Where is Leonardo’s Car – Using OSINT to trace vehicles

I love cars and I love OSINT. Sometimes I get to combine these passions. Not only for work, but also in little exercises that help sharpen my research skills.

A while back I posted a blog about using car spotting sites to find and track vehicles. The sites I discussed in that article where only the tip of the iceberg when it comes to finding information about specific vehicles online. Today, I want to walk you through other means of finding cars using unique identifiers such as license plates or VINs (vehicle identification numbers). There’s nothing fancy about what I’m going to show here. I’ll just follow the digital breadcrumbs using simple OSINT techniques.

For some reason I stumbled upon a Youtube video showing an Italian soccer player’s Ferrari. We’ve all been down that rabbit hole before. You start watching Youtube videos about cooking and end up somewhere completely different. Oh, the joys of the internet…

This video had a visible license plate and I was curious to see other places the car was spotted. My usual car spotter websites actually came up empty handed, no matter how I tried to enter the license plate number. So I took my search back to Google. Search engines actually OCR some of the images they index, so I entered the plate number and instantly received some results:

Next to the Youtube video that got me started, I found a blog in which the author posted multiple pictures of the car I was looking for. The plate number wasn’t listed anywhere as text on the website (checked through the developer tools as well: nothing came up), so Google must have OCRed it. Thumbs up to Google for this!

But wait, it gets even better. Google is not the only platform to OCR images, Facebook does so as well. So, I decided to take my search to Facebook and see if I could find further images of the vehicle there. Using the standard Facebook search, I entered the plate number. Keep in mind, throughout each search you might have to use different variations, adding spaces between characters or writing everything together.

The picture results are shown right away, as I have a direct hit in this query. Sometimes the picture results will not be shown in your main search results and you may have to click on the tab to the left to get to the image filter. Some guy on Facebook posted the Ferrari as his profile pic in April this year and this picture looks like it had the car at a repair shop or possibly a dealer.

Now, if this theory was right, the vehicle might not even belong to Leonardo Bonucci anymore. I could go looking for sales ads for such a vehicle and hope to find it. A lot of this would just be Googling and browsing through sales sites and would require a lot of tenacity and also a little bit of luck. Although, I still have an ace up my sleeve when it comes to Italian vehicles. This ace would allow me to find out more details on the Ferrari I was searching for.

I have a little app on my phone called iTarga. With this app, I can enter any Italian license plate and will receive further information on the vehicle. Here in Italy, vehicles are assigned license plates for life. Even if the car is sold, it keeps the plate numbers. Let’s see what iTarga tells me about Leonardo Bonucci’s Ferrari.

First date of registration, a VIN, insurance information (including insurance company and policy number) and the residence of the owner are among the things that can be found in the app. In our case, no insurance is listed. It is likely that the vehicle is not insured at the moment, adding to my suspicion that it is/was for sale. The owner’s residence is Milan, which happens to be the city Bonucci played in at the time most of the previously seen images were taken (he’s moved on to Juventus Turin now). These details give me further pivot points for my search. I could narrow down the results of sales ads to 2013 models and look in and around Milan or Turin (assuming it would be sold there). Or I could just simply Google the VIN.

Et voilà, I do receive results for sales ads. However, the vehicle offered here is a red Ferrari. I thought I was looking for a black one. And nowhere on the website can I find the VIN. See, zero results:

Yet again, a simple OSINT technique will help clear this up. Looking into the developer tools will enable you to search within parts of the website that aren’t directly visible to users. When checking the VIN there, I found that all uploaded images actually have the VIN in the file name.

Not only that, the URL also contains the VIN:

A little more research and everything makes sense. Bonucci originally drove the red Ferrari and had it wrapped in black foil. For the current sale, the black foil was apparently taken off again.

While this example utilized an Italian app, there are many similar sites for countries throughout the world (except in Germany…). The lesson to be learned here is to follow the digital bread crumbs. Sometimes seemingly simple OSINT techniques will lead you to your goal if you know how to combine them. And now you get an idea of how I spend my time when sitting in the passenger seat while my wife is driving. Googling license plates, checking car spotting sites and tracking the history of random exotic cars I see.

MW-OSINT / 16.10.2020

Social Media around the World

When most people speak of social media, the have the ‘Big 3’ in mind: Facebook, Instagram & Twitter. But social media is so much more than just these three platforms, especially when it comes to OSINT on intelligence targets that don’t speak English.

In OSINT investigations we often end up scavenging social media to find information on our intelligence targets. Who are they connected to? Where have they been? What are their interests? These and many more questions can be answered by having a look a person’s profile. However, social media is constantly evolving and platforms that were relevant yesterday may not be relevant tomorrow. When I ask my daughter about Facebook, she says: “Facebook is for old people”. Thus, she does not have an account there. You would most likely find her dancing in TikTok videos, as with many other Generation Z youths. So age clearly defines which social media platforms are used. Another defining factor is the cultural background someone has. Maybe Facebook was never that big in that person’s country. The following graphic shows the evolution of social media worldwide and how Facebook became the most used platform. However, in some countries other platforms still have the upper hand and not all ‘legacy’ platforms overtaken by Facebook have been shut down. In this article I would like to give a brief overview of some of the lesser known platforms that may be useful for OSINT investigations.

VKontakte & Odnoklassniki

If your intelligence target is from a Russian-speaking country or has a Russian cultural background (or is a right-wing idiot that thinks he is being censored on Facebook), chances are high you might find this person on one of the Russian Facebook clones. The platforms VKontake (‘in contact’) and Odnoklassniki (‘classmates’) are very similar to Facebook when it comes to the functionality offered and the basic OSINT research techniques that can be applied here.

Above you can see a VKontakte profile. A profile picture, some more detailed information including a birthdate and current residence city, a friend list as well as posts and pictures. Pretty much what you can find on an average Facebook profile. As with other social media platforms, a user can choose to alter the privacy settings to hide information, so some profiles may not have an open friend list or may not share all posts with the overall public. An interesting feature on VKontakte is in the top right of the image: information when the profile was last active. In OSINT this is really helpful to figure out if a user is still active on the platform, even if no current content is posted. In many cases this last activity will lead back to the use of VKontakte as a messenger. People might not post content anymore, but will stay В Контакте (in contact) with others through this platform. The search functionality of VKontake is in some ways superior to what we now have on Facebook. At the top of the page is a search box. Filling in a search term here will enable us to browse through different categories of results and narrow these down by adding additional filters.

As you can see, you can filter people by age range, birth date information and even their views on smoking an alcohol. Posts can be sorted by the number of likes or the mentioning of specific links. All in all, there are some pretty neat filters in here.

Odnoklassniki is very similar, having friend lists, a date of birth on most profiles and information when the user was last active. The good thing with both VKontakte and Odnoklassniki, is that they accept multiple language settings, so you can use the platform in English and also a couple of other languages. If you search for names in Latin script, it will also show you corresponding results in Cyrillic script.

The last activity is right underneath the profile name and the searches in Odnoklassniki offer filters just like in VKontakte. They even allow users to add holiday destinations, which are also a filter criteria.

As I mentioned, this article is just a quick overview of some foreign social media platforms. There lots of other cool OSINT techniques that can help research here, including third-party sites to search by profile pics or sites that help with geo-referenced searches. But let’s leave that for future blog posts. Another example I want to show is very popular in the Persian-speaking community.

Facenama

Facenama is a big social media platform mainly used in Iran. At quick glance on SimilarWeb shows that this site is also accessed from other countries, as there are Iranian communities throughout the world.

Facenama looks very much like Facebook. Even the coloring scheme is identical (to the old Facebook UI).

Unfortunately, there is no way to change the language settings, but luckily the Google translate browser extension works quite well here.

The search bar in the top right of the page will enable you to search for user profiles. Just remember that the default language is Farsi, so most profiles will be in Arabic script (including profile names) and typing will occur from right to left.

The profiles will have the same type of information we have seen in the Russian sites: date of birth, friends, posts and much more (if these aren’t hidden due to privacy settings). Remember that dates will be shown in Persian, so you’ll probably have to use a calendar converter to make sense of these dates.

I could go on for hours listing and showing social media platforms: Gab for right wing nut jobs, Stayfriends for old German people, NK for Polish people and don’t even get me started on Chinese social media. The bottom line is, that there is more out there than just the ‘Big 3’ (Facebook, Instagram, Twitter). Before you start investigating someone, you should figure out where you might find these people online. Their age, culture, language, country of origin and personal taste will affect their choice of which web platforms they use and these might not always be in English. So, in the ongoing discussion of what I would like to get better at in OSINT, I didn’t choose to learn programming languages such as Python to automate tasks. I’d rather get a better grasp of languages (Arabic, Farsi, Russian, etc.) in general and master tools that help translate to help bolster my research efforts.

MW-OSINT / 04.10.2020

Social media is dead, long live social media!

Is your intelligence target under 25 and not on Facebook? You might want to check the social media that kids nowadays are actually using!

My daughter always says: “Dad, Facebook is for old people!” It’s true, I’ve noticed that many people under the age of 25 aren’t on ‘traditional’ social media anymore. They are not on Facebook and they may give a confused look if confront them MySpace, GooglePlus or walkmans.

So, how and where do you find Generation Z on social media. Clearly, they still feel the urge to express themselves on the internet and they’re still out there, but mostly not with their real names. This makes OSINT much more challenging. On Facebook we could search for real names, we could search by phone number and in some cases we could find people through email addresses. Some of these techniques work on other social media platforms, some won’t. In any case, if you find a profile linked to one of your targets, you might come across further social media profiles that your intelligence target has backlinked on the one you have found.

I’ve noticed that many young people use TikTok, an app designer to share short music videos. It contains likes, friends and comments, similar to what we know from ‘traditional’ social media. Luckily, the TikTok app allows you to find profiles linked to phone numbers. For this, you need to install the app either on your burner phone or in an AndroidVM, then go to the profile page and tap the ‘add contact’ button on the top left. The red dot indicates that new contacts have been found.

Next up, choose the option in the middle, stating that would like to find contacts from your phone book. This of course means you have to add the phone numbers of your intelligence targets to the phone book first and give TikTok access to it.

Tapping ‘find contacts’ will show the amount of phone numbers that are linked to TikTok accounts and it also gives you the choice to follow them. It looks like some of my contacts are actually using TikTok.

If you have a nickname, even one derived from other platforms, these can be looked up in the app itself too. TikTok will only allow you to search for the beginning of the nickname and not for parts in the middle or last portion of the name. In the following screenshot I looked for nicknames containing ‘James’ and I was only shown names starting with ‘James’. The reason this is relevant, is that I have often found TikTok accounts to use prefixes or suffixes on their regular nicknames. So instead of just ‘James’, you might find the user as ‘xyz.james’ or ‘james.1982’.

However, there is a workaround for this. Just like with Instagram, there are many sites that scrape TikTok and display the accounts and in many cases the content as well. One of the ones I like to use is PlayTik. PlayTik allows you to search for hashtags and accounts. Let’s find an account that somehow uses ‘f1nd1ng’ in the nickname.

There we go, two accounts containing the searchterm. Now you can have a look at the profile and check out any videos this profile has uploaded (and publically disclosed). It looks like this particular profile also links to further social media and websites, like I had mentioned before. Plus, the profile contains a video. Feel free to watch it!

Facebook may be fading (soon), but others platforms will replace it. Thus: Social media is dead, long live social media! The new platforms are not just for young people, so go and try them out (research them) yourselves!

MW-OSINT / 13.09.2019

Unravelling the Norton Scam

It all starts with a bad sock puppet

If you have problems with Norton 360 or Norton Antivirus, please do not call +1-844-947-4746. You might end up with malware on your computer.

Do you have a look at the accounts that connect with you on Twitter or Medium? I do, and so does my buddy Sector035. In late April 2019, a new person followed Sector’s blog on Medium and he had a look at this new follower.

A weird URL? A nice picture of a female named Pierre? This profile was begging for further research. The URL led to a tech-support site that listed the following phone number: +1-844-947-4746. Sector didn’t even wait to check this number on his computer and immediately googled it on his cell phone. I guess that’s what you call OSINT curious.

It turns out that this phone number was listed on numerous obviously fake sites and blog posts offering tech-support. Out of curiosity, we decided to take a closer look at some of the sites, in order to see how they were connected to each other and possibly find out who was responsible for creating them. At the time we had no idea how time consuming and big this project would be! Among the sites using the phone number, we initially concentrated on these four:

Each site looked worse than the other. Horrible design, bad English and next to the aforementioned phone number, they all used the same address:

While Sector started to check the WHOIS information using DomainBigData, GoDaddy and Whoxy, I looked into to Google Street View and did a little reverse image searching on the photos. It turns out that all the photos used were either stock pictures or stolen off other people’s social media profiles and the address itself was in an inconspicuous housing area. Googling the address led us to more suspicious sites, some of them using a different phone number. Among these was one belonging to a company allegedly called Energetics Squad LLC. No records existed for such a company in the State of Illinois, nor in any other state. Keep this company in mind, as it will show up in a later blog post as well!

The WHOIS check didn’t always provide the exact name of the registrant, but we found another similarity: most of the websites had been registered around March 13-14, 2019 in India.

Using DNSLytics, Sector also checked the Google Analytics ID and found that the sites were not only linked by all of what was described above, they also shared a common tracking code (UA-code). At this point, it was time to start linking the information in Maltego.

What started with a bad sock puppet, led to googling information and from there to a deep dive into domain data, Google Analytics research, as well as pulling corporate records from official state registries. The hunt was on and upon finding all this correlating data, we couldn’t just let go and decided to push forward.

The Art of OSINT

Art is often considered the process or product of deliberately arranging elements in a way that appeals to the senses or emotions. OSINT is art and sometimes OSINT produces art.

This is big! We collected information on hundreds of different entities; including websites, names, phone numbers, email addresses and much more. When we started, relevant data was just dumped into a text file. To give you an overview of how much we collected and how much time and effort we spent on this project, Sector put together a little infographic.

We also used Hunchly during our collection and then realized we needed to structure the most important data and moved on to a spreadsheet. This worked for a while, until we felt the need to display links between entities in an easy and understandable way. A more visual approach was chosen, and Sector started what I call “The Art of OSINT”: a link chart.

Sector used Maltego for our case, but there are many alternatives you can use as well. I have grown fond of draw.io and others might use one of the various mind mapping apps and platforms. The idea behind link analysis in general is to evaluate relations and connections between entities. Link charts are the visualization of this data, which in many cases make it easier for an analyst to discover connections. Sometimes the connections are not direct, but indirect, linking entities to each other by a third-party individual.

Let us take a look at how link charts can be built from scratch. I previously mentioned a screenshot of one of the scam sites.

This site already has multiple pieces of information that can be connected. A name, a postal address, a phone number and lastly an email address. Each of these is the starting point for further OSINT investigtions. We found that the email address was used to register the domain allbagmanufacturing(dot)com. This domain also lists an Indian phone number in the WHOIS data.

It turns out, that the Indian phone number was also used to register the site roadrunneremailsupports(dot)com.

In conclusion, both sites are likely connected. However, how do we know that this phone isn’t just a burner phone or a random phone number? More OSINT research is required to verify if the phone number is existent and who it belongs to. I also mentioned a little social engineering coming up, didn’t I? If you were hoping to read about this topic now, I have to disappoint you. Rest assured, we have a nice story on social engineering later on. Now, let’s get back to our link analysis.

One piece of information leads to another and soon we find new leads and many connections between the entities. The chart itself has grown quite a bit in the meantime. At first glance, it will seem a bit chaotic. However, it is still easier to handle than relaying this information in a text-based form. For me, link charts have an artistic character. Each chart, whether built manually or automated, is one of a kind. Unique data, unique arrangements, all coming together to form a piece of modern art.

Put this on a large canvas, have Sector sign it and it would be something that could be found in the Guggenheim Museum of Modern Art. One day I plan to do exactly that. A vernissage on “The Art of OSINT”. Until then, let’s keep creating more masterpieces with our online investigations and link charts.

What’s the big deal? And who’s to blame?

Not all information can be found using OSINT. Sometimes a little social engineering can be useful.

We warned you not to call a certain phone number linked to multiple scam sites. So, what exactly happens if you call +1-844-947-4746? Although it was obvious that the sites we were looking at and this phone number were involved in some kind of scam, I was curious to see what the exact business model was. Before making the call, I googled a bit and found out that the number was toll-free in the US. After topping one of my burner phones, I dialed the number.

It rang shortly and a gentleman with an Indian accent answered and asked which kind of assistance I required. I explained that I had problems with Norton 360. I couldn’t activate it. The gentleman wanted my name and my phone number, upon which I provided him with some bogus data. At first, he asked if I was able to install software, as he wanted to use Supremo to check my computer. Playing dumb, I just told him that I didn’t understand what he wanted me to do and that I have never personally installed anything on this computer. My son always did all the IT-stuff for me.

Next up, he mentioned two URLs that I should try to visit: helpme(dot)net and 1234computer(dot)com. It turns out that both sites were also meant to give him remote access to my computer.

This is where another important lesson in OSINT (and life) can be learned: Proper preparation prevents poor performance! I had not expected this and therefore hadn’t set up a clean Windows VM to play around with. Sadly, I had to find an excuse why I couldn’t access the URLs he had cited. I promised to call back, but never managed to get through again. From there on, the number only redirected me to voicemail. Funny enough, the mailbox mentions a typical American name as the owner. Maybe this could be useful in the future.

In any case, it was clear how the scam worked. Unsuspecting, non-tech-savvy users would call the hotline and give the scammers remote access to their computers. From there on, the possibilities to do harm are countless. Ever since then, I have always made sure to have a clean Windows VM that I can use in such cases.

Since we now had an idea what the scam was about, we decided to push forward with our investigations. Sector found a phone number associated with the site energeticsquad(dot)com. This site belonged to an IT-company named Energetic Squad LLC in Illinois, which coincidentally was also located at the address we had already seen on multiple scamming sites. The number found by Sector was an Indian mobile and I decided to get in touch with this person to find out if he or she was in any way related to the aforementioned website or any of the other scam sites. I noticed that it was registered in WhatsApp, so I decided to have a little chat. Of course, I used a burner phone for this; to be more precise, I was running WhatsApp in an Android VM on my computer.

What struck me here, is that I was indeed on the right track. I never said that the company was a LLC, yet it was immediately mentioned by the person I was texting with. By this, the owner of the Illinois company offering local tech-support was using an Indian phone number. He wasn’t keen on giving me his name, but his reactions showed that me that the name I had was also likely correct. At least he didn’t deny it.

I offered to buy his domain and we went back and forth regarding the price. I was pushing hard, something I usually wouldn’t do in a real case, and soon he asked if I could come to the US for further negotiations. I assume he was just trying to put me off track and distracting from the fact that he was actually located in India. His English also wasn’t what I would expect from a native speaker. Of course, I was able to travel to the US and at this point he decided to end the conversation.

Being blocked by a scammer, because he accuses you of being a scammer. Now there’s a pot calling the kettle black! Of course, I burned bridges here and wasn’t able to reestablish contact after this. However, we did learn that the Indian mobile phone number was in fact connected to the website and the company. Also, we had likely found an actual name. Later on, we found out that the company Energetic Squad LLC had been registered in Illinois in the meantime, and that the name of their manager was the same name that is being used on the fake tech support voicemail. Everything and everyone is linked!

There were several other suspects which we had also contacted, and almost everyone acted as suspicious as the person I wanted to buy the domain from. In any case, these conversations gave us many new leads to follow up on and had us pivot toward social media profiles. Maybe these will provide some insight on the scam network.

The more, the better

What are backlinks and how are they used in the Norton scam? Our OSINT investigations lead us into the world of SEO.

The project started with a fake profile on Medium, which led us to several scam websites claiming to provide tech support. While the total number of these sites hasn’t risen much over the past months, entries promoting this scam on blogs, social media profiles and in comments on other websites have drastically risen. We see this as a crude search engine optimization (SEO) attempt.

Throughout the investigations we found individuals specializing in SEO, who where also likely linked to the network we were tracking. One of the sites that popped up in our search was yahoophonesupports(dot)com. Unlike previous fake sites that used Indian addresses and fake English-named personas, such as Nancy Wilson or Steven Dalton, this one was registered by someone named Jiya with a real looking email address. We had narrowed down our search to the city of Noida in India and the phone number used to register the site was definitely linked to our scam network. While most names used to register domains clearly came from fictitious peoples, maybe this was one was real.

Searching for the name on Google led to a result that fits the picture: Jiya from Noida offering SEO services.

Jiya mentions backlinks and blogger outreach, something we have seen in our scam as well. Let me explain the concept of backlinks, also known as inbound links. A backlink is a link on page A referring to page B. Most search engines interpret backlinks as votes on the popularity of a website. So, the more backlinks that lead to page B, the more popular this page seems and thus it will be rated higher in search results. The easiest way to create cheap backlinks is using free blogs, for example Medium. Googling the phone number used in this scam, we come up with over 1.000 Medium posts and sites, each also containing the link to one of the fake support sites, such as nortonhelpus(dot)com.

We are not only seeing this on Medium, but across many platforms. The amount of backlinks created clearly indicates that we are dealing with a large team of people, as this is probably not possible by one person or a small team alone. The number of search results for the phone number shown above has risen from 4,000 in May to about 21,000 this week and is still rising!

Of course, most posts come from obvious sock puppets, created with fake names and stolen profile pictures. Here’s Brad Pitt, alias James Rocky, offering tech support.

Once these sites are set up, clicking on the links can be automated, so that the target website (in this case all-emailsupport(dot)com) receives traffic, basically boosting its search index rating. The scam network is not expecting to generate any phone calls or support requests from these obviously fake Medium sites, these are just used in the SEO process for the actual scam site. We have been seeing these SEO-enablers on Twitter, Issuu and basically any platform that allows you to post information “quick and dirty”.

During our project we also looked at Google Trends, to figure out where the main targets of this scam were likely located. Obviously in the US, as the main tech support phone numbers were US toll-free numbers. However, we also noticed a British phone number. Google Trends allows you to look up search terms and see the interest over time and the region of interest for that specific search term. We were curious to see if any notable searches on one of the scam topics was googled in the UK often. We checked “activate Norton”, since this was one of the main services the scam network was allegedly offering: activating an expired Norton account.

Sure enough, the main regions that googled this term were the US and India. The US is obvious. This is the market the scammers are targeting. But why does India rank higher? One explanation could be testing the search term during the SEO-process. Coincidently, some of the peaks in the interest over time relate to time periods in which new fake sites were created. No notable searches were seen coming from the UK.

As mentioned before, most of the bogus blog posts were created really sloppy, as their sole purpose was to generate backlinks. We also found several sites linked to this scam that were apparently selling software. A lot of these sites were created using predefined templates, in some cases showing the shipping time or the general location on a map.

Apparently, the creator of this site had his location (likely based on a geolocation through Google) automatically added and didn’t bother to change it before putting the site online. In any case, he achieved his goal: A backlink to a scam site.

Now that we’ve learned how this scam uses SEO to promote their wrongdoings, is there anything that can be done to effectively tamper this scam network? How about Google, Bing, Yahoo and other major search engines take any site off their listing that features “+1-844-947-4746”. Except ours, we’re the good guys!

Mistakes on social media

Never be too careless with what you post on social media. It might come back to haunt you one day and in the case of our investigations it sure led to some key findings on the scam network!

The ultimate goal of our investigation was to find the people behind the scam. We had found fake personas and most of the WHOIS data was misleading as well. Of all the people we came across, there was still a reasonable amount of doubt that they were actually involved in the scam, or we couldn’t reveal their true identities for sure. In our hunt for the truth we also turned to social media, hoping to find someone connected to our case.

As previously mentioned, most names used to register websites were clearly fake. Nancy Wilson and Steve Dalton were among the favorites. However, we did notice that the majority of sites had Indian addresses in the WHOIS data. In particular, two addresses reoccurred often: one in Noida and one in Gurgaon (also known as Gurugram). We decided to find Facebook profiles that had visited both cities and that were also linked to East Peoria, the city in Illinois that the scammers posted on their websites. Maybe someone was careless enough and had checked into all three places. Of course, this was done a while back when the Facebook graph search was still working.

Jackpot! While we hadn’t expected to find anything on this wild query, one Facebook profile matched our search. This person, named Baljeet, was a very active social media user and constantly checked into places on Facebook. Among these check-ins were Noida, Gurgaon and East Peoria.

Intelligence collection always feeds on mistakes your adversaries or intelligence targets make, and Baljeet committed a major blunder here. Soon after, we found ourselves digging into Baljeet’s profile. It turns out, that he had frequently visited locations (marked in blue) in the vicinity of the office building (marked in red) which was found in the WHOIS data of the scam sites and in which we suspected the team of developers behind these sites to be located.

At the same time, we discovered that Baljeet registered a website with his full name and actual email address, which we had found in his Twitter timeline. This website was providing web design services to customers in the US. And guess what? The site was using the same address in East Peoria and also the same phone number as the scam sites! I decided to have a little chat with Baljeet on Facebook to confront him with this information.

The conversation went on for a while and Baljeet kept coming up with excuses and dodgy information on why his website was using the same address and phone number as the scam sites. After our talks, he actually took down both for a while, but the address is back on the website today. To us it is clear: Baljeet is definitely involved in this scam. You don’t just sit in India and coincidently use the same address (in the middle of nowhere) and phone number as multiple scam sites on your own site. You don’t just coincidently happen to work in vicinity of addresses we have attributed to the scam sites.

While looking into other social media profiles used by Baljeet, we came across other indicators that fit the overall picture. A couple days ago, he proudly posted the following on Twitter: A certificate of an advanced Google Analytics course (my browser was set to German when I opened the link).

As you can see, the information we won from social media research led to us a prime suspect in our case. Not only did we have a real name now, we also had a unique username, a pattern of life and enough information to verify his involvement in our scam.

Our social media research also led to many more key findings. One of the phone numbers we found in the scam sites’ WHOIS data was also posting job advertisements on Facebook.

A call center agent providing support to activate Norton? That’s exactly how our hunt started. Looks like it is all coming together! The information we gathered throughout our social media research was invaluable and had us dig even deeper into WHOIS data and source code from the scam websites, leading to unravelling the whole network behind it.

Tracing ownership

All roads lead to Rome. Or in our case to the Indian city of Noida and a certain phone number we found when we used Google Drive API and RiskIQ to generate more leads in our investigations of the Norton scam.

Above, we have shown you how we identified a possible scammer, visualized connections, contacted scammers and how they are using SEO to boost the ranking of their websites within search results. Here, we explain how we used several tools and techniques to connect the different domains and information, ultimately leading to a single person that sits in the middle of the web like a spider. Is it Baljeet, who was mentioned in the previously? Or are there more people working on this scam?

To answer that question, we went back to the beginning of our investigations and revisited the URL where it all started, the “nortonhelpus[dot]com” site. The website is not just a page that displays the scam phone number, it also contains a link that leads to a subdomain: “activate.nortonhelpus[dot]com”. Here we found two separate links to Norton software placed on a Google Drive.

The person uploadeding this was apparently “Nancy Wilson”, coincidentally the same name as the person who registered the domain name “nortonhelpus[dot]com”. With a common name like this, there wasn’t a lot we could do. We needed to find out what email address was connected to the Google account used, and who was in control of it.

The email within the Whois information for “nortonhelpus[dot]com” was “nw815624@gmail.com”, and to see whether this was the same email address used to upload the files, we had to find out who the actual owner of the Google Drive was. There are two ways of retrieving that kind of information. The first one involves using Firefox and opening up the Developer mode by pressing F12 after loading the page and then clicking on the Inspector. We searched for “@gmail.com” and were presented with a few options. While going over these options, we found the email address of the user hidden somewhere in the source code of the page.

If the owner’s email address cannot be found in the source code of the site, there is another way to move forward: the official Google Drive API. While using the older v2 version instead of the newer and currently active API v3, we were be able to retrieve information on the owner of said Google Drive, including an email address. The specific endpoint we needed to query was “drive.files.get”. After creating a developer account and a temporary project, and by using the online test form (https://developers.google.com/apis-explorer/#p/drive/v2/drive.files.get), we were able to query the endpoint and retrieve all the metadata of the uploaded files. The email address of “Nancy Wilson”, who we know is a fake persona.

Using the following command on the Google drive we were investigating, it pulls the results shown below.

Command: GET https://www.googleapis.com/drive/v2/files/1RCDQWugm5km0fDufLx60kUL_iPu5qj_L?key={YOUR_API_KEY}

Result:

Finding the email address “nancywilson962@gmail.com” gave us a new entry to start searching online for any information that could lead to the person behind the website. For this we turned to RiskIQ, where we used the PassiveTotal platform to query their massive data set of historical Whois information. Within RiskIQ you have the option to search on domain names, but also on email addresses or phone numbers that are inside the Whois registration information, which makes it extremely suitable to trace scammers via domain registration. When searching for the email address we sadly found out that only one domain was registered, the domain “amazing-wash[dot]com”. It appeared to be registered in India by a person called “Ramesh Kumar”, according to the Whois information he was living in Noida and with all this information we found a new phone number.

Though we hoped to find more domain names with our initial query, we simply pivoted on this phone number within RiskIQ for more mentions of it. And that is when we hit the jackpot.

There were more than a dozen domains connected to this single phone number. Even though we already found multiple domain names by searching for websites mentioning the phone number +1-844-947-4746 or using the address in East Peoria, all this information was connected via Whois information through a single phone number that had direct leads to the person placing the files on Google Drive. It was time to pivot again, this time on each and every phone number and email address that we found in the Whois information of all the domains. Upon finding new information, we retrieved related new domains too and so on and so forth. By using RiskIQ, manual Whois queries and going over the content of the websites, we were massively expanding the list of entities in our investigation. Starting with about 100 entities after our initial Google searches and manual labor, to well over 200 items by using the information within RiskIQ! This is how we ended up with a very large cluster of websites, fake aliases and email addresses, a bunch of new phone numbers, all circling around a person called “Ramesh Kumar”. The person with the phone number that connects more than 75% of all the collected data in this investigation in a direct or indirect way.

Of course, we couldn’t resist and tried to call Ramesh on the phone number we had discovered. After these calls, everything made total sense and we pretty much figured out how this whole network of scammers works.

Putting the pieces together

Gotcha! We found out who is responsible for this massive scam. Using OSINT and social engineering we tracked down the company behind the Norton Scam.

Time to finally unravel the Norton scam. Sector and I have decided to conclude our investigations and put the pieces together, after spending countless hours working on this case. Every time we thought we had figured it out, new information was found, taking us down another rabbit hole. Sometimes we spent days following a lead, just to find out that it wasn’t related to our case at all. As with most investigations, we were not able to solve all mysteries, but we are pretty sure we identified the company and some individuals behind this massive scam scheme.

We pointed out how everything led to specific Indian phone number (+91.9540878969). This number was used to register many of the domains we were looking into. Once more, I decided to make some phone calls to India. I found out that the number belongs to a web design office. The first four phone calls were answered by different men who did not understand English, so they hung up on me. My fifth phone call was more successful. I got a hold of a woman named Priya and told her that a friend of mine had recommend them and that I was looking to have a website set up for me. I had called the right place and I would need to speak to her boss, Priya explained. I also mentioned that the site was to be used as a scam site to obtain credit card data. This too was possible according to Priya. Soon afterwards I had a conversation with the boss, who remained nameless. If I was willing to pay roughly 150$ on PayPal, they would set up the site I needed. With these phone calls, we have proven that the web design office was responsible for setting up the type of scam sites that we have seen throughout our investigations.

During our research, we also came across a site which offered web design services to US customers and to which we had actually found legit websites they had created. This is something very common: using a US frontend to sell IT-services that are performed in India. So, not everything the team did was illegal or scam-related.

In order to promote the scam sites, another team was responsible for search engine optimization (SEO). The SEO team was most likely also located in the offices of the web design team, probably under the same leadership. Their job is to flood the internet with backlinks in order to promote the scam sites. So far, we have found more than 20,000 entries for this cause. From Facebook posts, to Medium blogs, to comments on non-related webpages; a large variety of backlinks were created in the past year.

The purpose of the scam is have the victims call one of the tech support phone numbers. Thus, a team of call center agents is required. Remember how the scam works? If an unsuspecting victim calls the number, they provide ‘assistance’ by obtaining remote access to the victim’s computer. In some cases malware is installed, in other cases they ask for credit card data in order to bill the customers for their service.

These call center agents were hired by a company named 4compserv, which is located at an address that was also used to register some of the identified scam domains. We suspect this is root of all evil, the company behind the scheme. Or at least some employees of the company, since we have also found evidence of 4compserv conducts legal business as well.

More evidence came up, which proves that the web design office and the call center are definitely related. Shortly after I had spoken to the boss of the web design office, I received a phone call from the number linked to the call center (+91.97117613). Unfortunately, I missed the call and haven’t been able to reach them ever since. Furthermore, one of the scammers I had personally texted with recently updated the CV on his website. Have a look at his current jobs:

While there are still some questions to be answered, our research has enabled us to have an overall understanding of the network and the techniques used to run their scam, as well as identifying the company most likely behind this scam: 4compserv in Noida, India.

Along the way, we would often stumble upon funny facts. Some of the scam developers were just so sanguine, they didn’t want to obscure their tracks. Such as the preferred use of the name ‘Nancy Wilson’ to register domains or create sock puppets. The original websites the scammers had set up were very crude, now it seems they are using nice looking WordPress templates, including chatbots. Usually, the chatbot would ask for a phone number, so the scammers can call back. And guess who you would be chatting with on all of these sites? Good ol’ Nancy!

We’re done! We managed to find the perpetrators behind all this. What started with a sock puppet on Medium led to unravelling a largescale scam network, targeting unsuspecting victims seeking tech support. We hope that our project may help counter the threat originating from this specific scam and raise awareness for similar schemes. Also, thanks to many of our readers for sharing the posts from this series on Twitter and LinkedIn, ultimately ranking the articles higher and higher on Google. Using OSINT and social engineering to enable counter-SEO against the scammer’s massive SEO effort!

Now it’s time to relax a bit…before we start the next awesome project!

Sector035/MW-OSINT – 28.08.2019

We explicitly decided to keep the disclosure of personal information on the investigated individuals to a minimum in these blog posts. However, the complete information gathered is available to law enforcement and/or the companies targeted by this scam upon request.

Unravelling the Norton Scam – Chapter 5

Never be too careless with what you post on social media. It might come back to haunt you one day and in the case of our investigations it sure led to some key findings on the scam network!

Chapter 1 – It all starts with a bad sock puppet

Chapter 2 – The Art of OSINT

Chapter 3 – What’s the big deal? And who’s to blame?

Chapter 4 – The more, the better

Chapter 5 – Mistakes on social media

The ultimate goal of our investigation was to find the people behind the scam. So far, we had found fake personas and most of the WHOIS data was misleading as well. Of all the people we came across, there was still a reasonable amount of doubt that they were actually involved in the scam, or we couldn’t reveal their true identities for sure. In our hunt for the truth we also turned to social media, hoping to find someone connected to our case.

Our social media research also led to many more key findings. One of the phone numbers we found in the scam sites’ WHOIS data was also posting job advertisements on Facebook.

Sector035/MW-OSINT – 15.08.2019

Pattern of Life Analysis in OSINT

OSINT isn‘t always about the actual content that is collected. Sometimes timestamps and other metadata can be just as useful in investigations.

In my previous job as a SIGINT analyst I wasn‘t always lucky enough to actually intercept the content of communications. Sometimes encryption methods made it impossible to actually see or hear what people were transmitting to each other. In such cases, the analysis of metadata (e.g. phone numbers, date and time, communication type, duration) became almost as important as the analysis of content itself.

We all have certain patterns of life that define how, when and where we communicate. Furthermore, specific events might also have an impact on our communications. In the SIGINT world, I learned how to read metadata in order to assess a target’s behavior and likely actions. Let me give you two quick fictional examples for this.

1. A mafia group regularly communicates using cell phones. They call each other almost daily, just like you or me would also make phone calls to close colleagues on a daily basis. However, every once in a while no phone calls are registered for a period of 2-3 days. Information received from other sources indicate that during these time periods meetings of mafia senior leadership take place and they all switch off their phones so they can‘t be tracked. So the next time a complete network goes silent, we can assume that there is a meeting again.

2. An employee working in a defense research and development lab is not allowed to use his or her phone during work hours. If this is known, then gathering metadata on this phone could easily reveal on which days the employee was actually at work.

Information like this can be used to create a pattern of life based on metadata. The same can also be done in OSINT. In some cases, it might be relevant to know a target‘s daily schedule. This is where social media such as Facebook and Twitter can be useful. Again, keep in mind we are not necessarily looking at the content of a post or (re)tweet, but at the time it was broadcasted. Maybe even the device it was sent from.

Let‘s have a look at some of my tweets over the course of a few months. Each dot represents one tweet. The x-axis shows the dates and the y-axis the time of day I tweeted something. Based on this information you can see when I would usually started my daily tweeting (green line) and when I finished for the day (red line). So my Twitter activities go from roughly 06:00am to 08:00pm.

Next you might notice a Twitter pause sometime in September (orange box). Guess what happened here?

I was on vacation and took a break from my Twitter account! Therefore, if you were to define my pattern of life based on my Twitter activities, you would easily detect any anomalies. A long pause: maybe I‘m on vacation again. A change of general tweeting times: maybe I‘m in a different time zone. No more tweets during the day: maybe I‘m not allowed to tweet at work anymore. Each time you detect such an anomaly, you should start digging into it and see what the actual reason for this change is.

In the aforementioned example I used Twitter, but pattern of life analysis in OSINT basically applies to any service in which communications or interactions have a time stamp, and that‘s basically every social media platform or message board.

So, the next time you notice I‘m offline for a bit, just wish me a happy and relaxed vacation!

MW-OSINT / 16.07.2019

I2PO: OSINT in Support of HUMINT Operations

In a previous post I explained a concept I named ‘Interdisciplinary Intelligence Preparation of Operations’ and how this could be used to support military operations.

This post will concentrate on the use of OSINT to prepare and monitor HUMINT operations. I will not distinguish between military intelligence HUMINT and sources used by law enforcement agencies or journalists. In both cases, getting access to a source and the preparatory work needed for this are quite similar. Each HUMINT operation starts with the identification and selection of a potential source, thus finding someone in vicinity of our actual intelligence target, who is able to consistently report key intelligence. In the past, even the acquisition of a source was accomplished by HUMINT means. A case officer heard or knew of someone who might have access to specific information and he then talked his way around to finally approach the potential source.

With more and more information being available online, especially through social networks, this approach can be done virtually in some cases. Scavenging Facebook, VKontakte, Instagram, but also LinkedIn and Xing can prove very valuable when searching for potential sources. Of course, this always depends on how outgoing a potential source is on the internet. Sometimes an approach solely through social media could be sufficient, at other times this will not produce any results at all.

The following diagram in theory depicts the steps for OSINT support to a HUMINT case. This scheme is roughly based on the general intelligence cycle with its different stages. We have planning & preparation, collection, processing and evaluation and lastly dissemination covered. In our case the information will be disseminated to the HUMINT operation, which itself will start the whole intelligence cycle over again.

For a better understanding, I have created a fictive case (well, some of it is true…). Let us assume we are part of police special commission in Hamburg focused on the Albanian mafia. The recent shooting of an Albanian national and member of the local Hells Angels, with ties to the Albanian mafia, caused an upstir among different mafia groups operating in the area. So far, no information has emerged on the background of the shooting and existing police sources struggle to provide any intelligence on this topic. The Key Intelligence Questions (KIQ) are ‘What are the current activities of the Albanian mafia in Hamburg?’ and ‘Are there signs of an uprising conflict between different mafia groups?’

Therefore, our special commission has decided to attempt to win additional sources within this network of mafia groups. The higher leadership in a mafia network will not easily cooperate, so someone on the perimeter, with insight into the core, has to be found. Instead of the traditional approach on the streets, we will use OSINT to pave the way ahead of any physical approach.

This leaves us with our initial intelligence objective: Recruiting a HUMINT source within this network to answer the KIQs. Before we start our hunt for sources there are a couple of things we need to know. Who are the key players, do they have nicknames? We should have in-depth knowledge about our targets, e.g. is there target-specific behavior or a specific language used? Having this information gives us a baseline, which we can use to start our OSINT research. Our first step is to identify the known key players and their online profiles. Luckily, most of them are active on Facebook and Instagram and they like showing off their flamboyant life style. Clubbing, exotic cars, girls and champagne seem to be a vital part of the thug life in Hamburg.

This chart depicts the results of the OSINT research on the core network of Albanian mafia in Hamburg, as it is visible on Facebook and Instagram. Now that we have found our potential intelligence targets online, we can survey their activities and figure out who is linked to them. There are many people surrounding this core network, so how can we identify someone who might be worth recruiting as a HUMINT source?

While reading comments to the pictures that these guys post, we stumble upon an individual who constantly idolizes the mafia leadership and their henchman und who frequently asks when he will be a part of ‘the inner circle’. ‘Soon’ is the most common reply and over the course of time he seems to get annoyed. Furthermore, a quick check in police databases reveals that he was registered on minor crimes and was not yet linked to the Albanian mafia. Let us draw a quick conclusion: We have a person with a criminal record, who has contact to senior leadership of the Albanian mafia and is increasingly aggravated on the fact that he is not fully accepted in the organization yet. That sounds like a promising HUMINT source to me!

Keep in mind that this whole procedure, especially the actual HUMINT work done afterwards, takes time. No quick success will come from this. Once we have acquired the source and he is reporting from within the network, our OSINT work does not stop. Now is the time to evaluate the HUMINT information with OSINT. As we have already seen, our targets are very active on social media and this also applies to our source. If our source tells us he had met with one of the bosses on a specific date or time, it could be validated through a Facebook or Instagram post.

One day our source tells us, that in the aftermath of the shooting, the Albanian mafia leadership had met with Chechen mafia leadership the previous evening. At first, this seems unbelievable, as we had assumed that these two groups were currently opposed to each other. One of the Albanian leaders posted about this the following day on Facebook:

This picture not only shows the Captains of the Albanian mafia, but also senior leadership of the Chechen mafia and our HUMINT source. We now know the meeting took place and we have the statement of our source on the topics of the meeting. It is vital that the source does not know we are tracking him and others on social media. We would not want any of this to be staged to back his statements and purposely give us false leads.

This short and fictive case shows how to use OSINT to enable HUMINT and to support HUMINT while an operation is ongoing. Of course, these techniques could also be applied by military HUMINT as well as journalists, as long as the targets and the potential sources are able to be located online.

OSINT supporting HUMINT: Another example of ‘Interdisciplinary Intelligence Preparation of Operations’, I2PO in short.

MW-OSINT / 03.09.2018