Page 1 of 1

Forward SPAM mails from spam assassin to another account with SAPM HEADER

Posted: 2019-05-24 07:44
by ajaykumar.k
Hi Guys,

I have recently implemented spam assassin in my organisation. It seems like working fine so far by tagging all suspicious mails as SPAM HEADER.
But my main intention is to forward all SPAM mails to separate account instead before reaching to mailbox of employees,
So, our system admin can monitor them and validate accordingly from that separate account .

Sorry for Type error/spelling mistakes & thanks in advance

Re: Forward SPAM mails from spam assassin to another account with SAPM HEADER

Posted: 2019-05-24 09:45
by mattg
catch with a global rule and forward to a new recipient, then delete the original

If subject contains 'Spam header'
AND
Recipient list not contain 'spam@example.com'

THEN
Forward to spam@example.com
Delete Message

Re: Forward SPAM mails from spam assassin to another account with SAPM HEADER

Posted: 2019-05-24 12:51
by palinka
ajaykumar.k wrote:
2019-05-24 07:44
Hi Guys,

I have recently implemented spam assassin in my organisation. It seems like working fine so far by tagging all suspicious mails as SPAM HEADER.
But my main intention is to forward all SPAM mails to separate account instead before reaching to mailbox of employees,
So, our system admin can monitor them and validate accordingly from that separate account .

Sorry for Type error/spelling mistakes & thanks in advance
A better strategy possibly. Set your spam delete threshold vey high, like 1000, then use rules to sort delete.

If X-hMailServer-Reason-Score > 7.99
Forward to spam@example.com
Delete

If X-hMailServer-Reason-Score < 8
Forward to spam@example.com

Then the user gets to decide grey area stuff AND admin gets to see and act on all incoming spam. This is what i do. This makes sure i can sort spam from ham. Even if the user ignores the spam (leaves it in the inbox) i can still make sure bayes picks it up. Make sure to change the subject so the user is notified on potential spam.

Once you build up a spam corpus for bayes and get your system running well, you can play with the number. I have my "rule delete threshold" set at 6. Since i can see all spam, I'm confident in the threshold.

Also read this incredibly informative thread: https://www.hmailserver.com/forum/viewt ... 20&t=26866