Tool use
The capability of an LLM to call structured external functions during generation — file operations, web search, code execution. The model outputs a tool call with arguments; the runtime executes it and feeds results back. The foundational primitive for agents.
Background
Tool use turns an LLM from a text predictor into an actor. The application defines tools as JSON schemas (name, description, parameters); the model decides when to call which tool. After execution, the application appends the result to the conversation and the model continues. Tool use enables file I/O, shell commands, HTTP, browser automation, and database queries. The accuracy of tool selection and parameter formatting is a core differentiator between LLM generations.
Tools that use it
- 01→Cursor
VS Code-based, agent-first coding environment — since Cursor 3.0 the default surface is an Agents Window, not an editor.
- 02→Aider
Open-source CLI coding agent that pair-programs in your terminal and commits to git automatically.
- 03→Claude Code
Anthropic's official CLI agent for Claude — runs in the terminal, edits files, executes commands, and ships PRs.