Hopefully Mdaemon will see merit in providing a tweaked version of Spam Assassin Rulesets for Mdaemon Customers on the basis that the default lets far too much through. Users can and should be able to add their own custom rules and I will do that for things like blocking new top level domains but I still feel that Mdaemon should introduce a more effective spam assassin ruleset as the defaults are just not adequate. Should I request this as a feature on the Mdaemon's Feature request website. It is clear that they customise Spam Assassin to move emails to spam if they contain malicious urls or have emails about Lots of Money and bitcoin (which currently scores 0.0 in Mdaemon's Spam Assassin). Looking at other Spam Prevent Companies it would appear that many of them use Spam Assassin and yet they block emails like the example highlighted above. There is no reason for such a low score.Ģ. Yes I could try Bayesian Learning Feature, but it does not negate my main point that Spam Assassin should be scoring high enough to put an email with a malicious malware link into the spam folder. This example alone demonstrates how ineffective Spam Assassin actually is. I suspect that Bayesian learning would not help here either. I cannot think of any legitimate reason for receiving an email with a malware url link in it, yet the spam assassin score is so low like 1 that we used to get emails like this regularly until I added this rule. Into the local.cf file for spam assassin to move any emails containing malware links into the spam folder. The default spam assassin rules are not adequate. I would however recommend that mdaemon consider creating a custom rule set for Mdaemon Clients for the following reasons:ġ. I will have a look at the Spam Assassin Custom Rulesets that you linked to and see anything here would help. I will see if I can use Bayesian where the only person that can add emails to this is myself, but I am not sure if this is workable as I only have about 2 dozen emails to put in over a period of two years, so I just don't have enough of a sample, unless I use the honeypot for this? Also never used the honeypot option as I never got around to it. I agree with them that treating Marketing and Spam emails differently makes sense as I don't want users to have a spam folder that they need to check for legitimate emails. Their recommendation for Marketing Emails is to unsubcribe to the emails which is something that I am working on with my clients as we do have a problem with marketing emails. I have always been advised by Zen Support in the UK to not use Bayesian as they claim it may end up causing more problems than it solves as users tend to fill the folder with marketing emails not actual spam, so I have never used it. We are using Mdaemon private cloud so have Outbreak, Spam Assassin, Cyren, Clam AV as well as all the Mdaemon settings like screenings, ptr record lookup etc. Is there a way to get this to work anywhere else? Or I am doing this wrong.īrad (MDaemon) Thank you for clarifying that the languages does not work in the local.cf file. I think the header has to explicitly state the language. The language one doesn't work as neither Spam Assassin nor Mdaemon seem to be able to autodetect languages. But this is what Mdaemon use for their spam filtering in addition to Outbreak Protection (which works brilliantly). I have never been happy with Spam Assassin because it doesn't seem to have any ability to read content of email and block on this basis.
We don't get a lot of spam as most of it gets dropped by Outbreak Protection, but Spam Assassin doesn't really seem to work very well as it very rarely blocks any email and when it does its usually because it has failed a spf check or something not because of content within the email. I have just added the top level domain rule yesterday. I have manually added the following to the local.cf rules file. Can anyone recommend where I can get source some good Spam Assassin Rules as the default configuration for Spam Assassin is simply not adequate.