ai-sdk-core

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

• Third-party content exposure detected (high risk: 0.80). This skill includes explicit examples and features that ingest public web content (e.g., the v5 migration note and examples showing a search tool like googleSearchRetrieval() and "sources" support for web references), meaning the agent may fetch and read untrusted third‑party content as part of its workflow.