codex gpt5.5 seems to be completely crazy in a clean session with a easy 'ask question' prompt
Open 💬 0 comments Opened May 29, 2026 by Meowzz95
What version of the Codex App are you using (From “About Codex” dialog)?
Version 26.519.81530 (3178)
What subscription do you have?
Pro $200
What platform is your computer?
Darwin 25.4.0 arm64 arm
What issue are you seeing?
feedback id: 019e72fb-5623-7e43-b6b5-886be3080d4e
with a very simple prompt and $ask-questions-if-underspecified skill, the output seems to be crazy, like it is mixing COT and actual response
What steps can reproduce the bug?
seems to be random and i notice low quality thread twice today
What is the expected behavior?
behave normally
Additional information
_No response_