---
title: "Tool Use | AgentRQ Glossary"
url: https://agentrq.com/glossary/tool-use
description: "Tool use is the ability of an AI agent to call external functions — such as reading files, running code, or sending messages — enabling it to affect the real world beyond text generation."
---

[Back to Glossary](https://agentrq.com/glossary/)

# Tool Use

**Tool use** is the capability of an AI [agent](https://agentrq.com/glossary/agent) to call external functions during inference, enabling it to interact with systems beyond pure text generation. When an agent uses a tool, it generates a structured call that the runtime intercepts, executes, and returns results from — allowing the agent to read files, run code, query databases, send messages, and take real-world actions.

## How Tool Use Works

1. The LLM is provided a list of available tools with their names, descriptions, and parameter schemas
2. During generation, the model decides to call a tool and outputs a structured tool call
3. The runtime (e.g., [Claude Code](https://agentrq.com/glossary/claude-code) or an [MCP client](https://agentrq.com/glossary/mcp-client)) intercepts the call
4. The tool executes and returns a result
5. The result is injected back into the model's context as a new message
6. The model continues reasoning with the tool result available

## Tool Use in Claude Code

[Claude Code](https://agentrq.com/glossary/claude-code) is built around tool use. Its core tools include:

- → `Read` — Read file contents
- → `Write` / `Edit` — Create or modify files
- → `Bash` — Execute shell commands
- → `Glob` / `Grep` — Search the codebase
- → `Agent` — Spawn sub-agents for parallel work
- → MCP tools — Any tools from connected [MCP servers](https://agentrq.com/glossary/mcp-server)

## Tool Use via MCP

The [Model Context Protocol](https://agentrq.com/glossary/mcp) standardizes tool use for remote services. AgentRQ exposes tools like `task`, `memory` and `skills` as [MCP tools](https://agentrq.com/glossary/mcp-tool), each taking an `action` and a `payload` (for example `task` with `create`, `reply` or `get`), which [Claude Code](https://agentrq.com/glossary/claude-code) calls just like any other tool in its repertoire.

## Safe Tool Use

Not all tools are equal in risk:

| Risk Level | Examples | Recommended Pattern |
| --- | --- | --- |
| Low | Read files, search code | Autonomous |
| Medium | Write files, run tests | Autonomous with logging |
| High | Deploy to production, delete data | Require [approval](https://agentrq.com/glossary/approval) |

AgentRQ lets you enforce approval gates for high-risk tool calls via [human-in-the-loop](https://agentrq.com/glossary/human-in-the-loop) workflows.

## Related Terms

- → [MCP Tool](https://agentrq.com/glossary/mcp-tool)
- → [MCP](https://agentrq.com/glossary/mcp)
- → [Claude Code](https://agentrq.com/glossary/claude-code)
- → [Agent](https://agentrq.com/glossary/agent)
- → [Approval](https://agentrq.com/glossary/approval)