9.5.1 MCP Roadmap: Server, Client, Capability

MCP is a protocol layer for connecting tools, resources, and prompt templates to model applications in a more standard way. It does not replace Agents or tools; it makes capabilities easier to expose and use consistently.

See the MCP Boundary First

Function Calling focuses on structured calls. MCP focuses on how external capabilities are discovered, described, called, and governed through a protocol.

Run a Capability Registry Check

Before implementing a real MCP Server, list what it exposes and what the Client may call.

server = {
    "tools": ["search_docs"],
    "resources": ["course://ch09-agent"],
    "prompts": ["study_plan"],
}

client_request = "search_docs"

print("server_ready:", all(server.values()))
print("can_call:", client_request in server["tools"])
print("boundary:", "server exposes, client calls")

Expected output:

server_ready: True
can_call: True
boundary: server exposes, client calls

If the boundary is vague, permissions and debugging will be vague too.

Learn in This Order

Step	Read	Practice Output
1	MCP concept	Explain why a protocol layer reduces integration mess
2	MCP architecture	Distinguish Host, Client, Server, tools, resources, prompts
3	Server development	Wrap one capability with clear input, output, and errors
4	Client integration	Discover and call server capabilities safely
5	Ecosystem	Connect MCP to IDEs, databases, browsers, knowledge bases, and Agents

Evidence to Keep

Keep this page’s proof of learning as a small evidence card:

Capability

resource, prompt, or tool exposed by server

Contract

schema, transport, permissions, and error shape

Call Trace

discovery, invocation, response, and failure handling

Failure Check

incompatible schema, missing auth, unsafe tool, or server error

Integration Action

validate server contract before adding autonomy

Pass Check

You pass this chapter when you can draw the Host-Client-Server relationship and explain what the Server exposes, what the Client calls, and where permissions are checked.

The exit mini project is a course-materials MCP Server design: one search tool, one resource URI pattern, one prompt template, and one failure-handling rule.

Check reasoning and explanation

1. A passing answer describes the agent loop: goal, plan, tool call, observation, memory or state update, and stop condition.
2. The evidence should include a trace that another developer can inspect, not only the final answer.
3. A good self-check names one safety or reliability control such as tool schemas, permission boundaries, retries, evaluation cases, or a human-review point.