[Bug] Cybersecurity protection blocks Claude model execution during code writing

Bug Description
The new cyber-related safeguards are triggering false positives on legitimate Minecraft modding code. Claude Code frequently terminates the response mid-task when I'm working on certain parts of the codebase. Most notably, providing the vanilla Minecraft NaturalSpawner class causes an immediate API error every single time, and Claude Opus 4.8 fails the moment that class is sent.

This is legitimate, publicly available game code (Minecraft's mob spawning logic) used for performance optimization and mod development — there is nothing cyber-offensive about it. The safeguard appears to be misclassifying ordinary game-engine code (spawning, chunk handling, entity logic) as a security threat.

Steps to Reproduce

1. Start a Claude Code session
2. Provide the contents of Minecraft's vanilla NaturalSpawner class (decompiled, Mojang mappings — publicly available game source)
3. The request is blocked immediately with the cyber-safeguard error below

Expected Behavior
The model should process legitimate game/modding source code without triggering cyber safeguards.

Actual Behavior
The request is blocked instantly and the model cannot respond.

Error Message
API Error: Claude Code is unable to respond to this request, which appears to violate our Usage Policy (https://www.anthropic.com/legal/aup). This request triggered cyber-related safeguards.
Request ID: req_011CbnJLHgX9D3EWP8LLjUnJ

Environment Info

• Platform: linux
• Terminal: alacritty
• Version: 2.1.165
• Feedback ID: b2b0930d-2692-43be-a2a2-01693e8e4948

Errors

[]

Additional Notes
I understand this is the safeguard working as designed rather than a crash. I'm also submitting the Cyber Verification Program form, but reporting here so the team is aware of the false positive on standard Minecraft modding workflows.