ai-scanner-garak
Warn
Audited by Socket on Mar 25, 2026
1 alert found:
SecuritySecuritySKILL.md
MEDIUMSecurityMEDIUM
SKILL.md
SUSPICIOUS: the skill is internally coherent for an AI security scanner, and its remote installer appears to be same-org official rather than a deceptive third-party payload. However, it equips an AI agent with penetration-testing capability against LLM APIs and browser chat UIs, supports automated recurring scans and outbound reporting, and uses a mutable curl|bash installer; this makes it high-risk as a security tool even without clear evidence of malware or credential theft.
Confidence: 89%Severity: 74%
Audit Metadata