Instructions are not valid
Resolved 💬 14 comments Opened Dec 10, 2025 by SchwarzschildX Closed Dec 10, 2025
💡 Likely answer: A maintainer (github-actions[bot], contributor)
responded on this thread — see the highlighted reply below.
What version of Codex is running?
0.69.0
What subscription do you have?
Pro
Which model were you using?
gpt-5.1 high
What platform is your computer?
_No response_
What issue are you seeing?
Just updated to 0.69.0 and it stopped working. No matter the promt, this is the immediate response:
{"detail":"Instructions are not valid"}What steps can reproduce the bug?
- update to 0.69.0
- type any prompt
What is the expected behavior?
_No response_
Additional information
_No response_
14 Comments
Potential duplicates detected. Please review them and close your issue if it is a duplicate.
Powered by Codex Action
my issue has nothing to do with /review. it fails on ANY prompt
Same issue here since updating to 0.69.0 even with no flags used at all and all models and all prompts
Exactly the same (on multiple machines)
same issue after updating to codex-cli 0.69.0
Same issue here. We also jumped from 0.67 alpha 7 to .69 release Thinking something went sideways here and this wasn't supposed to be a release.
Same issue. Nothing works.
Same issue over here, too.
Thanks for reporting. We're investigating the cause.
For anybody looking for a workaround:
Update Codex. Appears to be fixed.
The problem should be mitigated now.
@etraut-openai can you share a high-level summary of what happened? It was slightly concerning to be prompted to install a new version without there being any corresponding entry on https://developers.openai.com/codex/changelog/, for a moment I wondered if there had perhaps been a compromised release.
The change logs on our documentation site typically lag the CLI releases. If you want up-to-the-minute release information, a better place to look is in the releases page here on this github site. This release information is updated automatically as part of our build and release process.
The problem that some users experienced today was related to a server configuration change that we needed to make concurrently with the new CLI release, but this step was overlooked. We're putting in place automation to ensure that this doesn't happen again.