This repository is a tutorial companion for building, evaluating, and training tool-calling LLM agents with a deliberate focus on agentic safety. The main story is: understand what makes an agent ...