[Rspamd-Users] How to reject email subject/body message?

[Stanislav]

Hi,

I'm more or less familiar with mail server, and, overall, how it works. 
I do not consider myself a guru tho, haha. I'm pretty much hosting a 
mailbox for myself only at this point; everything was configured by me 
from scratch a long time ago. I used to have just Postfix as mta using 
built-in basic checks and rbls to reject spam. I've recently started 
noticing legitimate emails getting rejected, so I'm experimenting with 
this now.
I didn't use SpamAssassin or any other similar solution in the past. 
Pretty much Rspamd is the first one I've decided to try.

What I'm expecting to achieve is the following:
  - Rspamd renders a message, and based on some key patterns or text like 
"you won million dollar" rejects the email.

I know it's inefficient, but a server with a "single" mailbox works just 
fine for me. Currently, most of the spam Rspamd can handle, but I'm 
still getting some of it, even tho I did configure Bayes training with 
Dovecot/Sieve.

Best

> Hi there,
> 
> On Fri, 20 Oct 2023, Steve Witten wrote:
>> On Fri, Oct 20, 2023 at 11:48 AM Stanislav via Users wrote:
>>> 
>>> I'm relatively new to spam filtering systems.
>>> Could somebody please point me to the documentation or give an 
>>> example
>>> of how to perform the following actions:
>>> - Reject an email with specific text in a body message.
>>> - Reject an email with specific text in a subject.
>> 
>> See:
>> 
>> https://rspamd.com/doc/modules/regexp.html
> 
> If I may make an analogy, if the OP had asked someone to teach him how
> to swim then Mr. Witten's answer, although accurate, somewhat resembles
> throwing the questioner from the ten metre board over the diving pool.
> 
> Stanislav: can you tell us how familiar you are with the structure of
> 
> (1) emails in general and
> 
> (2) MIME-structured emails in particular?
> 
> It's especially important to know what you're dealing with before you
> try to filter mail because otherwise the results may be disappointing.
> 
> --
> 
> 73,
> Ged.