Bug: Model Behavior: changes API style of existing API when adding to it despite beeing told not to

[FreddyE1982]

I had a nice simple api that used dot structures only like

something = something.something
something.something = something

I asked the agent to add some functionality to that api and explicitly told it NOT to change the api style explaining things as above.

It changed the whole api anyway to use functions and complicated structures instead.

PS:

You really need to make the model better adhere to specific user instructions

[github-actions]

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

• Keeps creating branches after being explicitly told not to #5101
• bug: model behavior: model randomly "simplifies" existing logic, breaking stuff in the process #5096
• bug: codex web: agent randomly ignores parts of AGENTS.md #5093
• codex still keeps trying to do minimal implementations of stuff even if explicitly told not to in AGENTS.md and prompts #4087

Powered by Codex Action

[etraut-openai]

Thanks for the bug report. We've shared the feedback with the team members who are responsible for training our codex models. We're tracking this feedback and similar reports so we can make the model more robust in future iterations.

If you see other model behavior that you'd like to report, please use the recently-added /feedback command in the CLI. That will provide us with additional details including logs and conversation text.