What is Tool Use?
Tool use refers to an AI agent's ability to interact with external systems — calling APIs, executing code, querying databases, writing files, or performing any operation — extending its capabilities beyond text generation through protocols like MCP.
WHY IT MATTERS
A language model without tools is sophisticated autocomplete. With tools, it becomes an agent that can act on the world. Tool use is what transforms 'I recommend creating a file called config.yaml' into actually creating it.
Modern tool use follows a standard pattern via MCP: the developer defines available tools with schemas on an MCP server, the agent discovers and invokes them through the MCP protocol, the server executes the tool, and the result is returned for further reasoning.
The tools available to an agent define its capability boundary. A coding agent with access to read, write, and execute tools can refactor entire codebases. The critical question is: which tools should the agent have access to, with what argument constraints, and at what rate?
HOW POLICYLAYER USES THIS
PolicyLayer governs all MCP tool use through YAML-defined policies. When an agent invokes any MCP tool, PolicyLayer evaluates the call — checking the tool name against allow/deny lists, validating arguments against constraints, and enforcing rate limits — before forwarding it to the server. This ensures tool use stays within authorised boundaries. No code changes to the agent or server.