Anthropic API Incorrectly Flags Question About Data Quarantine Logic as Potential Malware Evasion

[charles-dyfis-net]

Bug Description
Incorrect AUP violation detection from a question related to debugging code that deliberately "quarantines" inert data samples by putting them into a holdout set when processing against them fails, and then reintroduces them later -- I presume that whatever model was used for checking misinterpreted this as being a request related to building evasion tools for malware.

Environment Info

• Platform: darwin
• Terminal: ghostty
• Version: 1.0.70
• Feedback ID: 1db608b4-8dff-4640-bf28-9a07702b1202

Errors

[]

[github-actions]

Found 3 possible duplicate issues:

1. [BUG] protein research & Usage Policy #2687
2. Anthropic API Usage Policy Violation Blocking Research Request #3518
3. GitHub CLI API Request Blocked Incorrectly as Usage Policy Violation #4870

This issue will be automatically closed as a duplicate in 3 days.

• If your issue is a duplicate, please close it and 👍 the existing issue instead
• To prevent auto-closure, add a comment or 👎 this comment

🤖 Generated with Claude Code

[charles-dyfis-net]

I'm not sure those apply -- they're related, but determining whether the same fix that addresses those bugs would fix this one as well calls for knowledge of implementation details that, to the best of my knowledge, aren't public.

[github-actions]

This issue has been inactive for 30 days. If the issue is still occurring, please comment to let us know. Otherwise, this issue will be automatically closed in 30 days for housekeeping purposes.

[github-actions]

This issue has been automatically closed due to 60 days of inactivity. If you're still experiencing this issue, please open a new issue with updated information.