nano-banana-prompting
Pass
Audited by Gen Agent Trust Hub on Nov 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from users, such as existing prompts and answers to clarification questions, which are then interpolated into a final prompt for the nano-banana image generation tool.
- Ingestion points: Step 1 (Gather Reference Materials) and Step 2 (Clarify Intent with Questions) in SKILL.md.
- Boundary markers: Absent. The instructions do not define clear delimiters or specific instructions for the agent to ignore hidden commands within user-provided prompts.
- Capability inventory: The skill uses the AskUserQuestion tool and invokes an external tool/skill named nano-banana.
- Sanitization: Absent. There is no mention of filtering, escaping, or validating the user-provided text before it is used to construct the final prompt for the image generator.
Audit Metadata