image-generation

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

• Third-party content exposure detected (high risk: 0.80). The skill supports Gemini "search grounding" (see gemini.sh --search-grounding flag in SKILL.md and the tools: [{"google_search": {}}] example in references/api-details.md), which pulls open web search/image results (untrusted third‑party content) into the model's generation flow and can materially influence outputs and subsequent behavior.