de-ai-documents
Warn
Audited by Socket on Sep 18, 2026
1 alert found:
AnomalyAnomalySKILL.md
LOWAnomalyLOW
SKILL.md
SUSPICIOUS. The skill is internally coherent as a document-rewriting guide, but its stated purpose is to help disguise AI-generated artifacts as human-authored, including manipulation of notebook execution metadata and authenticity signals. There is little classic malware behavior—no credential access, exfiltration, or untrusted installs—but the deceptive objective and trace-shaping instructions make it a meaningful security/integrity risk.
Confidence: 91%Severity: 62%
Audit Metadata