Page 8 of 8~112 min topic

MCP in code

Ship and explain the MCP stdio client

Page 8 packages proved vs unproved evidence so another engineer can run, trust, or reject the MCP stdio client calling one local tool.

~14 min this pageMastery and shipping

1Learn the idea

Read

Assemble the ship record

A shippable lab artifact includes: how to run it, the metric result (discovery count; successful call; timeout on hung server), the failure you can still reproduce (calling a tool not in the discovered list, or hanging on stderr noise), the security gate for auto-approving every discovered tool including filesystem writes, and a rollback note. The user decision it supports remains: connect to a local MCP server, list tools, and invoke one allowlisted tool safely.

Read

Freeze the evidence

npm run typecheck
npm test

Read

Run the credentialed smoke test only in an approved environment:

RUN_LIVE_EVALS=1 npm run test:live


Expected evidence: **The client discovers echo, calls it once, prints “hello”, and closes the transport.**. Store this beside the fixture version so scores remain meaningful after content changes in `mcp-in-code`.

Read

Explain limits without apology

State operating limits for the MCP stdio client in plain language: fixture size, offline vs live dependencies, and what would require a new eval set. Shipping mcp-in-code is honest scoping, not maximal confidence language.

Read

Lab notebook: proved vs unproved

Fill this table in your notes for the MCP stdio client:

  • Proved on local stdio server exposing one read-only echo tool: …
  • Unproved beyond the fixture: …
  • Metric that blocks release: discovery count; successful call; timeout on hung server
  • Failure still reproducible: calling a tool not in the discovered list, or hanging on stderr noise
  • Security gate: auto-approving every discovered tool including filesystem writes
  • Rollback: …

Ship the narrative only when the unproved list is honest. Reviewers trust narrow claims that support connect to a local MCP server, list tools, and invoke one allowlisted tool safely more than maximal language that collapses under the first production oddity.

Read

Worked judgment

Hand your ship note to a peer and ask them to recreate a proved/unproved ship note with rollback without watching you type. If they cannot, your evidence is still tribal knowledge. Tighten the run command and the metric line until a stranger can validate the MCP stdio client against local stdio server exposing one read-only echo tool.

Read

Why this stage matters for the MCP stdio client

At the mastery and shipping stage for mcp-in-code, the job is narrower than finishing a product demo. You are creating one progressive evidence piece about local stdio server exposing one read-only echo tool that later pages inherit without redefining success. Keep that fixture small enough to inspect by hand, keep outputs copy-pasteable as text, and refuse to narrate this baseline as if it were a production SLA: direct function call without MCP discovery.

For this page specifically, success looks like a proved/unproved ship note with rollback while still centering the user decision to connect to a local MCP server, list tools, and invoke one allowlisted tool safely. If you cannot point to a file, command, or assertion that proves that for the MCP stdio client, stay on this page instead of advancing.

Glossary: tool · Glossary: structured output · Cheatsheet: production ops signals

Previous

Go deeper

Before you start

Why this matters

List two things this chapter proved on the fixture and two things it did not prove about the MCP stdio client. If you cannot name the gaps, you are not ready to ship the narrative—even if the code runs.

In the wild

See how this idea shows up as a product and a company — then come back to the lesson. Skills transfer across vendors.

Check your understanding

Page assessment

Answer from memory. Completion is saved from this evidence, not from opening the next page.

1. Are proved and unproved lists both non-empty?
2. Is rollback concrete (command or version pin)?
3. Would a stranger reproduce the metric on the fixture?
4. Does the note still center the decision: connect to a local MCP server, list tools, and invoke one allowlisted tool safely?

All responses are required.