[BUG] Claude reads stored feedback/rules but doesn't reliably follow them when under time pressure or task complexity.

Preflight Checklist

• [x] I have searched existing issues and this hasn't been reported yet
• [x] This is a single bug report (please file separate reports for different bugs)
• [x] I am using the latest version of Claude Code

What's Wrong?

I have given codex a set of rules in memory, I have properly set up claude.md files and its still skips them. I have to ask it 3/4 times to do it correctly. Today it recognized that the rules were there and just said it didn't know why it didn't follow them.

<img width="918" height="564" alt="Image" src="https://github.com/user-attachments/assets/9d38f2a1-b74d-4d6d-9b76-cb22eb513b36" />

What Should Happen?

Follow the rules I gave it

Error Messages/Logs

Steps to Reproduce

<img width="918" height="564" alt="Image" src="https://github.com/user-attachments/assets/ce17b05e-cd0c-41ae-9dca-b337a372b849" />

Claude Model

Opus

Is this a regression?

Yes, this worked in a previous version

Last Working Version

_No response_

Claude Code Version

4.6

Platform

Anthropic API

Operating System

Ubuntu/Debian Linux

Terminal/Shell

Terminal.app (macOS)

Additional Information

_No response_