tool-use

To enable LLMs to call external tools, this skill uses structured tool calls where the model outputs a tool name and parameters, the system executes the tool, and results are fed back into the reasoning process to retrieve real-time data without hallucination.

Function calling connects LLMs to external APIs by applying structured tool schemas that let the model invoke services like databases or weather systems, fetch current information, and process the returned results within its reasoning loop.

To perform automated actions safely, this skill provides observable execution with clear feedback loops, error handling, and fallbacks to maintain reliability when integrating LLMs with messaging platforms, databases, or other external systems.

Yes, structured tool schemas are required for API integration with an LLM agent because they define the tool name and parameters the model must output to execute the function and return the result for processing accurately.

LLMs hallucinate real-time data when they rely solely on text generation, but using structured tool calls to query external data sources and feeding the results back into the reasoning process prevents this by returning factual information.