nano-banana-character-consistency

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

• Third-party content exposure detected (medium risk: 0.30). SKILL.md + scripts/imagegen.py: outsider-authored free text flows directly from the user-provided --prompt (and optional --system) into the request body for POST /openapi/v1/chat/text or POST /openapi/v1/generation/image without any selection or filtering of specific items.