scenario-model-training
Warn
Audited by Socket on Sep 3, 2026
1 alert found:
AnomalyAnomalySKILL.md
LOWAnomalyLOW
SKILL.md
SUSPICIOUS: The core behavior is coherent with Scenario model training and data appears to flow to official Scenario services, not obvious exfiltration endpoints. The main concerns are transitive skill installation, mutable unpinned install source, and the ability to trigger paid/state-changing training actions without strong approval boundaries.
Confidence: 87%Severity: 52%
Audit Metadata