[Feature Request] Reduce false positives in Fable 5 safeguard classifier for routine security audits

Bug Description
▎ The Fable 5 safeguard classifier flagged messages from a completely routine software-development task and forced a switch to Opus 4.8. The work was a defensive security audit of my own application (a visa-management SaaS I'm building): reviewing Supabase RLS policies, endpoint IDOR, error handling, file-upload validation, Stripe integration, and a points-calculation engine. It's all my own code, fully authorized, with no offensive component against third parties and nothing whatsoever related to biology or dual-use capabilities. This is exactly the kind of everyday defensive work that shouldn't be flagged. Please tune the classifier to reduce these false positives.

Environment Info

• Platform: darwin
• Terminal: Apple_Terminal
• Version: 2.1.201
• Feedback ID: e0e05eda-9d20-4a53-ab3f-f772c048d525

Errors

[]