RuntimeAgent systems topic

Tools and MCP

Tools let models request deterministic functions; MCP standardizes how compatible hosts discover and call external tools, resources, and prompts. Good tool design minimizes ambiguity, constrains inputs, makes side effects visible, and returns results the model can interpret reliably.

Apply the lifecycleReview sources

Questions to answer

Resolve these before adding tools, frameworks, or automation.

  • Is the capability read-only, reversible, consequential, or externally stateful?
  • What schema and descriptions make correct tool selection likely?
  • How are authentication, authorization, timeouts, retries, and output size controlled?

Implementation lifecycle

Build the evidence in this order

  1. 01

    Contract

    Define narrow inputs, outputs, errors, side effects, and authorization requirements.

  2. 02

    Connect

    Expose the capability through the simplest appropriate application or protocol boundary.

  3. 03

    Verify

    Test discovery, selection, malformed input, permission denial, timeout, and duplicate execution.

Artifacts to maintain

  • tool schema
  • MCP server
  • auth policy
  • approval rules

Continue through AgentAtlas

Evidence ledger

Primary documentation behind this guide

Model Context Protocol

MCP specification

The protocol model, capability negotiation, tools, resources, prompts, transports, authorization, and security considerations.

OpenAI

Connect Codex to MCP servers

How Codex discovers, configures, authenticates, and uses MCP servers and their capabilities.

Next topic

Multi-agent systems