Re: New spamming trick?

2014-10-10 Thread Axb

On 10/10/2014 01:46 PM, Martin Gregorie wrote:

I've recently noticed what may be a new spamming technique: sending mail
to Yahoo Groups with an invalid group name - since Yahoo! doesnt! seem!
to! use! SPF, this intentional backscatter gets delivered to the forged
recipient address with the payload in the returned message text.

There are two ways of recognising it:

- the List-id: header is set to UnknownList.yahoogroups.com
- the user part of the To address is alphanumeric soup



pls pastebin a sample



Re: New spamming trick?

2014-10-10 Thread RW
On Fri, 10 Oct 2014 12:46:50 +0100
Martin Gregorie wrote:

 I've recently noticed what may be a new spamming technique: sending
 mail to Yahoo Groups with an invalid group name - since Yahoo!
 doesnt! seem! to! use! SPF, this intentional backscatter gets
 delivered to the forged recipient address with the payload in the
 returned message text. 
 
 There are two ways of recognising it:
 
 - the List-id: header is set to UnknownList.yahoogroups.com

I had 

List-Id: UnknownList.yahoogroupes.fr

Note the e in groupes - probably the first-part, UnknownList.yahoo,
would be consistent.


Re: New spamming trick?

2014-10-10 Thread Benny Pedersen

On October 10, 2014 1:46:50 PM Martin Gregorie mar...@gregorie.org wrote:


- the List-id: header is set to UnknownList.yahoogroups.com
- the user part of the To address is alphanumeric soup


Did yahoo dkim sign it ?

List sender domain as blacklist_from then, or maybe its even blacklist_to 
*@yahoogroups ?


Re: New spamming trick?

2014-10-10 Thread Martin Gregorie
On Fri, 2014-10-10 at 14:26 +0200, Axb wrote:
 On 10/10/2014 01:46 PM, Martin Gregorie wrote:
  I've recently noticed what may be a new spamming technique: sending mail
  to Yahoo Groups with an invalid group name - since Yahoo! doesnt! seem!
  to! use! SPF, this intentional backscatter gets delivered to the forged
  recipient address with the payload in the returned message text.
 
  There are two ways of recognising it:
 
  - the List-id: header is set to UnknownList.yahoogroups.com
  - the user part of the To address is alphanumeric soup
 
 
 pls pastebin a sample
 
 
Here you go:  http://pastebin.com/aqhcTZxH

I've replaced my address is these by example.com or example.isp.com but
the message is otherwise unchanged.

RW: you're right (just had another from Yahoo UK - I'm about to change
the rule to match UnknownList.yahoo 

Benny: Yes they did - after all, how can they tell a bouncing message
due to a fatfingered address from one that was crafted to bounce?

The examples I've seen so far have apparently been equity pumping scams.
Is this also a common feature?


Martin




Re: New spamming trick?

2014-10-10 Thread jdebert
On Fri, 10 Oct 2014 12:46:50 +0100
Martin Gregorie mar...@gregorie.org wrote:

 I've recently noticed what may be a new spamming technique: sending
 mail to Yahoo Groups with an invalid group name - since Yahoo!
 doesnt! seem! to! use! SPF, this intentional backscatter gets
 delivered to the forged recipient address with the payload in the
 returned message text. 

This is actually quite old. The only differences are what you describe
later.

Another old trick is to send to moderated groups as non-members and
have the group moderators reject the messages. 

Yahoo hasn't yet figured out how to not bounce such messages, it seems.




Re: New spamming trick?

2014-10-10 Thread David Jones
 On Fri, 10 Oct 2014 12:46:50 +0100
 Martin Gregorie mar...@gregorie.org wrote:

  I've recently noticed what may be a new spamming technique: sending
  mail to Yahoo Groups with an invalid group name - since Yahoo!
  doesnt! seem! to! use! SPF, this intentional backscatter gets
  delivered to the forged recipient address with the payload in the
  returned message text.

 This is actually quite old. The only differences are what you describe
 later.

 Another old trick is to send to moderated groups as non-members and
 have the group moderators reject the messages.

 Yahoo hasn't yet figured out how to not bounce such messages, it seems.

Yep.  Regular backscatter that my servers block.  You need something that
can detect and block backscatter.  MailScanner does this and an excellent
prebuilt VM to check out is http://efa-project.org/.  I have only seen one
commercial product do backscatter detection.  There may be others but I
have been using MailScanner for so long that I never needed to look for
other solutions.


Re: New spamming trick?

2014-10-10 Thread Axb

On 10/10/2014 06:59 PM, Martin Gregorie wrote:

On Fri, 2014-10-10 at 14:26 +0200, Axb wrote:

On 10/10/2014 01:46 PM, Martin Gregorie wrote:

I've recently noticed what may be a new spamming technique: sending mail
to Yahoo Groups with an invalid group name - since Yahoo! doesnt! seem!
to! use! SPF, this intentional backscatter gets delivered to the forged
recipient address with the payload in the returned message text.

There are two ways of recognising it:

- the List-id: header is set to UnknownList.yahoogroups.com
- the user part of the To address is alphanumeric soup



pls pastebin a sample



Here you go:  http://pastebin.com/aqhcTZxH

I've replaced my address is these by example.com or example.isp.com but
the message is otherwise unchanged.

RW: you're right (just had another from Yahoo UK - I'm about to change
the rule to match UnknownList.yahoo

Benny: Yes they did - after all, how can they tell a bouncing message
due to a fatfingered address from one that was crafted to bounce?

The examples I've seen so far have apparently been equity pumping scams.
Is this also a common feature?


Thanks for the sample...

Was wondering why I didn't see any

had an ancient Postfix header_check regex rule

/^X-Yahoo-Newman-Property: groups-bounce/   REJECT

 (I have no use for Yahoogroups mail)



Re: New spamming trick?

2014-10-10 Thread Martin Gregorie
On Fri, 2014-10-10 at 20:17 +0200, Axb wrote:
 Thanks for the sample...
 
 Was wondering why I didn't see any
 
 had an ancient Postfix header_check regex rule
 
 /^X-Yahoo-Newman-Property: groups-bounce/ REJECT
 
Does this only appear in Yahoo groups bounce messages? If so , I'll add
it to the rule and/or replace my current List-id name match.

I searched for information but only found people saying 'I dunno' only
with more verbosity. Apparently Yahoo doesn't publish and descriptions
for these headers.

   (I have no use for Yahoogroups mail)
 
Same here and extend it to include Google Groups too.


Martin






Re: New spamming trick?

2014-10-10 Thread Benny Pedersen

On October 10, 2014 6:59:40 PM Martin Gregorie

Benny: Yes they did - after all, how can they tell a bouncing message
due to a fatfingered address from one that was crafted to bounce?


the mailerdaemon is dkim signed, the attached msg is not signed, so its not 
sent from yahoo imho



The examples I've seen so far have apparently been equity pumping scams.
Is this also a common feature?


Ahh note the isp send you a dsn back for undelivered, here the isp is 
really yahoo, hopefully i am right, anyway its yahoo spam, block the url in 
bounce msg attachment with clamav


Re: New spamming trick?

2014-10-10 Thread Axb

On 10/10/2014 08:39 PM, Martin Gregorie wrote:

On Fri, 2014-10-10 at 20:17 +0200, Axb wrote:

Thanks for the sample...

Was wondering why I didn't see any

had an ancient Postfix header_check regex rule

/^X-Yahoo-Newman-Property: groups-bounce/   REJECT


Does this only appear in Yahoo groups bounce messages? If so , I'll add
it to the rule and/or replace my current List-id name match.


honestly, I couldn't sign that - my rule dates back to 2006 and I've 
never had a complaint - it's a works for me



I searched for information but only found people saying 'I dunno' only
with more verbosity. Apparently Yahoo doesn't publish and descriptions
for these headers.


   (I have no use for Yahoogroups mail)


Same here and extend it to include Google Groups too.


I don't remember GooGroups bounces being an annoyance.. but one never 
knows..




Re: New spamming trick?

2014-10-10 Thread Martin Gregorie
On Fri, 2014-10-10 at 20:49 +0200, Benny Pedersen wrote:
 On October 10, 2014 6:59:40 PM Martin Gregorie
  Benny: Yes they did - after all, how can they tell a bouncing message
  due to a fatfingered address from one that was crafted to bounce?
 
 the mailerdaemon is dkim signed, the attached msg is not signed, so its not 
 sent from yahoo imho
 
True enough: I thought you were asking if the bounce message had been
signed, which it had - by Yahoo. As that message is only an attachment
that originally came from elsewhere, I'd have thought a DKIM sig on it
was irrelevant.

  The examples I've seen so far have apparently been equity pumping scams.
  Is this also a common feature?
 
 Ahh note the isp send you a dsn back for undelivered, here the isp is 
 really yahoo,

Of course. I see it because the sender was forged, but I wouldn't call
it Yahoo spam unless you can tell me how Yahoo is meant to tell a
misspelt group name from one that's a deliberate mismatch. 


Martin





Re: New spamming trick?

2014-10-10 Thread Martin Gregorie
On Fri, 2014-10-10 at 21:03 +0200, Axb wrote:
 On 10/10/2014 08:39 PM, Martin Gregorie wrote:
  On Fri, 2014-10-10 at 20:17 +0200, Axb wrote:
  Thanks for the sample...
 
  Was wondering why I didn't see any
 
  had an ancient Postfix header_check regex rule
 
  /^X-Yahoo-Newman-Property: groups-bounce/  REJECT
 
  Does this only appear in Yahoo groups bounce messages? If so , I'll add
  it to the rule and/or replace my current List-id name match.
 
 honestly, I couldn't sign that - my rule dates back to 2006 and I've 
 never had a complaint - it's a works for me
 
OK, understood. Thanks.

  I searched for information but only found people saying 'I dunno' only
  with more verbosity. Apparently Yahoo doesn't publish and descriptions
  for these headers.
 
 (I have no use for Yahoogroups mail)
 
  Same here and extend it to include Google Groups too.
 
 I don't remember GooGroups bounces being an annoyance.. but one never 
 knows..
 
I don't know about GG either - only that I don't/won't use them while
NNTP: newsreaders suit me much better than the web forum type of
interface.


Martin