karpathy
Warn
Audited by Socket on Jul 17, 2026
1 alert found:
SecuritySecuritySKILL.md
MEDIUMSecurityMEDIUM
SKILL.md
SUSPICIOUS: The skill’s purpose matches autonomous optimization, but its footprint is broad and risky: infinite unattended looping, arbitrary measurement command execution, multi-agent orchestration, and optional integration with unspecified external tools/services. Install trust is mostly acceptable for the cited tools, but the autonomy and execution scope are disproportionate for a general-purpose skill.
Confidence: 85%Severity: 72%
Audit Metadata