Agent Loop

The core execution cycle of an AI agent: receive input → reason about the task → decide on an action → execute the action → observe the result → reason again → repeat until the task is complete or a termination condition is met. The agent loop is what distinguishes agents from single-shot LLM calls—the model iterates, using tool results and intermediate observations to refine its approach. Loop design determines agent capability: how many iterations are allowed, what termination conditions exist, how errors trigger retries, and whether the agent can request human input mid-loop.