langchain

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

• Third-party content exposure detected (high risk: 1.00). The skill's instructions (Step 4: "Document Loaders and Text Splitters") explicitly include WebBaseLoader and UnstructuredHTMLLoader for ingesting web pages, and the RAG/chain examples show the agent retrieving and using that fetched content as context for answers, so arbitrary public web content could be ingested and influence agent behavior.