neon-functions

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

• Third-party content exposure detected (medium risk: 0.30). Neon Functions are exposed via a public HTTP/WebSocket/SSE runtime handler whose request body/streamed input (e.g., the messages JSON posted to fetch) is read by the agent at runtime, meaning outsider-authored free text can reach the LLM through direct caller POSTs.