You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@spamassassin.apache.org by Nels Lindquist <nl...@maei.ca> on 2004/03/18 18:46:21 UTC
New Habeas information (on behalf of Des Cahill)
Hi SA Community -
A lot of you have been asking us for an update on Habeas' response to
the current spammer spoofing attack on Habeas Sender Warranted Email.
Here's what's going on:
a. Legal action.
A full legal and technical investigation of the matter has been
underway since January 12, 2004 (the day after this all began) and
legal action is expected to begin shortly. The support from the email
community to date has been phenomenal and has been instrumental in
our investigation. Any examples of the spam should be emailed, with
full header and body, to report@habeas.com. Any additional
information regarding this matter may be reported to habeas at
support@habeas.com. The legal model can be very effective against
spammers, but it is a slow process.
b. Implementation of whitelist reference in SpamAssassin.
Spam Assassin 3.0, scheduled for release in 6-8 weeks (I believe)
will incorporate an improved Habeas ruleset that will automatically
refer to the Habeas Whitelist (aka the HUL or Habeas Users List) in
addition to the Warrant Mark. Use of this version of SpamAssassin
should defeat these Habeas Warrant Mark spoofing attacks.
c. Enhancements to SpamAssassin 2.6x.
Habeas is working with the SA community to implement a software patch
that will augment the Habeas ruleset of SpamAssassin 2.6x with the SA
3.0 functionality for Habeas mentioned in b. above (i.e., checking
whitelist). Details on the ruleset will be released shortly as soon
as we complete testing.
I should add that some interesting ideas for additional future
ruleset enhancements have been floated to Habeas from the SA
community - we are intrigued and will be taking a look into them.
[ed. note -- see below for my suggestion in this regard]
We appreciate your support as we transition fully away from the
header plus blacklist model to header plus whitelist model in SA.
We're already using header plus whitelist in our ISP relationships.
Clearly the tactics of spammer/hackers (i.e., zombie PCs) have
rendered the header plus blacklist combination ineffective. We
appreciate the difficulties this has given all of you and we hope for
your continued support as we transition (quickly!) to a more secure
Habeas implementation in SA.
thanks -
Des Cahill
CEO
Habeas, Inc.
> -----Original Message-----
> From: technical-discussion-bounces@habeas.com
> [mailto:technical-discussion-bounces@habeas.com]On Behalf Of Nels
> Lindquist
> Sent: Wednesday, March 17, 2004 4:46 PM
> To: technical-discussion@hil.habeas.com
> Subject: [Technical-discussion] Some suggestions on behalf of
> SpamAssassinusers
>
>
> Hi there.
>
> There's been quite a bit of discussion about Habeas on the SA list of
> late due to the recent rash of spammers utilizing the Habeas SWE to
> cut through content filters.
>
> A lot of people are growing impatient and setting their HABEAS_SWE
> scores to zero, and some misguided folk are actually assigning it a
> positive score based on the belief that they receive no legitimate
> mail with the Habeas SWE! While I understand the legal process can
> take a while (grinding slow but exceedingly fine, etc.) it might be
> nice for you guys to provide a status update on your website (or even
> a post to the SA list!) a little more often so people at least
> understand that the process is ongoing.
>
> In the interim, though, I'd like to make a more technical suggestion.
> Since the Pharmacourt infringers have been using hijacked open relays
> (perhaps infected with some MyDoom or Beagle variant?) merely putting
> IP addresses into the HIL is something of an excercise in closing the
> barn door after the horse has gone (though certainly not a waste of
> time, since there are other horses).
>
> It's been my observation that the URLs referenced in the Pharmacourt
> spam are far less variable than the relays they use (makes sense--
> domains do have to be registered and that costs money), and that
> lends itself to the creation of a custom SpamAssassin ruleset which
> is demonstrably quite effective. Working on the suggestions of SA
> list participants, I've been building the following based on URLs
> referenced within infringing e-mails I've received:
>
> uri PHARMACOURT_BIZ /\b(?:affiliatedrugs|affiliateddgrugs| \
> charterdrugs|fifthdimensionrx|firstassist| \
> majesticdrugs|moderndrugstore|pharmacourt| \
> pharmawarehouse|valuepointmeds)\.biz\b/i
> describe PHARMACOURT_BIZ Includes a link to spammer
> www.pharmacourt.biz
> score PHARMACOURT_BIZ 3.0
>
> meta HABEAS_VIOLATOR_LOCAL (!HABEAS_VIOLATOR &&
> HABEAS_SWE && \
> PHARMACOURT_BIZ)
> describe HABEAS_VIOLATOR_LOCAL Spammer known to abuse Habeas mark
> score HABEAS_VIOLATOR_LOCAL 16.0
>
> Over the past few weeks, these local rules have been catching
> infringing e-mails about twice as often as HABEAS_VIOLATOR (which is
> triggered by an infringer on the HIL) and very few infringing
> messages have made it to my Inbox.
>
> So here's what I propose: Habeas should create (or contribute to) a
> custom SpamAssassin ruleset! (see
> http://www.merchantsoverseas.com/wwwroot/gorilla/sa_rules.htm for
> examples of other rulesets in common usage.)
>
> Since Habeas receives lots of reports of infringing mail, you could
> pull out URLs used by the infringers and generate a more
> comprehensive ruleset along the lines of what I've outlined above.
>
> By updating such a ruleset as new URLs are discovered and making it
> available for download, SA users can be better protected from
> infringing spam while still retaining the benefit of properly scoring
> legitimate Habeas SWE users. Additionally, by posting notices of
> updates to the SA discussion list, Habeas would be seen to be
> actively participating in the community which can't help but incur
> goodwill for the company. I wouldn't be surprised if more people
> were encouranged to report their infringing mail, too, if they
> perceive a direct benefit to themselves by doing so.
>
> So there you have it. Food for thought, anyway!
Nels Lindquist <*>
----
Quidquid latine dictum sit altum viditur.
Whatever is said in Latin, sounds profound.
Re: New Habeas information (on behalf of Des Cahill)
Posted by John Andersen <js...@pen.homeip.net>.
On Thursday 18 March 2004 08:46, Nels Lindquist wrote:
> > Since Habeas receives lots of reports of infringing mail, you could
> > pull out URLs used by the infringers and generate a more
> > comprehensive ruleset along the lines of what I've outlined above.
> >
> > By updating such a ruleset as new URLs are discovered and making it
> > available for download, SA users can be better protected from
> > infringing spam while still retaining the benefit of properly scoring
> > legitimate Habeas SWE users.
This seems to be on a potentially promissing track.
But I submit that these offending URLs would be more effectivly
tracked with DNS like approach which, I believe, could be handled
by something like Razor.
In fact I recall Vipul mentioning this sort of use on the Razor list.
It would/could/should be a separate database than the existing
razor database, but the result would be the same.
Submit a url to an engine, engine extracts interesting parts,
hashes same, submits hash to a razor like engine, and gets
a simple binary answer back.
In fact, I would bet this could be done by Bind, because dns is
really fast, and it really does not care what its looking up.
--
_____________________________________
John Andersen