nano-banana-pro

Pass

Audited by Gen Agent Trust Hub on Mar 13, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: The script scripts/generate_image_boluobao.py attempts to retrieve an API key by executing the pass command (pass show api/boluobao) if the key is not found in environment variables. This is a common pattern for users of the standard Unix password manager.
  • [EXTERNAL_DOWNLOADS]: The fallback script scripts/generate_image_boluobao.py communicates with apipark.boluobao.ai to generate images and subsequently downloads the resulting image files to the local system.
  • [DATA_EXFILTRATION]: User-provided text prompts and, optionally, local input images are transmitted to external AI services (Google and Boluobao) to fulfill the image generation and editing requests.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied text and images which are passed directly to large models for processing.
  • Ingestion points: User prompts via --prompt and image files via --input-image in scripts/generate_image.py.
  • Boundary markers: The skill uses basic prompt templates described in SKILL.md but lacks strict technical delimiters in the execution scripts.
  • Capability inventory: Limited to network API calls for image generation and local file writing for saving the generated image.
  • Sanitization: No specific sanitization or filtering of the input prompt is performed before transmission to the API.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 13, 2026, 06:59 AM