[BUG] Flagging a prompt puts Claude in a dysfunctional state

Preflight Checklist

• [x] I have searched existing issues and this hasn't been reported yet
• [x] This is a single bug report (please file separate reports for different bugs)
• [x] I am using the latest version of Claude Code

What's Wrong?

When Claude rejects a prompt, the message does not explain the reason and does not explain how much information "rollback" will lose. Once the message has been written, Claude rejects all further prompts, even a single question mark.

What Should Happen?

Claude should allow ignoring that and continuing. Barring that, Claude should provide an option to ignore the flagged prompt and continue accepting new prompts. Ideally, there should be an explanation of what part of the AUP was apparently violated and a process for requesting human review.

Error Messages/Logs

Steps to Reproduce

I got he message after requesting a song parody of Maria centered around scolia.

Claude Model

None

Is this a regression?

Yes, this worked in a previous version

Last Working Version

_No response_

Claude Code Version

2.2553.1 (Claude desktop app for Windows)

Platform

Anthropic API

Operating System

Windows

Terminal/Shell

Windows Terminal

Additional Information

This seems to happen for any rejected prompt.