hermes-integration
Warn
Audited by Socket on Apr 18, 2026
1 alert found:
AnomalyAnomalySKILL.md
LOWAnomalyLOW
SKILL.md
SUSPICIOUS: The skill is broadly consistent with Hermes integration, and the upstream project appears legitimate, but it grants a wide, credential-bearing MCP/CLI layer the ability to read conversations and autonomously send messages or run scheduled tasks. Main concerns are broad scope, versionless external-tool trust, and autonomous cross-platform actions rather than clear malware or credential theft.
Confidence: 84%Severity: 68%
Audit Metadata