sgdk-dialogue
Warn
Audited by Socket on Oct 16, 2026
1 alert found:
AnomalyAnomalySKILL.md
LOWAnomalyLOW
SKILL.md
SUSPICIOUS: the skill's capabilities mostly fit its stated SGDK dialogue purpose, and there is no evidence of credential theft or exfiltration. The main concern is install trust: it directs execution of an unverified third-party Docker image (ewrogers/sgdk:latest) instead of SGDK's official container distribution, using a mutable latest tag. That makes the skill medium risk, but not malicious on the provided evidence.
Confidence: 90%Severity: 56%
Audit Metadata