Interact with an agentic environment using text inputs
Auto‑fix buggy code with reinforcement‑learning agent
Generate dumpling images based on mood