Lesson 9 of 20

MCP & Agent Skills

Model Context Protocol

4 min read

The Model Context Protocol (MCP) has emerged as the universal standard for connecting AI models to external tools and data sources. Introduced by Anthropic in late 2024, it's now adopted by ChatGPT, Gemini, Copilot, and most major AI platforms.

Why MCP Matters

Before MCP, every platform had its own tool format:

# OpenAI format
{"type": "function", "function": {"name": "search", "parameters": {...}}}

# Anthropic format
{"name": "search", "input_schema": {...}}

# Custom formats everywhere else...

MCP provides one standard for tools, resources, and prompts that works everywhere.

The MCP Ecosystem

As of December 2025:

  • Thousands of public MCP servers
  • Major adopters: Claude, ChatGPT, Gemini, Copilot, Cursor
  • Categories: Databases, APIs, file systems, browsers, dev tools

Core Concepts

MCP defines three primitives:

1. Tools

Functions the model can call:

{
  "name": "search_web",
  "description": "Search the web for information",
  "inputSchema": {
    "type": "object",
    "properties": {
      "query": {"type": "string", "description": "Search query"}
    },
    "required": ["query"]
  }
}

2. Resources

Data the model can read:

{
  "uri": "file:///workspace/README.md",
  "name": "Project README",
  "mimeType": "text/markdown"
}

3. Prompts

Reusable prompt templates:

{
  "name": "code_review",
  "description": "Review code for best practices",
  "arguments": [
    {"name": "language", "description": "Programming language"}
  ]
}

How It Works

┌─────────────┐     stdio/HTTP     ┌─────────────┐
│   AI Host   │◄──────────────────►│ MCP Server  │
│ (Claude,    │                    │ (Your tool) │
│  ChatGPT)   │   JSON-RPC 2.0     │             │
└─────────────┘                    └─────────────┘

The protocol uses JSON-RPC 2.0 over stdio or HTTP:

# Request from host to server
{
  "jsonrpc": "2.0",
  "method": "tools/call",
  "params": {
    "name": "search_web",
    "arguments": {"query": "MCP protocol"}
  },
  "id": 1
}

# Response from server
{
  "jsonrpc": "2.0",
  "result": {
    "content": [{"type": "text", "text": "Search results..."}]
  },
  "id": 1
}
ServerPurposeUse Case
filesystemFile operationsRead/write local files
postgresDatabase accessQuery databases
githubGit operationsPRs, issues, code
brave-searchWeb searchFind information
puppeteerBrowser controlWeb automation
slackTeam communicationSend messages

Finding MCP Servers

# Official MCP registry
https://github.com/modelcontextprotocol/servers

# Community servers
https://mcp.so/servers

# NPM packages
npm search mcp-server

Nerd Note: MCP is to AI tools what USB is to hardware—one standard that just works. Build your tools once, use them everywhere.

Next: Building your own MCP server. :::

Quick check: how does this lesson land for you?

Quiz

Module 3: MCP & Agent Skills

Take Quiz
FREE WEEKLY NEWSLETTER

Stay on the Nerd Track

One email per week — courses, deep dives, tools, and AI experiments.

No spam. Unsubscribe anytime.