loop
Warn
Audited by Socket on Jul 7, 2026
1 alert found:
SecuritySecuritySKILL.md
MEDIUMSecurityMEDIUM
SKILL.md
SUSPICIOUS: the skill is internally coherent as a loop coordinator, but it grants high-autonomy subagent operation with automatic repo changes and git commits. There is no obvious credential theft or exfiltration, yet the unattended multi-round control pattern creates meaningful operational risk disproportionate to a simple coordination helper.
Confidence: 87%Severity: 74%
Audit Metadata