Add sensitive data detection flow in bot configuration

A new flow has been added to the chatbot's configuration for detecting sensitive data from the bot's output. This enables the bot to prevent replying when it detects any potentially sensitive data, thus enhancing the bot-user interaction security.

nemoguardrails_config/bot_flows.co

@@ -12,5 +12,12 @@ define flow self check output
     bot refuse to respond
     stop
 
+define flow detect sensitive data on output
+  $has_sensitive_data = execute detect_sensitive_data(source="output", text=$bot_message)
+
+  if $has_sensitive_data
+    bot refuse to respond
+    stop
+
 define bot refuse to respond
   "⚠️ I'm sorry, I can't respond to that. (This message was blocked by the guardrail)"