ai-langchain-patterns

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

• Third-party content exposure detected (high risk: 0.72). The skill’s runtime workflow can ingest outsider-authored free text via retrieval/tooling paths (e.g., web_search_retriever/WebBaseLoader fetching public pages or slack_read_channel/slack_search returning other users’ messages), and that retrieved page_content/message text is then fed into the LLM as {context} in the LCEL/RAG pipeline.

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

• Potentially malicious external URL detected (high risk: 0.90). The DocumentEnricher component calls an external API (https://enrich.internal/api) at runtime and injects the returned "enrichment" into the LLM context, so content fetched from that URL can directly control prompts.