchestration.

50+ tutorials and implementations for Generative AI Agent techniques, from basic conversational bots to complex multi-agent systems.

The most comprehensive Claude Code guide: agentic workflows, hooks, skills, MCP servers, quizzes, and production-ready templates. 430K+ lines.