continuous-learning
Warn
Audited by Socket on Sep 19, 2026
1 alert found:
SecuritySecuritySKILL.md
MEDIUMSecurityMEDIUM
SKILL.md
SUSPICIOUS: the skill is internally coherent as a pentest memory aid, but its actual effect is to strengthen an AI agent’s offensive-security capability over time. There is no installer or credential-stealing behavior, yet the combination of attack-pattern retention, defense-evasion content, persistent memory, and broad execution-capable tools makes it high risk for misuse.
Confidence: 85%Severity: 72%
Audit Metadata