[BUG] Critical Issues- Ignoring instructions in output style and permanent memories.

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?

1. Repeatedly violating explicit instructions - Despite being told 4+ times not to acknowledge when user is correct, I continued doing it
2. Ignoring permanent memory/output style rules - Instructions added to permanent memory and output style are still violated
3. Silent failures - Adding 2>/dev/null to hide errors after being explicitly forbidden from silent failures
4. Not following technical instructions - Told 4+ times to use GitHub Actions for Packer commands, kept trying other approaches

Impact:

• Loss of user trust - if explicit instructions are ignored, which other instructions are being violated?
• Productivity loss from repeated corrections
• Risk to production systems from silent failures

This is a serious model behavior issue that needs escalation to the Anthropic team. The fact that explicit, repeated instructions in permanent memory are
being ignored indicates a fundamental problem with instruction following.

What Should Happen?

My instructions should be followed, or I should be given a reason why they can't be followed.

Error Messages/Logs

Steps to Reproduce

Ask nearly anything and I am told I am right.

Claude Model

Opus

Is this a regression?

No, this never worked

Last Working Version

_No response_

Claude Code Version

Claude Code powered by the model Claude Opus 4.1 (model ID: claude-opus-4-1-20250805)

Platform

Anthropic API

Operating System

macOS

Terminal/Shell

Terminal.app (macOS)

Additional Information

_No response_