helpmelearn
Warn
Audited by Socket on Sep 6, 2026
1 alert found:
AnomalyAnomalySKILL.md
LOWAnomalyLOW
SKILL.md
SUSPICIOUS: the core tutoring/book-building workflow is coherent and the package-manager install advice is proportionate, but the optional finishbook.sh materially expands the skill into autonomous code-and-file generation with Claude Code running in detached tmux under bypassPermissions. There is no clear credential theft, exfiltration, or malicious install path, so this is not malware, but it is a medium/high-risk skill because it enables unattended actions beyond a typical learning assistant.
Confidence: 91%Severity: 62%
Audit Metadata