LLM agent managing a simulated GPU inference cluster
Step through an agentic file‑editing environment