C-Command Software Forum

Facebook Account spam getting through

I’m getting tons of “Facebook Account (blah blah)” messages from scammers trying to get me to click and provide my password and all. I’ve clicked on several of these to mark them as spam, but they keep getting through. Near as I can tell, the scammers are sending the spams from Facebook itself or fooling a mailer to do so since they come from @facebookmail.com – the same address good notifications come from.

Here is an entry from the log:

Predicted: Good (26)
Subject: Facebook Account Update
From: update+igbyovabkaur@facebookmail.com
Identifier: YKgXqDCyjhmx7Ae8P5Rfgg==
Reason: P(spam)=0.188[0.471], bias=0.000, x-mimeole:ProducedByMicrosoftMimeOLEV6.00.3790.1830(0.995), R:^138(0.995), U:pinaoeeiili(0.995), S:Facebook(0.040), F:Facebook(0.070), U:facebook(0.076), U:facebook(0.076), login(0.080), login(0.080), MT:Windows-1252(0.870), guide(0.141), guide(0.141), ^a-style-marginbottom15px(0.141), ^a-style-marginbottom15px(0.141), R:^77(0.843)
Date: 2009-12-26 12:39:03 -0600

Is there any way to get Spamsieve to kill these but not messages from family and friends? I’ve got a second mail rule running after spamsieve’s rules to send them to trash, but that only works when the subject is the same. I could turn off all facebook notifications, but I don’t check it a lot and family uses it. Ugh. I suppose I could email Facebook support and tell them to not allow usernames with “update” in them :slight_smile: (scammers would change the names, I guess… never ends…)

Thoughts and ideas?

Did you train them all as spam?

The e-mail address does not prove that. I suggest that you send in a report.

Generally speaking, that will happen automatically if you train the spammy ones as spam and the good ones as good.

I don’t recommend that kind of rule because it makes it likely that you’ll forget to train the ones that get through SpamSieve as spam. Plus, if you want to block messages by subject, SpamSieve has a feature for that.

Thanks for the reply, Michael. I will begin training all spammy ones as spam. I was worried I’d blow my ratio of good/bad with the numbers of these coming in. :slight_smile:

I’ll also remove the rule from Mail and look into setting it up in Spamsieve’s feature. Thanks!

You must always correct all the mistakes or the accuracy will deteriorate. If you have not been doing this, you should reset the corpus and start over.