[Bug] Overly aggressive security safeguard triggering on benign security/cyber terminology
Status Closed — not planned
Reported on v2.1.222
Maintainer reply None cached
Activity 2 comments · opened Aug 5, 2026 · closed Sep 8, 2026
Bug Description
Keeps falling back into the security safeguards, and downgrades the model, which is rather annoying, it almost seems that when it sees the word cyber, or security, it will immediatly be triggered; there was no malicious thing going on here, just me trying to improve my product.
Environment Info
- Platform: darwin
- Terminal: zed
- Version: 2.1.222
- Feedback ID: a6e14a2e-03e9-4686-a32e-63cc86417e04
Errors
[]This issue has 2 comments on GitHub. Read the full discussion on GitHub ↗