edit-image-nano-banana

Pass

Audited by Gen Agent Trust Hub on Aug 19, 2026

Risk Level: SAFEDATA_EXFILTRATIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [DATA_EXFILTRATION]: The skill transmits local image files to Fal.ai storage infrastructure to enable cloud-based image processing. This operation is the primary function of the skill and requires the user to provide a FAL_KEY via environment variables for authentication with the service.\n- [EXTERNAL_DOWNLOADS]: The skill retrieves processed image files from Fal.ai URLs generated by the API response and saves them to the local output directory.\n- [INDIRECT_PROMPT_INJECTION]: The skill possesses an indirect prompt injection attack surface as it ingests and processes external image data which could contain malicious instructions designed to influence the model's output.\n
  • Ingestion points: Untrusted data enters the context through local image files or public HTTPS URLs provided in the image_paths parameter.\n
  • Boundary markers: The skill does not implement specific delimiters or instructions to the model to ignore potential instructions embedded within the source images.\n
  • Capability inventory: The skill has the ability to read local files, perform network requests to the Fal.ai API, and write files to the local system via scripts/edit.sh.\n
  • Sanitization: No validation or sanitization of the visual or textual content within the input images is performed before processing.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 19, 2026, 04:39 PM
Security Audit — agent-trust-hub — edit-image-nano-banana