multi-agent-orchestration
Warn
Audited by Socket on Jun 19, 2026
1 alert found:
AnomalyAnomalySKILL.md
LOWAnomalyLOW
SKILL.md
SUSPICIOUS. The skill’s purpose broadly matches orchestration, but it expands the agent’s autonomy and allows task content to flow to unspecified external agents. There is no direct malware behavior, no credential theft path, and no concrete installer risk in the provided text, but the external-agent routing and autonomous continuation make the overall risk medium.
Confidence: 84%Severity: 56%
Audit Metadata