What is the initial problem:
Using some "Simple Agent after new mail has arrived" - "reply to sender" action: The agent gets triggered (again), if the database is replicated (manually) on the same server.
And these agents are decentralized. It would be better to have one central single point, where the mailadresses , which should do a reply, would be deposited and the reply answer be managed.
Such a single central point could be a rule on the router for example.
Background and actual workaround:
Simple action agents are intended to run against new documents. By making a new replica, all the documents are seen as new so the agent runs against them again.
The way to work past that is to either make it an "Before new mail" arrives, then the router will run the code and reply as it gets deposited into the users mail file or to expand the agent, adding a flag to the documents that have been processed and have the code check and see if that flag is set and not send the email if it is.
But there is no workaround for centralizing all the reply-to agents to a single point, e.g. a router rule and to have them at a glance.
