vercel-ai-sdk

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

• Third-party content exposure detected (high risk: 0.80). The skill explicitly fetches and ingests content from external public sources — e.g., it instructs using WebFetch to check https://ai-sdk.dev/docs/, includes a "searchWeb" tool example that would retrieve web results, and shows MCP integration that calls httpClient.tools() to dynamically fetch tools from external MCP servers — so the agent will read/interprete third‑party public content as part of its workflow.