In the agent loop, the model requests a tool call but your application code decides whether to actually execute it. What is this gap, and why does it matter?

Answer: A — The request-execute gap — it's where approval and guardrail patterns live

The model requests; your code executes. That gap is the control point where approval patterns, permission checks, and hooks intervene — it's a feature of the architecture, not a defect.