image-gen
Pass
Audited by Gen Agent Trust Hub on Oct 18, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process user-provided text prompts and batch manifests to generate images, creating a surface for potential instruction injection in external data.\n
- Ingestion points: Command-line prompt arguments, text-based batch files, and JSON manifest files processed by
generate_image.pyandnano-banana-generate.py.\n - Boundary markers: The skill does not implement specific delimiters to separate user prompts from system instructions, as the prompts are the primary intended input.\n
- Capability inventory: The skill has the capability to perform network requests to Google's Gemini API and write generated image files to the local file system.\n
- Sanitization: Input prompts and manifest data are processed without specific sanitization or filtering.\n- [EXTERNAL_DOWNLOADS]: The skill identifies necessary dependencies for its operation.\n
- Documentation and error messages reference the installation of
google-genaiandpillowfrom standard package registries. These are well-known, trusted libraries required for Gemini API interaction and image processing.
Audit Metadata