effect-ai-streaming

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

• Third-party content exposure detected (high risk: 0.80). The skill's required workflow (e.g., the "Complete Example" and the "Source Parts" section) shows LanguageModel.streamText producing Response.StreamPart entries like "url-source" and "document-source" which are parsed via Prompt.fromResponseParts and merged into conversation history, meaning untrusted/public third‑party content (URLs/documents) is ingested and can directly influence subsequent prompts and actions.