Btw I found the same issue with the status node - it filters direct nodes from which it will ignore status updates but if I put a change in-between the status and the node producing the status update, I've got an endless loop.
I.e. this setup:
If I enable the change node and then trigger one of the injects, loop is created. Especially since the status is ever changing in the debug node 398 - it's a message counter, status each time now, status all sends out message, debug 398 counts that message, new status, status all captchas that status update and sends out a message, debug 398 counts that message, updates its status, status all captchas that update .... and I'm dizzy.
