LangChain RAG Pipeline

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

• Third-party content exposure detected (high risk: 1.00). The skill includes explicit web loaders (WebBaseLoader/CheerioWebBaseLoader fetching pages like "https://docs.langchain.com" in the "Document Loaders" examples) and shows those retrieved documents being fed into the model/agent (see "ex-rag-with-agent" where retriever results are joined into prompts), which means untrusted third-party web content can be ingested and materially influence agent behavior.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The skill's runtime examples show loading external web pages (e.g., WebBaseLoader / CheerioWebBaseLoader fetching https://docs.langchain.com) and injecting the fetched documents into the model context/prompts, so that URL can directly control agent instructions at runtime.