mutation-design-aav
Warn
Audited by Socket on Mar 25, 2026
1 alert found:
SecuritySecuritySKILL.md
MEDIUMSecurityMEDIUM
SKILL.md
SUSPICIOUS. The skill is internally coherent for computational AAV mutant design, but it relies on opaque shared-file downloads for a PyTorch oracle model that is loaded with torch.load(), creating a high supply-chain and arbitrary code execution risk. No credential theft or exfiltration is evident, so this is not confirmed malware, but the install/execution trust is poor and the biological optimization target is sensitive.
Confidence: 91%Severity: 86%
Audit Metadata