gpt-image

Warn

Audited by Socket on Aug 12, 2026

1 alert found:

Anomaly
AnomalyLOW
SKILL.md

SUSPICIOUS: the skill’s purpose is coherent, but it expands trust through a transitive skill install and routes prompts, images, and login data through inference.sh rather than direct OpenAI APIs. This looks more like a medium-risk third-party gateway/integration than confirmed malware.

Confidence: 84%Severity: 62%
Audit Metadata
Analyzed At
Aug 12, 2026, 10:39 PM
Package URL
pkg:socket/skills-sh/halt-catch-fire%2Fsuperpowers%2Fgpt-image%2F@2ccae34d2a82b5e258d0fbf112db101927eae820ae364b695cfd3ea9a690fe06
Security Audit — socket — gpt-image