Complete guided curriculum

AI Agents & Model Context Protocol

A 12-module, 152-lesson path from agent fundamentals to building, securing, testing and operating MCP-powered systems in production.

Module 1 · 10 lessons

AI Agent Fundamentals

Agents, workflows, loops, tools, function calling and first-agent architecture.

Open module →
Module 2 · 10 lessons

MCP Fundamentals

Why MCP exists, host-client-server architecture, JSON-RPC, lifecycle and compatibility.

Open module →
Module 3 · 15 lessons

MCP Core Capabilities

Tools, resources, prompts, schemas, elicitation, completion, pagination and caching.

Open module →
Module 4 · 15 lessons

Build an MCP Server

Python SDK setup, tools, resources, prompts, FastAPI, PostgreSQL and testing.

Open module →
Module 5 · 11 lessons

Transports & Clients

stdio, Streamable HTTP, local and remote clients, discovery, caching and multiple servers.

Open module →
Module 6 · 15 lessons

Agents with MCP

OpenAI Agents SDK, hosted/local/remote MCP, approvals, filtering and practical agents.

Open module →
Module 7 · 15 lessons

Agent Patterns

Memory, context, handoffs, planner-executor, evaluator-optimizer and multi-agent orchestration.

Open module →
Module 8 · 14 lessons

MCP Security

OAuth, authorization, least privilege, prompt injection, SSRF, destructive tools and tenancy.

Open module →
Module 9 · 11 lessons

Testing & Observability

MCP Inspector, automated tests, tracing, failure monitoring and regression evaluation.

Open module →
Module 10 · 15 lessons

Production MCP

Deployment, Docker, scaling, retries, timeouts, rate limits, caching and cost optimization.

Open module →
Module 11 · 13 lessons

Advanced MCP

Apps, extensions, registry, tasks, managed auth, machine-to-machine auth and versioning.

Open module →
Module 12 · 8 lessons

Framework Integrations

OpenAI Agents SDK, Claude, VS Code, Cursor, LangChain/LangGraph and Semantic Kernel.

Open module →