blip-2-vision-language
Pass
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No security issues detected. The skill provides legitimate documentation and code samples for image captioning and visual question answering using established libraries like HuggingFace Transformers and Salesforce LAVIS. All dependencies and model references target reputable sources. While the skill defines workflows that process external images and user-provided text, which is an inherent surface for indirect prompt injection in multimodal models, it does not implement any dangerous capability chains that would allow such inputs to compromise the system.
Audit Metadata