nano-banana-pro
Pass
Audited by Gen Agent Trust Hub on Mar 29, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: Fetches official libraries from standard package registries.
- Includes
google-genaiandpillowfor image processing and API communication. - [COMMAND_EXECUTION]: Executes image generation logic via a Python script.
- Uses
uv runto launch the scriptscripts/generate_image.pywhich performs the core logic. - [PROMPT_INJECTION]: The skill processes external data which serves as a surface for indirect instructions.
- Ingestion points: The script
scripts/generate_image.pyaccepts text prompts and local image files for processing. - Boundary markers: Absent; user prompts are passed directly to the model content list.
- Capability inventory: File system read/write access (for input/output images) and network access to Google API endpoints.
- Sanitization: None; the script passes input data directly to the API client library.
Audit Metadata