I'm not seeing any documentation for AI guardrail sets "Stop threshold" configuration. I understand the idea behind it, but some documentation or additional help text would be welcome. After reviewing the code (with Claude's help), it seems that at the current time, the 3 plugins that are included can all ever just contribute 0 or 1 to the aggregate score. Therefore, _any_ value of "Stop Threshold" between 0 and 1 is equivalent. I also understand that other guardrail plugins can be developed that provide fractional scores. -mike