firebase-ai-logic-basics

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

• Third-party content exposure detected (high risk: 0.80). SKILL.md explicitly references "Search Grounding with the built in googleSearch tool" (public web search results) and instructs passing external file URLs to the Gemini API in the Multimodal section, so the agent will ingest and act on untrusted third‑party web/user-generated content that can influence its behavior.