Agent Loops And Why They Are Risky

An agent loop is when an agent does the same thing over and over because it did not realize the task was done.

Famous case: an agent that was supposed to send 'one' email sent 1,000 because each retry felt like a 'first try' to it.

Three ways to prevent loops

• Hard limit on retries (max 3)
• Rate limits per minute
• Human approval required if same action retried

The big idea: Agent loops can blow through your budget fast. Always set retry limits.

Practice this safely

Try this with a school, hobby, or family example where the stakes are low. Use the AI output as a draft you can question, not as the final answer.

1. 1Ask AI to explain runaway in plain language, then underline anything that sounds uncertain or too broad.
2. 2Give it one detail from "Agent Loops And Why They Are Risky" and ask for two possible next steps plus one reason each step might be wrong.
3. 3Check rate limit against a trusted source, teacher, adult, expert, or original document before you use it.