Skills
Navigate
skills/inference-sh/skills/nano-banana/Gen Agent Trust Hub

nano-banana

Pass

Audited by Gen Agent Trust Hub on Sep 19, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses the belt CLI tool to interact with image generation models. The execution is restricted to the belt command via the allowed-tools configuration.
  • [EXTERNAL_DOWNLOADS]: The documentation references the installation of the belt CLI and additional agent skills from the vendor's official GitHub repositories and NPM registry.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted input which could be manipulated to influence the behavior of the underlying image generation models.
  • Ingestion points: The prompt string and images array (containing external URLs) within the JSON payload for the belt app run command in SKILL.md.
  • Boundary markers: The input is structured within a JSON object, providing basic delimitation.
  • Capability inventory: The skill invokes the belt command to process data and generate/edit images through remote APIs.
  • Sanitization: No specific sanitization or filtering of the prompt or image source URLs is described in the skill instructions.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 19, 2026, 10:23 AM
Security Audit — agent-trust-hub — nano-banana