What are the infrastructure requirements?

The server is built to run on Bun or Node.js and requires a PostgreSQL database to manage users, OAuth clients, and session tokens.

Can I customize the tools provided by the server?

Yes, the server uses the standard MCP TypeScript SDK, allowing you to register any custom tools or resources within the generated src/mcp/server.ts file.

Why does this skill use OAuth 2.1 with PKCE?

OAuth 2.1 with PKCE is the modern security standard for authorization, ensuring that only authenticated users and registered clients can execute your MCP tools.

What is the benefit of a remote MCP server?

A remote MCP server allows AI agents to access your custom tools from anywhere over HTTPS, removing the restriction of local stdio connections and enabling cloud-based workflows.

Which agents can connect to this server?

Any client supporting the MCP Streamable HTTP transport can connect, including Claude Code, Claude Desktop, and various IDE extensions.

Remote MCP Server Deployer

Name: Remote MCP Server Deployer
Author: mattwoodco

bymattwoodco

•

API 개발

Deploys a production-ready Model Context Protocol (MCP) server featuring secure OAuth 2.1 authentication and PKCE support.

This skill automates the creation of a standalone, remote MCP server, enabling agents like Claude Code or VS Code to access custom tools via secure Streamable HTTP. It implements a complete OAuth 2.1 flow including dynamic client registration and PKCE, providing a robust security layer for remote tool execution. Built on a modern stack of Bun, Hono, and Drizzle ORM with a PostgreSQL backend, it handles everything from database schema setup for token management to the .well-known metadata endpoints required by the MCP specification, making it ideal for developers who need to expose internal tools to AI agents safely over the internet.

주요 기능

01Streamable HTTP transport for remote agent connectivity via URL

02Type-safe environment configuration and validation with Zod and Bun

033 GitHub stars

04Automated database schema generation using Drizzle ORM and PostgreSQL

05Built-in .well-known discovery endpoints (RFC 8414 and RFC 9728)

06Full OAuth 2.1 implementation with PKCE and dynamic client registration

사용 사례

01Exposing internal company tools to Claude Code securely over the public internet

02Building a multi-tenant MCP service that requires user authentication and consent

03Creating a centralized hub for custom AI tools accessible from any MCP-compatible environment

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add mattwoodco/skills mcp-server

For use in Claude.ai and ChatGPT

Download Skill