The hour is serious, reflection ...

MisterM

MisterM

Verified User
Joined
Jul 31, 2022
Messages
350
Hello

I know, my subject can be translated differently, but, it's like crazy, frankly.

I just received a fake message from Orange to unsubscribe.

I use rspamd (@zEitEr) is it pass with success, hey yes, even worse, I make a test of scanning on Tinycp (https://tinycp.com/page/show/rbl-check) is his ip of sending is not recognized in the 103 rbl, not badly known in the list.

What to do, ha yes some will say to me, there are fuzzy lists, they are add, others paying that I would not quote.

Come back in spamassassin, but yes, it also passes, believe me, what to do

Mz
 
It can always happen that some spam does pass and is not in any list yet.
Get the headers and report them to spamcop.

You can always get the sender where it's coming from.
You can block that ip in for example your firewall or blacklist.

Spammers always wander around, get other vps systems or something with other ip's and thus are not present in RBL's (yet).
 
MisterM said:
How do I do it?
Click to expand...
That depends on the mail client you're using.

As for the final spam score, that's high. But I as far as I know rspamd is working a bit different than Spamassassin. And Spamassassin normally had marked anything higher as 5.5 of 7.5 as spam anyway.
 
@Richard G question ,
Is the Exim version to a crucial importance, knowing that for the installation, they use the version 4.5 is not the latest?

Because I have just been alerted again that a blatant spam is still going on, and the score is at its highest.
SpamTally: Final spam score: 20
Click to expand...
Mz
 
Both Spamassassin and rSpamd work better after you train them. So you might feed them with 200+ SPAM and 200+ HAM emails. And even after that SPAM might pass through. As spammers are growing too, they use new technics to send SPAM.
 
@zEitEr so for you, you have to let it work ?
How do you manage the "Bayesian statistics" part in rspamd ?
Thank you for your precious help
Mz ;)
 

Attachments

  • 2022-11-04 21.38.10.png
    2022-11-04 21.38.10.png
    88.1 KB · Views: 104
That's all default from the installation. Nothing changed. If your page looks differently, then it might be a subject for an investigation. And posting its screenshot here might be a good beginning.
 
As of BAYES SPAM and HAM, those are numbers of learned emails. You should train rSpamd. You might use the webinterface of the plugin for it.
 
I will add one thing, it is a pity, and I saw it with mailcow, that when a message passed without being well filtered or that we were tired to see it in his mailbox or we forgot to unsubscribe, we can send it in the junk folder and send a signal that told rspamd, because it is him who manages the antispam in mailcow, to send data of the message in the other data of the thousands of messages sent.

Just an idea to advance this kind of "crap" that we receive every day and that disturbs people who are not used to this kind of message.
 
Great @zEitEr

One thing in your setting file.

Folder "INBOX.teach-isspam"; and "INBOX.teach-isnotspam"; can be , ATTENTION , a question , "INBOX.spam" and "INBOX(Reception)".
 
The back-end script is running automatically with cron, since that I would not suggest using INBOX for teaching HAM. You might do it the way you described, but you should be ready that SPAM messages from INBOX might be learned as HAM.
 
You must log in or register to reply here.
Back
Top