nano-banana-2
Pass
Audited by Gen Agent Trust Hub on Mar 12, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses the
infshCLI tool, a vendor-specific utility, to handle model execution and account login.- [EXTERNAL_DOWNLOADS]: The skill documentation references theinferenceshPython SDK and provides instructions for installing vendor components vianpxfrom theinference-sh/skillsrepository.- [PROMPT_INJECTION]: The skill exhibits vulnerability to indirect prompt injection due to the ingestion of external and untrusted data. - Ingestion points: User input through the
promptfield and web-retrieved information whenenable_google_searchis used. - Boundary markers: No explicit markers or specific instructions are provided to help the model distinguish between control instructions and data.
- Capability inventory: The skill is granted access to the
Bash(infsh *)capability for interacting with the generation platform. - Sanitization: There is no evidence of input validation or sanitization before processing strings provided to the underlying model.
Audit Metadata