version: 1.1.0-SNAPSHOT SafeGuardAdvisor lacks any logs It would be nice if the log informs which amongst the specific sensitiveWords causes the failure. I guess it would be essential for production logs. At least there should be traces that it acted upon the request and response.
Comment From: sunyuhan1998
That's absolutely right. I agree that adding logs to indicate which sensitive words were matched makes sense. I've tried to push a PR to support this: https://github.com/spring-projects/spring-ai/pull/3589
Comment From: ilayaperumalg
@AmritSDutta Thanks for reporting. In this case, you need a custom advisor to log anything useful in the context of the use case. We definitely don't want to facilitate the sensitive information being available in the default implementation of the SafeGuardAdvisor.