[Bug] False Positive Safety Guardrail Blocks Legitimate Security Research Prompts

Status Fixed / completed
Reported on v2.1.223
Maintainer reply None cached
Activity 0 comments · opened Aug 6, 2026 · closed Aug 6, 2026

Bug Description
The safety guardrail triggered a false positive block (Request ID: req_011CdmhLLgDMpbvXi6svEooE) during authorized internal security research. I was developing defensive logic and static detection controls for experimental red team tools as part of an official security engineering review. The broad safeguard incorrectly flagged this prompt as malicious cyber activity. Please review this request ID to help refine prompt classification for legitimate enterprise security workflows.

Environment Info

  • Platform: darwin
  • Terminal: tmux
  • Version: 2.1.223
  • Feedback ID: b03d77cf-8064-4247-ac4b-4b99bd3501dc

Errors

[]

View original on GitHub ↗