nano-banana-prompt-builder
Pass
Audited by Gen Agent Trust Hub on Oct 16, 2026
Risk Level: SAFE
Full Analysis
- [COMMAND_EXECUTION]: The
INSTALL.mdfile includes shell commands (mkdir,cp,chmod) intended as manual installation instructions for the user. These commands are limited to local directory management and do not involve remote execution or automated agent activity. - [INDIRECT_PROMPT_INJECTION]: The skill ingests user-provided scene descriptions to interpolate into prompt templates. While this constitutes an input surface, the skill lacks the capabilities (such as network access or file system writes) required for an injection to be exploited, rendering the risk negligible.
- [SAFE]: The skill's behavior is consistent with its stated purpose as a creative tool for prompt engineering. No evidence of obfuscation, credential harvesting, or unauthorized data access was found across the 16 analyzed files.
Audit Metadata