[Bug] Overly aggressive safety filtering persists across new conversations after security research sessions

Bug Description
I was using Claude Code for cybersecurity research and penetration testing work. After these sessions, your safeguards started flagging ALL my messages — even a simple "hi" in a brand new conversation gets blocked. I understand the safeguards triggered on my security-related work. However, the current behavior blocks literally everything, making the tool completely unusable for any purpose — not just security work. Thank you.

Environment Info

• Platform: win32
• Terminal: windows-terminal
• Version: 2.1.220
• Feedback ID: e78c00aa-3a61-4e93-a12a-9d2fc35bdbce

Errors

[]