Re: base64 encoded subjects

2020-02-07 Thread hamann . w
>> I'm seeing a lot of spam with base64 encoded subjects:
>> 
>> Subject:
>> =?UTF-8?B?RnVsbCBkZW50YWwgY292ZXJhZ2UgZm9yIGZhbWlsaWVzIGFuZCBzZW5pb3JzLCBjb3ZlcnMgYWxsIHByb2NlZHVyZXM=?=
>> 
>> Subject: =?UTF-8?B?V2VhciB5b3VyIE11bHRpLVRvb2wgYXJvdW5kIHlvdXIgd3Jpc3Qu?=
>> 
>> 
Hi,

I live in a part of  the world where =?UTF-8?B? is normal  everywhere 
utside the
realm of US Ascii
I would, however, treat UTF16 as a spam indicator

Best regards
Wolfgang



Re: base64 encoded subjects

2020-02-07 Thread Bill Cole
On 7 Feb 2020, at 21:16, Benjamin Toll wrote:

> On 2020-02-07 18:07, RW wrote:
>>
>> There's already a sub-rule for this
>>
>> __SUBJECT_ENCODED_B64Subject:raw =~ /=\?\S+\?B\?/i
>
> Oh. Thank you very much.
>
> It looks like I haven't been reading enough of the docs after all.

If you use SA, "perldoc Mail::SpamAssassin::Conf" is your best friend.

(Or it's mine at least, but I'm running low on surviving friends.)

-- 
Bill Cole
b...@scconsult.com or billc...@apache.org
(AKA @grumpybozo and many *@billmail.scconsult.com addresses)
Not For Hire (currently)


Re: base64 encoded subjects

2020-02-07 Thread Benjamin Toll

On 2020-02-07 18:07, RW wrote:


There's already a sub-rule for this

__SUBJECT_ENCODED_B64   Subject:raw =~ /=\?\S+\?B\?/i


Oh. Thank you very much.

It looks like I haven't been reading enough of the docs after all.



Re: base64 encoded subjects

2020-02-07 Thread RW
On Fri, 7 Feb 2020 16:29:38 -0600
Benjamin Toll wrote:

> I'm seeing a lot of spam with base64 encoded subjects:
> 
...
> SA is scoring the messages pretty high based off the body, but a lot
> of spams with the base64 is still getting through. I thought it
> wouldn't be too hard to write a rule to catch these, but clearly I
> don't know what I'm doing:
> 
> header   BRT_BASE64_SUBJECT Subject =~ /=\?UTF\-8/
> 



There's already a sub-rule for this 

__SUBJECT_ENCODED_B64   Subject:raw =~ /=\?\S+\?B\?/i


so 

meta BRT_BASE64_SUBJECT __SUBJECT_ENCODED_B64

would do. Note that it's the ?B? part that makes it base64. 


Re: base64 encoded subjects

2020-02-07 Thread Alan Hodgson
On Fri, 2020-02-07 at 16:29 -0600, Benjamin Toll wrote:
> I'm seeing a lot of spam with base64 encoded subjects:
> 
> Subject:
> =?UTF-8?B?RnVsbCBkZW50YWwgY292ZXJhZ2UgZm9yIGZhbWlsaWVzIGFuZCBzZW5pb3JzLCBjb3ZlcnMgYWxsIHByb2NlZHVyZXM=?=
> 
> Subject: =?UTF-8?B?V2VhciB5b3VyIE11bHRpLVRvb2wgYXJvdW5kIHlvdXIgd3Jpc3Qu?=
> 
> 
> SA is scoring the messages pretty high based off the body, but a lot of
> spams with the base64 is still getting through. I thought it wouldn't be
> too hard to write a rule to catch these, but clearly I don't know what
> I'm doing:
> 
> header   BRT_BASE64_SUBJECT Subject =~ /=\?UTF\-8/
> 
> This doesn't trigger on any of these spams. Am I going about this the
> wrong way or I'm I just that bad at writing regexs? Any suggestions
> would be appreciated.
> 

SA decodes those before rule matching.

Try Subject:raw =~ 


base64 encoded subjects

2020-02-07 Thread Benjamin Toll
I'm seeing a lot of spam with base64 encoded subjects:

Subject:
=?UTF-8?B?RnVsbCBkZW50YWwgY292ZXJhZ2UgZm9yIGZhbWlsaWVzIGFuZCBzZW5pb3JzLCBjb3ZlcnMgYWxsIHByb2NlZHVyZXM=?=

Subject: =?UTF-8?B?V2VhciB5b3VyIE11bHRpLVRvb2wgYXJvdW5kIHlvdXIgd3Jpc3Qu?=


SA is scoring the messages pretty high based off the body, but a lot of
spams with the base64 is still getting through. I thought it wouldn't be
too hard to write a rule to catch these, but clearly I don't know what
I'm doing:

header   BRT_BASE64_SUBJECT Subject =~ /=\?UTF\-8/

This doesn't trigger on any of these spams. Am I going about this the
wrong way or I'm I just that bad at writing regexs? Any suggestions
would be appreciated.