nano-banana-2
Pass
Audited by Gen Agent Trust Hub on Sep 2, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill references external components and installation instructions from GitHub repositories associated with the service provider (inference-sh and belt-sh). These tools are required for the skill's primary function of interacting with the Nano Banana 2 image generation model.
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted external data.
- Ingestion points: User-supplied prompt strings and external image URLs (up to 14) defined in the input schema within SKILL.md.
- Boundary markers: No explicit delimiters or instructions to ignore embedded commands are present in the provided prompt interpolation examples.
- Capability inventory: The skill utilizes the Bash tool to execute belt CLI commands, which includes network communication with the inference.sh API.
- Sanitization: The instructions do not define sanitization or validation logic for the content of the user-provided prompt or the data retrieved from external image URLs.
Audit Metadata