hi there,

I've just had an idea and I thought i'd look into it.

I'm thinking about setting up a user called 'spamlist' now when someone sends a e-mail to spamlist@domain.com it executes a command, the command gets out the header information from the e-mail, ip address it came from or other information and then put it on a blacklist that procmail, or postfix, or program of some kinda, i'm not sure what one will do this then forwards it into a mail box and all the e-mails get stored there for a month and a cron job deletes it once a month.

So users can define what e-mails are spam and what are not and if a mistake is made the e-mail will still be stored and the clients can contact the IT support staff to recover the e-mail.

questions:

has another program already done this?
Can someone give me some information on executing commands via e-mails?
Can anyone see any problems not listed below?

problems:
the head information will have information from the user the sent the e-mail to spamlist@domain.com not who sent it to the user (i'm sure that the header information from the orginal sender will still be there, we'll just have to set the program to dismiss the infomration from @domain.com)

Any thoughts or help would be great

The biggest problem - ALL the header information in spam (including first three Received-From: ) can be forged. That's why it won't work well.

how do anti spam agents determine whats spam and whats real mail then?

Well, by inconsistency of headers (from the viewpoint of RFC), and by content analysis. For example, "Buy now and get 20% more for free" is somewhat suspicious..