Best way to implement AI Guardrails in Xano

How can I ensure that users can ask just appropriate questions to my Xano AI agent?