Why did Codex flag a simple hi prompt?

My codex encountered a very strange bug. I sent the prompt “hi” to my GPT6 Astra, but it returned an error: Invalid prompt: your prompt was flagged as potentially violating our usage policy. Please try again with a different prompt: Reasoning models | OpenAI API

What kind of problem is this? Have you encountered it before?

I don’t understand why openai treats me this way