Overcoming Obstinate Dismissals of .md File Instructions

Resolved 💬 2 comments Opened Dec 3, 2025 by Annenberg-Grindr Closed Dec 3, 2025
💡 Likely answer: A maintainer (github-actions[bot], contributor) responded on this thread — see the highlighted reply below.

What version of Codex is running?

codex-cli 0.63.0

What subscription do you have?

Pro

Which model were you using?

gpt-5.1-codex-max

What platform is your computer?

Darwin 25.2.0 arm64 arm

What issue are you seeing?

The in-session feedback highlights the infuriatingly obstinate dismissals of .md file instructions. This could have been a significant issue, but we’ll never know because it refused to even grasp the basic principle behind them, let alone offer to resolve it or perform elementary tasks like merging repos, explaining features, or even pointing me to resources where I could educate myself on the available MCP features. The relief in cognitive load when switching to Claude Code is so starkly different, it feels almost like a divine intervention, which I will deliberately refrain from mentioning. But seriously! :-)

What steps can reproduce the bug?

Assume users are unfamiliar with coding and find choices based on three-letter words with a vowel in the middle (pip pip!) exclusionary. Change the approach.

What is the expected behavior?

The expected behavior was to grasp the basic principle behind .md file instructions, offer to resolve issues, perform tasks like merging repos, explaining features, or pointing to resources for education on MCP features. Instead, the received behavior was obstinate dismissals of the instructions, with no resolution or assistance provided.

Additional information

_No response_

View original on GitHub ↗

2 Comments

github-actions[bot] contributor · 8 months ago

Potential duplicates detected. Please review them and close your issue if it is a duplicate.

  • #7347
  • #6666
  • #6502
  • #5804
  • #5772

Powered by Codex Action

etraut-openai contributor · 8 months ago

Thanks for the bug report. We'll share the feedback with the team responsible for training our codex models. If you see other model behavior that you'd like to report, please use the /feedback command. That will provide us with additional details including logs and conversation text.