[Feature Request] Allow user override for model selection on safety-flagged prompts

Bug Description
It would be a great feature if one could use the Fable model to scan an internal codebase for potential attack vectors. Fable currently automatically flags such prompts for cybersecurity or biology topics and switches to Opus.

Environment Info

• Platform: darwin
• Terminal: ghostty
• Version: 2.1.170
• Feedback ID: 3b698c66-ccf1-44c7-af93-81ea70df80cb

Errors

[]