langgraph-python-quickstart

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

• Third-party content exposure detected (high risk: 0.70). The required workflow fetches and follows live docs from a public web URL at runtime (HTTP/Docs MCP), so the LLM may ingest outsider-authored free text from that page.

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

• Potentially malicious external URL detected (high risk: 1.00). The skill explicitly instructs at runtime to "Fetch that page" and implement the docs from https://docs.langchain.com/oss/python/langgraph/quickstart, so external content at that URL would directly control the agent's implementation and behavior.