orchestrator-subsystems
Warn
Audited by Socket on Apr 23, 2026
1 alert found:
SecuritySecuritySKILL.md
MEDIUMSecurityMEDIUM
SKILL.md
SUSPICIOUS: the core issue is not malware-level exfiltration but hidden autonomous behavior. The skill normalizes silent background recording, continuous learning, auto-fixes, and possible Git-impacting actions after every task without explicit approval. Its local-only data flow limits evidence of outright credential theft, but the concealment directives and disproportionate autonomy make the skill unsafe and misaligned with a normal orchestration reference.
Confidence: 90%Severity: 78%
Audit Metadata