gemini-nano-banana-2

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The prompt explicitly instructs passing API keys as a --api-key argument (and notes "use if user provided key in chat"), and shows examples embedding KEY in command lines, which requires the LLM to accept and include secret values verbatim — a high exfiltration risk.

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.90). The skill explicitly supports Google Web Search and Google Image Search grounding (see SKILL.md --google-search / --image-search flags and scripts/generate_image.py where search tools are added to the GenerateContentConfig), meaning it fetches and ingests open web/image results which the model uses to influence image generation and decisions, creating a clear avenue for indirect prompt injection from untrusted third‑party content.