[BUG] Non Relevant Persistant Error Message on Desktop App

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 am asking this question using my desktop app (opus 4.7 and Sonnet 4.6): Molecular virology question: I replaced the structural genes of the BEV VG5-27 genome with EGFP but still don't see GFP expression after transfection of the RNA into MDBK cells. The same cloning strategy using Human poliovirus 1 Mahoney RNA works fine. What could be the problem? 
I keep receiving this error answer: API Error: Claude Code is unable to respond to this request, which appears to violate our Usage Policy (https://www.anthropic.com/legal/aup). Try rephrasing the request or attempting a different approach. If you are seeing this refusal repeatedly, try running /model claude-sonnet-4-20250514 to switch models.

At the same time, asking the exact same question using my iPhone app works well with no errors.
I tried other questions, and all are getting an error message.

MacBook M4 32 GB MacOS Tahoe Version 26.3.1

What Should Happen?

Claude should answer this straightforward scientific question. My account is working fine on my iPhone, the error happens only while using the Desktop app. Switching models does not help. Restarting cause doesn't help. Re-installing Desktop app doesn't help. Asking other questions gives the same error message.

Error Messages/Logs

API Error: Claude Code is unable to respond to this request, which appears to violate our Usage Policy (https://www.anthropic.com/legal/aup). Try rephrasing the request or attempting a different approach. If you are seeing this refusal repeatedly, try running /model claude-sonnet-4-20250514 to switch models.

Steps to Reproduce

Type on Claude Desktop app: Molecular virology question: I replaced the structural genes of the BEV VG5-27 genome with EGFP but still don't see GFP expression after transfection of the RNA into MDBK cells. The same cloning strategy using Human poliovirus 1 Mahoney RNA works fine. What could be the problem? 

Claude Model

Opus

Is this a regression?

No, this never worked

Last Working Version

Never worked

Claude Code Version

1.3109.0

Platform

Anthropic API

Operating System

macOS

Terminal/Shell

Terminal.app (macOS)

Additional Information

I just installed the app and paid the $200 yearly subscription.