Glossary — Agentic AI

What is an MCP Tool?

1 min read Updated

An MCP tool is an executable capability exposed by an MCP server, described with a name, description, and JSON Schema parameters, that AI agents can discover and invoke through the Model Context Protocol.

WHY IT MATTERS

MCP tools are the action primitives of the protocol. They represent things an agent can do — send a payment, query a database, create a file, check a spending limit. Each tool has a schema describing its inputs and expected outputs.

What makes MCP tools powerful is standardization. A tool defined once on an MCP server is automatically available to any compatible agent. The agent's LLM sees the tool's name, description, and parameters — enough to decide when and how to use it.

Tool design matters enormously. Well-designed tools have clear names, comprehensive descriptions (the LLM reads these to decide when to use the tool), strict input validation, and informative error messages. Poorly designed tools lead to agents misusing them.

HOW POLICYLAYER USES THIS

PolicyLayer exposes spending controls as MCP tools — check_spending_limit, validate_transaction, get_spending_history. Any MCP-compatible agent can use these tools to enforce financial policies without custom integration.

FREQUENTLY ASKED QUESTIONS

How do MCP tools differ from function calling?
Function calling is how the LLM generates structured invocations. MCP tools are how capabilities are described and served over the MCP protocol. MCP tools are invoked through function calling — they're complementary, not competing concepts.
How many tools should an MCP server expose?
Keep it focused. Each tool added to an agent's context consumes tokens and adds cognitive load for the LLM. Group related operations and use clear naming to help the model choose correctly.
Can MCP tools have side effects?
Yes, and they often do — that's the point. A payment tool transfers real money. This is why tool descriptions should clearly indicate side effects, and why spending controls are critical for financial tools.

FURTHER READING

BUILD WITH POLICYLAYER

Non-custodial spending controls for AI agents. Setup in 5 minutes.

Get Started