nano-banana-2

Pass

Audited by Gen Agent Trust Hub on Jul 13, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses the runcomfy CLI tool to interact with the Google Nano Banana 2 model. User inputs are encapsulated in a JSON payload during invocation, which is a documented strategy to mitigate shell injection risks.
  • [CREDENTIALS_UNSAFE]: Authentication is handled via API tokens stored in a restricted configuration file (~/.config/runcomfy/token.json with 0600 permissions) or through an environment variable (RUNCOMFY_TOKEN), adhering to industry standards for secure secret management in CLI tools.
  • [EXTERNAL_DOWNLOADS]: The skill relies on the @runcomfy/cli package, which is the official utility for the RunComfy service. This is a well-known service dependency necessary for the skill's primary function.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 13, 2026, 04:10 AM
Security Audit — agent-trust-hub — nano-banana-2