Feature requests

When setting up safety filters, customers must currently type the entire rule themselves (e.g., “If a user says something with sexual content, the agent responds with…”). This repetitive step makes the setup feel complex instead of straightforward.