[Bug] Overly aggressive safety guardrails blocking legitimate code security audits

Bug Description
Fable is unusable... Each time I start an audit of my own code, I am blocked by guardrails... even if I use a prompt in my own folder, which is a git folder...
Even when simply discussing with the chat about this issue, Fable's safety measures are triggered...
Here's what Claude responds: The prompts to the sub agents are dense with words like "SQL injection", "RCE", ... A classificator whom is scanning cybersecurity content is triggered on these words without distinguish offensive and defensive...
What is the purpose of you saying we can audit code if there is no context possible...

Environment Info

• Platform: darwin
• Terminal: WarpTerminal
• Version: 2.1.170
• Feedback ID: 4a25afe7-b47c-4b82-9f1f-eef23e53cc9e

Errors

[]