[Bug] Safeguard filters contaminate conversation state and prevent model recovery

Status Closed — not planned
Reported on v2.1.221
Maintainer reply None cached
Activity 1 comment · opened Aug 4, 2026 · closed Sep 8, 2026

Bug Description
The safeguards are shit and trigger on things that the model decides to do itself and taint conversations so that you can't switch back. It's infuriating. At no time did I ask for security, it completed PHP work building a plugin and decided to do a security scan and then broke the whole conversation

Environment Info

  • Platform: darwin
  • Terminal: xterm-256color
  • Version: 2.1.221
  • Feedback ID: b3b4ff6c-a877-40a6-91a8-37dfa529d127

Errors

[]

View original on GitHub ↗

This issue has 1 comment on GitHub. Read the full discussion on GitHub ↗