vercel-ai-sdk

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

• Third-party content exposure detected (high risk: 0.90). The SDK explicitly supports server-side tools that "search the web" and call external fetch functions (references/tools.md and references/streaming.md) and includes SourceUrlUIPart/FileUIPart URL parts and streamed tool-output chunks that are incorporated into the model's message stream, so arbitrary public web content can be ingested and interpreted by the agent.