nano-banana-2

Pass

Audited by Gen Agent Trust Hub on May 28, 2026

Risk Level: SAFENO_CODEPROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • [SAFE]: The skill instructions do not contain malicious patterns or security vulnerabilities.
  • [NO_CODE]: The skill is composed of markdown text and does not provide any internal scripts or binary files.
  • [PROMPT_INJECTION]: The skill includes an interface for passing user-generated prompts to an external API. 1. Ingestion points: 'prompt' field in the 'belt' command and Python SDK. 2. Boundary markers: No delimiters are shown in the provided examples. 3. Capability inventory: Interactions with the 'belt' tool to perform remote image generation. 4. Sanitization: Input sanitization is not mentioned in the provided text.
  • [EXTERNAL_DOWNLOADS]: References installation instructions and documentation from the vendor's public GitHub repository.
  • [COMMAND_EXECUTION]: Uses the 'belt' CLI tool for interacting with the service, which is a permitted tool for this skill.
Audit Metadata
Risk Level
SAFE
Analyzed
May 28, 2026, 05:54 PM
Security Audit — agent-trust-hub — nano-banana-2