nano-banana-builder
Pass
Audited by Gen Agent Trust Hub on Sep 23, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill facilitates the ingestion of user-provided prompts for image generation and refinement, creating a surface for indirect prompt injection. \n
- Ingestion points: Found in the
ImageGeneratorcomponent inSKILL.mdand thegenerateImageandeditImageserver actions inreferences/advanced-patterns.md. \n - Boundary markers: The provided code examples do not include explicit delimiters or instructions to ignore embedded commands. \n
- Capability inventory: The skill is limited to cloud storage writes (Vercel Blob, S3) and network requests to Upstash Redis; it does not request local file system access or shell command execution. \n
- Sanitization: Prompts are passed directly to the AI model without sanitization logic. \n
- [EXTERNAL_DOWNLOADS]: The skill includes instructions to download and install standard libraries from well-known, trusted organizations for AI SDKs and cloud services. \n
- Evidence: Instructions to install
@ai-sdk/google,@vercel/blob, and@aws-sdk/client-s3inSKILL.mdandreferences/configuration.md.
Audit Metadata