nano-banana-prompt

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

• Third-party content exposure detected (high risk: 0.90). This skill explicitly instructs the agent to "research" external information (e.g., "Novel/Film Scene 3D Poster: Research: First retrieve information about this work..." and "Coordinate Visualization: 1. Identify the location from coordinates 2. Research what exists at these coordinates"), which requires fetching open/public third-party content and therefore exposes the agent to untrusted third-party content.