Guard-railing notification

[peuportier]

Dear Team,

I hope you're having a great day.

It would be beneficial to implement a stricter guardrail system, similar to the one used by OpenAI, to manage violations. Additionally, it would be useful for administrators to have a way to track users who attempt to violate guidelines. Accounts with excessive violations could face suspension, with a staged approach including various warnings depending on the context. Please ensure this is aligned with our company policies and provide a link to the relevant policy document for example.

Thank you.