assistant-x-openclaw-jarvis-voice-assistant
Warn
Audited by Socket on Sep 12, 2026
1 alert found:
AnomalyAnomalySKILL.md
LOWAnomalyLOW
SKILL.md
SUSPICIOUS. The skill’s capabilities are largely coherent with a local voice-assistant purpose, and its documented data flows are mostly localhost/OpenClaw-native rather than third-party exfiltration. The main concern is install/provenance trust: the skill is branded under ara.so/Hermes Skills but instructs cloning a personal GitHub repo, plus it asks for broad OpenClaw admin scope and offloads model provenance to external README links. No confirmed malicious behavior or credential-harvesting endpoint is shown.
Confidence: 89%Severity: 52%
Audit Metadata