agent-architecture

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

• Third-party content exposure detected (high risk: 1.00). The skill explicitly defines and uses a fetch_url tool (see "Step 2: Tool-Using Agent" and the LangGraph example) that fetches arbitrary web URLs and injects the retrieved page text into the agent's message loop, so untrusted public web content can be read and materially influence agent actions.