link laundering – Dr John's Tech Talk

Intro
I’ve been seeing an uptick in brief spams which provide links to a very legitimate site: the Google search engine!

The details
I’ve been getting a lot – several per day – that look like this one:

From: [email protected]
To: [email protected]
Subject: Legal drugs forum

Legalize!!! Read about strongest legal drugs in the world, and buy it online: https://www.google.com/url?q=http%3A%2F%2F%77%77%77.le%67a%6C%69z%65r%2EDRJinfo%2F&sa=D&usg=AFQjCNG0coaOvXJMkOn0nEMvP-dl11XKnQ

Attention: MDMB(N)-BZ-F is not allowed now!

Here’s another example which appears to be a different spam campaign using the same technique which I received several weeks after initially posting this article:

From: [email protected]
Subject: Turn your bedroom into paradise of satisfaction
 
https://www.google.com/url?q=http%3A%2F%2F%73lip.h%65al%69DRJn%67%73%65%63%75re%65%73hop.%65%75%2F&sa=D&usg=AFQjCNFeP_XevUiXV-m-DtxAJVi3SMRtVQ

I’ve changed the links slightly so no one gets in trouble by actually following it.

The link is changed each time and so is the sender.

How to report this?
I have been reporting these to Google directly on their page to Report malicious software, https://www.google.com/safebrowsing/report_badware/.

I have reported five to then of these and have never received a response from Google. It seems the best we can hope for is that Google engineers are sufficiently annoyed by my reports that they begin to agree hey there’s a problem here and maybe people will think less of us if we continue to do nothing.

Why this is particularly devastating
Because the malware link uses this combination:

– https (which encrypts everything)
– a very legitimate web site, www.google.com
– malware

It is very tricky to defeat. Many URL filters, e.g., those used on explicit proxies, cannot peer into https traffic and so have to make a single judgment for a whole site, even one as complicated as www.google.com. Either it is all good, or it is all bad. Who would have the courage to categorize Google as a source of malware and hence block all users from it?

So these perpetrators have engaged in what amounts to link laundering. Some of the URI is encoded in hex, I suppose to help avoid detection and create many valid patterns that are hard for Google to stamp out.

This started over a month ago and is stronger than ever today, so we know at press time Google, in spite of all its advanced technology, does not have a handle on it.

If you see something similar I suggest to report it directly to Google. They may need a little more motivation than I can single-handedly provide them.

Conclusion
Link laundering is now an avenue to sneak spam through. It uses links that point to the Google search engine itself. It seems to have eluded them or been under their radar in spite of many reports. Let’s hope the bad guys don’t have the upper hand permanently.

Appendix
If you are interested in how the URL looks decoded I figured there would be decoders available on the Internet and indeed there are. For instance at http://meyerweb.com/eric/tools/dencoder/

So the URL mentioned above decodes as (again just slightly obfuscated to not make good people do bad things by mistake:

https://www.google.com/url?q=http://www.legalizerDRJ.info/&sa=D&usg=AFQjCNG0coaOvXJMkOn0nEMvP-dl11XKnQ

References
enom-originated spam is discussed here.

In my case – not much! It appears that despite providing top-notch content the only “readers” are those trying to profit from me. To use the word “scam” may be a bit strong, but any outfit that demands money upfront to supposedly help you make money is highly suspect in my playbook.

So I’ve heard from Tina. It goes like this:
Admin – I’ve checked out http://drjohnstechtalk.com/blog/2011/06/grep-is-slow-as-a-snail-in-sles-11/ and I really like your writing style like in your post Grep is Slow as a Snail in SLES 11 | Dr John’s Tech Talk. I am looking for blog authors who would like to write articles as either a full time job or part time job (for some extra money). I think your writing style would work very well. You receive pay per article, anywhere from $5 to $50 per article depending on the topic, article length, etc… If interested you can find more information at www.onlinehomewriter.net.

Please do me a favor and do not follow that link. It redirects you secure.signup-way.com, some strange-looking URL that McAfee categorizes as Malicious Sites, High Risk. So I don’t think I’ll be going there.

Then there’s Tony:
Blog Admin – If your blog isn’t bringing in as much money as you would like it to check out my site www.QuickCashBlogging.com. We show blog owners how to maximize their blogs earnings potential. Tony

McAfee verdict: Spam site, medium risk. That’s just great.

The McAfee URL checker I use is http://www.trustedsource.org/en/feedback/url.

Clearly these people have program trolling the Internet for new domains and new blogs, trying to squeeze some $$ from them. Unfortunately I’m not sure any person who could benefit from the information has read my blogs. So I feel I am making negative progress – instead of elevating the level of discourse on the Internet helping it to be used for more spam and scams.

I just feel bad for humanity. Is this the best we can do? A well-meaning person embarks on a quixotic journey to provide better technical information on some topics, and the average response from my fellow human beings is to try to take advantage of a hopefully vulnerable and naive newbie? I am literally concerned for us as a race.

August 16th Update
The spam and scam started as a trickle. Now it’s raining spam in my inbox. I continue to be disappointed. In email the ratio of spam to “ham” may be about five to one, so not knowing any better you could expect a similar ratio with WordPress blogs. Not so! Of my fifty comments, not counting the ones from myself, the legit comments number about two-and-half, more like a twenty to one ratio. I will probably use a WordPress plugin to cut them off, but since I started on this public service mission, here are some more scams.

This one is spam as it was posted to my Sample page:
HTC is a well known name in the smartphone segment. The company has come up with smartphones boasting of exquisite features and HTC EVO 4G is one of the most potent… It came from an address ending in @mail.ru . One of many hats is as spam fighter. Let me tell you you see an a sender address mail.ru and you’re talking pure spam. The IP resolves to Latvia, however, and that fact hardly inspires confidence either.

To my post WordPress, Apache2, Permalinks and mod_rewrite under Ubuntu I got a comment The Best Way To Fix Acid Reflux. Now that’s a closely related topic!

Another one claims to help if I’m looking for information about babies (very relevant for a tech blog. yeah, right!).

Very many fall into the generic flattery category. Like this one:
Hey There. I found your blog using msn. This is a very well written article. I will be sure to bookmark it and come back to read more of your useful information. Thanks for the post. I’ll definitely return.

Or this:
I agree with your Gnu Parallel Really Helps With Zcat | Dr John's Tech Talk, great post.

I had to investigate those a little bit as I almost fell for one the first time. Then you realize that it’s so generic – except for the one where he obviously just pasted in the title of my post programatically – that it could be used for any blog post.

An equally popular scam are the SEO scams – Search Engine Optimization. I think the point of those scams is to shake a little money from you for supposed help to improve your blog’s ranking in the search engines.

Returning to the flattery scams, how do I know for sure this isn’t real, genuine flattery of my wonderful posts? I’ll tell you. There are a couple unambiguous clues and another strong hint.

Let’s start with the strong hint. Since I haven’t told anyone about my blog, pretty much the only way someone’s going to find it who has legitimate interest in its content is through Google or another search engine. So, in the web server access log, where I am recording the HTTP_REFERER (what URL the browser visited just before hitting my blog post), I should expect to see one of the search engines. I should not see some random web site mentioned because there is simply no good reason for browser controlled by a human being to go from someone else’s web site directly to my web site. And yet that is precisely what I am seeing. I would give examples but it would only serve to promote their web sites, so i will refrain from even an example.

But even more damning is to examine how long the poster has spent on my site. A human being has to read the post, contemplate its meaning, then type in a comment to finally post, right? It could rarely be done in under a minute. WordPress tells me the IP of the poster of the comment. I take that IP and search for it in the access log using grep. I am seeing that these comments are being made in one second after the web page was first downloaded. One second. It is not humanly possible. But for a program, piece of cake.

Here’s a real example (some of this may be cut off, depending on your browser):

109.169.61.16 - - [12/Aug/2011:06:31:51 -0400] "GET /blog/2011/06/gnu-parallel-really-helps-with-zcat/ HTTP/1.0" 200 22757 "http://blahblah.net/invest/" "Mozilla/4.0 (compatible; MSIE 7.0; Windows NT 5.1; .NET CLR 1.1.4322; .NET CLR 2.0.50727; .NET CLR 3.0.04506.30)"
109.169.61.16 - - [12/Aug/2011:06:31:51 -0400] "GET /blog/ HTTP/1.0" 301 340 "http://blahblah.net/invest/" "Mozilla/5.0 (Windows; U; Windows NT 6.1; en-US) AppleWebKit/532.2 (KHTML, like Gecko) Chrome/4.0.221.7 Safari/532.2"
109.169.61.16 - - [12/Aug/2011:06:31:51 -0400] "POST /blog/wp-comments-post.php HTTP/1.0" 302 902 "http://drjohnstechtalk.com" "Mozilla/4.0 (compatible; MSIE 7.0; Windows NT 5.1; .NET CLR 1.1.4322; .NET CLR 2.0.50727; .NET CLR 3.0.04506.30)"
109.169.61.16 - - [12/Aug/2011:06:31:52 -0400] "GET /blog/ HTTP/1.0" 200 110806 "http://blahblah.net/invest/" "Mozilla/5.0 (Windows; U; Windows NT 6.1; en-US) AppleWebKit/532.2 (KHTML, like Gecko) Chrome/4.0.221.7 Safari/532.2"

Also this example illustrates the other damning evidence of lack of human involvement in the comment. A real browser run by a real human being has to pull in a lot of objects to display a single WordPress page. You’ve got stylesheets, external javascript pages and even the image at the top. They should all be requested, and be recorded in the access log. But a programmatically controlled browser needs far less! It needs the HTML of the blog page, and then the page it POSTS the comment to. Perhaps a third page after the POST to show it the POST was successful or not. And that’s exactly what I’ve seen in all the spam/scam comments I’ve checked out by hand, not just the flattery scams. They are all using the absolute minimum page accesses and that simply screams non-human access! I am, unfortunately, not really so special as they would have me believe! And the SEO scams are just annoying advertising. Most of the rest is what I’d call link laundering, where they’re using the legitimacy of my site to try to get links to their shady sites included, by trickery, carelessness or any means. And some are just using it as pure spam to my inbox since that’s where the comments go for review and they don’t even care if I approve their spam for public viewing or not.

Possible Explanation
My hypothesis is that there are specially constructed advanced searches in Google you can do to find new WordPress blogs. You can download the results and programatically loop through them and attempt to post your spam and scams. It’s pretty easy to program a browser like curl using PERL to post to a WordPress blog. Even I could do it! And that low barrier to entry jibes with the level of professionalism I perceive in these scams, which is to say, pretty low, like something I would cook up by my lonesome! Misspellings, poor English, blatant calls-to-action are par for the course, as well as source IPs from remote regions of the world that have no possible interest in my arcane technical postings.

Now you could argue that a real browser could have cached some of those objects and so upon a return visit it might only access a minimum set of objects and hence look a bit like a program. To that I say that it is rarely the case that all objects get cached. And even if they did, you still have to take time to type in your comment, right? No one can do that in a second. The access lines above span the time from 6:31:51 to 6:31:52!!

The Final solution
I think I’ve made my point about the spam. I have followed Ryan’s advice and activated a plugin called Akismet. Their site looks fairly professional – like they know what they are doing. An API key is required to activate the plugin, but that is available for free for personal blogs. I’ll append to this blog whether or not it works!

Feb 28th update
600 spam comments later, 20 in the last few hours alone, I am sooo tired of rotten apples abusing the leave a comment feature, even though I am protected from approving the comments, it is still filling up my database. So I have taken an additional step today and implemented a Captcha plugin. This supposedly requires some human intelligence to answer a simple math problem before the post is allowed. I’ll post here about how well it is or isn’t working.

September update
Well, the captcha plugin has stopped virtually all spam, except one random comment. A user wishing to post a comment has to solve a very simple math/language problem. I recommend this approach. I suppose eventually the scammers will catch up with this defense, but in the meantime I am now enjoying peace and tranquility in my seldom-visited but formerly frequently spammed blog!