firebase-ai-logic-basics

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

• Third-party content exposure detected (medium risk: 0.30). Skill.md and its referenced client guides describe a direct LLM runtime workflow where prompts entered/used at the app level are passed to Firebase AI Logic via generateContent/sendMessage/generateContentStream, meaning outsider-authored free text can be ingested when an app forwards user input as the prompt.