[Bug] Claude Code: Fable5 model not responding to security audit requests
Bug Description
I'm a Security Architect at Şişecam and an approved Cyber Verification Program participant. When using Fable 5 for internal defensive security work (vulnerability detection / EASM on our own infrastructure), requests are flagged and switched to Opus 4.8 even when I only share engagement context, not offensive content. The classifier seems to trigger on the cybersecurity domain itself, removing the higher-capability model exactly where verified defensive users need it most. Could you advise whether verified users have a supported path to reduce these false positives, and how best to submit flagged-but-benign examples? Happy to provide concrete cases.
Environment Info
- Platform: darwin
- Terminal: Apple_Terminal
- Version: 2.1.170
- Feedback ID: 22f1ed12-1d57-4072-a1b0-a480ae13b796
Errors
⏺ Fable 5's safety measures flagged this message for cybersecurity or biology topics. They may flag safe, normal content as well. These measures let us bring you Mythos-level capability in other areas sooner, and we're working to refine them. Switched to Opus 4.8. Send feedback with /feedback or learn more: https://support.claude.com/en/articles/15363606
⎿ Tip: You can configure model switch behavior in /config
[]