›_ Codex Issues
Codex Issues / #6804

Codex got very lost debugging a C crash that resulted from stubs that Codex had itself created

Resolved 💬 2 comments Opened Nov 17, 2025 by jsavin Closed Nov 17, 2025
💡 Likely answer: A maintainer (etraut-openai, contributor) responded on this thread — see the highlighted reply below.

What version of Codex is running?

codex-cli 0.58.0

What subscription do you have?

Business

Which model were you using?

gpt-5.1-codex

What platform is your computer?

macOS

What issue are you seeing?

See uploaded logs and title of this issue. Codex burned many tokens trying to debug a memory-related crash in a C library (legacy dependency for the large C application refactor I'm working on). The cause turned out to be incomplete stubs that Codex had itself created. It took multiple sessions for it to finally find the root cause and is now making progress again, but it should have been "smart" enough to check this, or at least to document the stubs itself. I added guidance in AGENTS.md to avoid this situation in the future.

What steps can reproduce the bug?

Uploaded thread: 019a8e67-95d2-7c91-8f94-47341d795a21

What is the expected behavior?

_No response_

Additional information

_No response_

View original on GitHub ↗

2 Comments

etraut-openai contributor · 9 months ago

Thanks for the bug report. We'll share the feedback with the team responsible for training our codex models. We're tracking this feedback and similar reports so we can make future models more robust and efficient. And thanks for using the /feedback command. That additional information is useful.

jsavin · 9 months ago

Thanks, Eric. I appreciate the quick response.

On Nov 17, 2025, at 3:55 PM, Eric Traut @.***> wrote: etraut-openai left a comment (openai/codex#6804) <https://github.com/openai/codex/issues/6804#issuecomment-3544377031> Thanks for the bug report. We'll share the feedback with the team responsible for training our codex models. We're tracking this feedback and similar reports so we can make future models more robust and efficient. And thanks for using the /feedback command. That additional information is useful. — Reply to this email directly, view it on GitHub <https://github.com/openai/codex/issues/6804#issuecomment-3544377031>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAN6VAUVRCXACPHUHDXJEL335JOALAVCNFSM6AAAAACMMPIUU2VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZTKNBUGM3TOMBTGE>. You are receiving this because you authored the thread.