nanobanana
Warn
Audited by Socket on Apr 23, 2026
1 alert found:
AnomalyAnomalySKILL.md
LOWAnomalyLOW
SKILL.md
SUSPICIOUS. The core image-generation purpose matches the skill, and the official Gemini endpoint/model references are coherent. The main risk is the custom gateway design: it allows prompts, images, and API credentials to be sent to arbitrary non-Google endpoints, with auto auth potentially forwarding both bearer and API-key credentials. That is a meaningful data-flow and credential-forwarding concern, though not confirmed malware.
Confidence: 89%Severity: 63%
Audit Metadata