Tracks · Learn Claude

What is an MCP server, and how do you build one?

An MCP server is how Claude reaches outside the conversation — it exposes tools, resources and prompts over a standard protocol, so one integration works in any MCP client. This track covers tool schemas the model picks correctly, when a capability should be a resource instead, stdio versus Streamable HTTP, the initialization handshake, auth, and error shapes that make retries safe. Assumes Foundations.

What you learn in this track

Tool schema design

Narrow, orthogonal tools beat one god-tool with a mode flag, because the model picks tools by their description.

Side-effect hints (read vs. destructive)

readOnlyHint and destructiveHint tell the client what it may do without asking you first.

MCP initialization handshake

The MCP handshake is where client and server agree on what exists — most integration bugs are visible right here.

Transport: stdio vs. Streamable HTTP

Pick stdio for local processes and Streamable HTTP for remote servers; SSE is legacy and still supported.

Resources vs. tools vs. prompts

If a capability only reads, expose it as a resource. Reaching for a tool first is the most common early MCP mistake.

Versioning & auth models

Default to OAuth 2.1 the moment a tool touches per-user data.

Idempotency & error shapes

Surface provider errors verbatim; collapsing everything into a 500 makes retries unsafe.

Guides in this track

• Which MCP transport — stdio, SSE or Streamable HTTP?
• What is MCP, and what is an MCP server?

How the Socratic coach teaches this

Each module is a Socratic dialogue: the coach probes your reasoning with scenario questions, surfaces weak points, and generates flashcards and study notes from your own gaps. Mastery is tracked per concept, so you can see which ideas you actually handle and which ones still need work.

See where this track sits in the curriculum on the six tracks and why they are in this order, or open the ClaudeMaster.ai hub to start it.