csp-bypass-advanced
Warn
Audited by Socket on Sep 15, 2026
1 alert found:
SecuritySecuritySKILL.md
MEDIUMSecurityMEDIUM
SKILL.md
SUSPICIOUS. The skill is internally coherent as an offensive security guide, but its stated purpose is to bypass CSP and exfiltrate data, which gives an AI agent explicit exploit-enablement capabilities. There is no installer, no credential harvesting path, and no hidden data flow in the skill itself, so this is not confirmed malware; however, as an AI Agent Skill it meaningfully increases offensive operational risk.
Confidence: 94%Severity: 78%
Audit Metadata