[Bug] Fable 5.1 overly aggressive security safeguards causing false positives for legitimate security work

Bug Description
Subject: Fable 5.1 false positives are blocking legitimate security work I've been using Claude for months, recently moved to the Max plan, and have been an advocate for the product. I'm a CVP participant and an attested cybersecurity professional, and I want to flag a problem that is now affecting my willingness to keep paying. I have not been able to complete a single task on Fable 5.1. The safeguards redirect me even on clearly non-risk work. As one example, designing the architecture for an agentic purple team (SOC and red team) is defensive design with no request to generate offensive artifacts, and it still gets redirected. Because nearly all of my work sits in the security domain, the topic-based routing appears to trigger on almost every session I run. I understand the intent behind the safeguards. My concern is the false-positive rate for professionals whose entire workflow is security. In its current state, Fable is effectively unusable for me, which makes the Max subscription hard to justify. I'd welcome the chance to share session-level examples so your team can see the pattern and use it to reduce false positives. I want to keep advocating for Claude, and I'm hoping this feedback helps get there.

Environment Info

• Platform: darwin
• Terminal: Apple_Terminal
• Version: 2.1.277
• Feedback ID: 441a708b-6bc8-4f62-b742-e0f986c782a8

Errors

[]