It was three strikes, I would get notified over email for each one and the third strike got my account banned. I guess the strikes are one step up from the automatic guardrail refusal messages, I would get one after getting too many refusals in a row or I guess when I did work the model deemed 'particularly sensitive'?

I tried to appeal, hoping their automated system would take pity on me but predictably it was instantly denied lol

Now granted, I was doing actual offensive cybersecurity work without being enrolled in the cyber program, and I have plenty of spare alternate accounts anyway so it wasn't that big of a deal. It actually already net me a few 0days, so their guardrails aren't very effective.

I’m not the person you’re replying to, but for me just a few attempts at doing cybersecurity related things (not even RE) with ChatGPT triggered a warning by email. A few days later I was banned. No appeal.