Choosing MCP's Shape for the Job
Four questions, who initiates the call, how sensitive the data is, how long the work runs, and whether a human needs to see it, turn a vague request into a specific choice of primitive, transport, auth path, and extension. Map seven operational use case families, developer tools, data access, enterprise systems of record, long running workflow automation, interactive UIs, reusable workflows, and machine to machine integration, to the MCP primitive, transport, and extension that fit each one. Choose between a tool, a resource, and a prompt by asking who controls the action: the model, the application, or the user. Recognize when MCP is the wrong tool for a job and name what a team should reach for instead. Reason about the four operational concerns, authorization path, cache scope, consent, and observability, that any use case design has to answer. Read a server/discover result and a tools/call exchange and connect their cache hints and extension declarations back to the use case that produced them. A team that has learned MCP's message shapes still hits a harder question the moment a real project lands on their desk: given this job, which of the protocol's shapes actually fits? The wire rules hold equally for a five second lookup and a twenty minute pipeline, so nothing in the JSON-RPC envelope forces the choice by…
Choosing MCP's Shape for the Job: Four questions, who initiates the call, how sensitive the data is, how long the work runs, and whether a human needs to see…
This free lesson is part of the AI Engineering from Scratch curriculum. Read the full explanation, run the lesson code, and verify the result in the interactive reader or from the repository source.
Browse the complete course catalog or open this lesson on GitHub.