The vulnerability stems from the SMTP test function's error handling. The original implementation (pre-patch) in SMTP.ts line 50 simply re-threw raw errors ('throw e;'), potentially exposing internal server details in error responses. This allowed two attack vectors: 1) SSRF via controlled SMTP server configuration pointing to internal services, and 2) sensitive information disclosure through error messages containing server responses. The patch replaced this with generic error messages and logging, confirming this was the vulnerable code path. The CWEs (200, 209, 918) directly map to this function's behavior of exposing error details and enabling internal network requests.