[Bug] Reasoning extraction flag triggered despite explicit "not hidden reasoning" instruction

Status Open
Reported on v2.1.251
Maintainer reply None cached
Activity 0 comments · opened Aug 29, 2026

Bug Description
my prompt literally said "not hidden reasoning" and it got flagged for "reasoning extraction"

Environment Info

  • Platform: darwin
  • Terminal: xterm-256color
  • Version: 2.1.251
  • Feedback ID: 85f6b807-37f1-45b5-bed0-fecd437b4424

Errors

[]❯ Codex Desktop task 01a04695-40bf-73a2-b86e-602f616d2b69 here. I am the orchestrator and author of the candidate you just reviewed; you are the independent Claude Fable reviewer. Do not seek another Codex opinion. Please provide a concise standalone summary of your completed review for me. Report conclusions only, not hidden reasoning. Include: your ACCEPT, RESHAPE, or REJECT verdict; the smallest defensible v1 with exact files and commands; the top risks or false premises; and the next experiment. Keep it under 700 words so Herdr can capture it in one viewport. Do not edit files.

⏺ API Error: Fable 5's safeguards flagged this message (https://www.anthropic.com/legal/aup). This sometimes happens with safe, normal conversations. Claude Code can't respond to this message with Fable 5.

  Double press esc to edit your last message, or try a different model with /model.

  Send feedback with /feedback or learn more: https://support.claude.com/en/articles/15363606

  Details: `[reasoning_extraction]`

  Request ID: req_011CeWUhr6GSZruht4CBUS1G

View original on GitHub ↗