Unboxing AI: Build a Remote MCP Server from Zero to Deployed with OAuth
2h 56mGeneral2025-07-31
Authors
Morten Rand-Hendriksen
Senior Staff Instructor, Speaker, Web Designer, and Software Developer
Course details
MCP, the Model Context Protocol, enables you to connect pretty much anything to AI. In this livestream series-turned-course, LinkedIn Principal Instructor Morten Rand-Hendriksen unboxes MCP. You’ll learn what MCP is and why it matters. Then get hands on and learn how to build MCP servers from scratch, use cutting edge features like elicitation and sampling, manage security and authorization, and deploy remote MCP servers for easier access and use.
Learning objectives
Understand what MCP is and why it matters.
Create your own MCP servers from scratch.
Use cutting edge features like elicitation and sampling.
Explore MCP security challenges and authorization.
Deploy remote MCP servers for easier access and use.
Learning objectives
Understand what MCP is and why it matters.
Create your own MCP servers from scratch.
Use cutting edge features like elicitation and sampling.
Explore MCP security challenges and authorization.
Deploy remote MCP servers for easier access and use.
Skills covered
OAuthProgramming FoundationsGenerative AIArtificial Intelligence FoundationsArtificial Intelligence (AI)Software DevelopmentOne-Off
Concepts
0. Introduction
- 01 - Why MCP changes everything
1. MCP - The Model Context Protocol
- 02 - Understanding what MCP means
- 03 - The problem MCP solves
- 04 - MCP's revolutionary approach
- 05 - Real-world examples of MCP in Claude
- 06 - Safety concerns and safe usage of MCP
- 07 - Why MCP makes automation easier
2. Building an MCP Server
- 08 - How to build your own MCP server
- 09 - Setting up your MCP development environment
- 10 - SDKs and setup
- 11 - Installing and using UV
- 12 - Running an MCP server locally
- 13 - Using the MCP Inspector
- 14 - Testing an MCP server in VS Code
- 15 - Using Resources and Prompts
- 16 - Building an MCP server from a template
- 17 - Writing tools, resources, and prompts
- 18 - Testing your MCP server
- 19 - Using an MCP server in Claude Desktop
- 20 - Customizing your MCP server to fit your needs
- 21 - Extending your MCP server with GitHub Copilot
3. MCP Elicitation and Sampling
- 22 - Elicitation and sampling - The bleeding edge of MCP
- 23 - Understanding MCP elicitation
- 24 - Why and when elicitation is useful
- 25 - Demo - MCP elicitation in VS Code
- 26 - Why elecitiation matters
- 27 - MCP elicitation in code
- 28 - Testing a new MCP elicitation
- 29 - MCP sampling - The server uses AI tools
- 30 - Demo - MCP sampling in VS Code
- 31 - Innovating with elicitation and sampling
4. MCP Authorization and Security
- 32 - Why authorization and security matter
- 33 - The MCP authorization spec
- 34 - Implementing MCP authorization
- 35 - The MCP OAuth 2.0 flow visualized
- 36 - Best-practices for MCP authorization
- 37 - Options for MCP authorization
- 38 - MCP and security
- 39 - MCP security risk - Vibe coding
- 40 - MCP security risk - Excessive permisisons
- 41 - MCP security risk - Agent error
- 42 - MCP security risk - Prompt injection
- 43 - MCP security risk - Confused deputy
- 44 - MCP security risk - Session hijacking
- 45 - Authorization and security are job #1
5. Deploying Remote MCP Servers with OAuth
- 46 - Why remote MCP servers are so useful
- 47 - Deploying a rempote MCP server on Cloudflare Workers
- 48 - Testing a remote MCP server
- 49 - Adding tools and updating a remote MCP server
- 50 - Implementing GitHub OAuth 2.0 in a remote MCP server
- 51 - Testing MCP OAuth in local development
- 52 - Setting up MCP OAuth on Cloudflare
- 53 - Testing a remote MCP server with OAuth
Conclusion
- 54 - Build MCP servers and pave the cowpaths of agentic AI