vision
Pass
Audited by Gen Agent Trust Hub on Jun 13, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The tool facilitates image analysis by communicating with established AI providers including OpenAI, Alibaba Cloud (DashScope), and Volcengine (Doubao). All default endpoints are official service URLs.- [SAFE]: Credential management follows best practices by utilizing environment variables for API keys rather than hardcoding sensitive information.- [SAFE]: The skill's operation is transparent and aligns with its stated purpose of reading local image files and providing descriptions via vision models.- [SAFE]: No evidence of obfuscation, prompt injection, unauthorized persistence, or privilege escalation was found in the provided files.
Audit Metadata