gpt-image
Warn
Audited by Socket on Apr 25, 2026
1 alert found:
AnomalyAnomalySKILL.md
LOWAnomalyLOW
SKILL.md
Purpose and capabilities are mostly aligned for an image-generation skill, but trust hinges on a third-party CLI and the skill explicitly permits arbitrary proxy endpoints. That makes credential and data routing the main concern: benign if pointed at official OpenAI endpoints, suspicious if used with custom gateways.
Confidence: 84%Severity: 58%
Audit Metadata