nano-banana

Pass

Audited by Gen Agent Trust Hub on Sep 23, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes user-provided text prompts for image generation, which is a standard surface for indirect prompt injection. The impact is limited to the content of the generated image.
  • Ingestion points: User-provided prompt strings in the JSON request body (SKILL.md).
  • Boundary markers: None explicitly defined; relies on external API safety filters.
  • Capability inventory: Shell command execution for network requests (curl) and file processing (jq, base64) in SKILL.md.
  • Sanitization: Not applicable at the skill level; relies on the destination API's content filtering.
  • [COMMAND_EXECUTION]: The skill provides bash commands for the agent to execute, including curl for network interaction and jq/base64 for handling JSON and image data. These are standard operations for the skill's functionality.
  • [EXTERNAL_DOWNLOADS]: The skill makes network requests to official Google Gemini API endpoints (generativelanguage.googleapis.com) to generate images.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 23, 2026, 05:06 PM
Security Audit — agent-trust-hub — nano-banana