Showing posts with label OSINT. Show all posts
Showing posts with label OSINT. Show all posts

Monday, March 4, 2019

Analysing a massive Office 365 phishing campaign


Last week, a friend of mine reached out with a query: a contact in his address book had sent him a suspicious email. As it turns out, it was. In this blog post, we'll have a quick look at an Office 365 phishing campaign, which turned out to be massive. This type of phishing has been on the rise for a while now (at least since 2017), and it's important to point out, as seemingly attacks are only increasing.


Analysis

As mentioned earlier, Office 365 (O365) phishing isn't new, but it is definitely prevalent. A high-level overview of a typical attack is as follows:

Figure 1 - High-level overview of typical O365 phishing
















A typical flow of such an attack may be as follows:


  1. An attacker sends an O365 spearphishing email, likely from a spoofed or fake email address;
  2. The user is enticed to click on the link, or open the attachment which includes a link;
  3. The user will then unknowingly enter their credentials on the fake O365 page;
  4. Credentials get sent back to the attacker;
  5. Attacker will access the now compromised user's mailbox; and,
  6. The cycle repeats: the attacker will send spearphish emails to all of the compromised user's contacts - with this difference, it's coming from a legitimate sender.
This is exactly what happened to a friend of mine: he got sent an email from a legitimate email address, which was a contact in his address book - only the sender never intentionally sent this email! 

Let's have a look at the infection chain.

The initial email

The initial email sent looked as follows:

Figure 2 - "P.AYMENT COPY"












Clicking on the "OPEN" button would redirect you to a legitimate but compromised Sharepoint (part of O365) webpage. Seeing as a legitimate business has been compromised, I won't post the link here. Its web administrators have been notified.


Figure 3 - "Access OneDrive"













The PDF document

Next step is hosting a PDF named "INVOICE.PDF", which entices the user to access OneDrive to view the shared file. If the user were to click on "OPEN PDF HERE":


Figure 4 - "Login with Office 365"















URI: https://happymachineit[.]info/Michael/b4fb042ba2b3b35053943467ac22a370/OFE1.htm

The final landing or phishing page


Finally, clicking on "Login with Office 365" will redirect the user to the final phishing page, which will look as follows:

Figure 5 - Final landing page
















The final landing page is as follows:
https://happymachineit[.]info/Michael/b4fb042ba2b3b35053943467ac22a370/7hsfabvj2b0b9rguzbzw910d.php

When entering credentials, they will be sent off to the attacker, and the cycle from Figure 1 will repeat itself. Note that other scenarios are possible, for example:
  1. The attacker may try to (re-)sell credentials that have been gathered so far on criminal forums
  2. The attacker may send more targeted spearphishes to potentially interesting victims
  3. The attacker may attempt to access other services or accounts using the same user/password combination
In short, there's countless other possibilities.

The phishing infrastructure

Avid readers will have noticed the phishing website uses a valid SSL certificate, which has the following details:


  • Subject DN: CN=happymachineit.info
  • Issuer DN: C=US, ST=TX, L=Houston, O=cPanel, Inc., CN=cPanel, Inc. Certification Authority
  • Serial: 169382499542171049850152621295591104087
The SSL cert was issued by Comodo in January. Details can be found on Censys.io.

An additional email address is connected with "happymachine": fudtoolshop@gmail.com

The phishing website encountered here, https://happymachineit[.]info, is hosted on the following IP: 178.159.36[.]107

Pivoting on that IP brings us to the following SSL certificate details:

emailAddress=ssl@server.localhost.com, CN=server.localhost.com

This means the certificate is a local and self-signed one. In other words, if you are accessing a secure website, and you see "server.localhost.com" as the SSL certificate, do NOT trust it. This is sometimes from an automatic setup from the hosting provider.

As a side-note, a search for the Common Name (CN) mentioned above with Censys currently yields 473 (unexpired certs) results: https://censys.io/certificates?q=%28server.localhost.com%29+AND+tags.raw%3A+%22unexpired%22&

Performing a search with RiskIQ's PassiveTotal as well as VirusTotal, and after filtering results, we obtain a whopping total of 875 unique Office 365 phishing sites, hosted on that IP alone! It appears this campaign has been active since December 2018.

Searching a bit further, it appears the whole ASN (which is a collection of IP prefixes controlled by a single entity, typically an ISP), AS48666 is in fact riddled with Office 365 as well as other phishing sites. Using URLscan.io we can quickly gauge the ASN is hosting multiple phishing sites for Office 365 as well as Adobe:

Figure 6 - AS48666 hosting badness










General Info:

  • Geo: Russian Federation (RU) — 
  • AS: AS48666 - AS-MAROSNET Moscow, Russia, RU 
  • Registrar: RIPENCC

As shown in this blog post, one IP address can host tons of phishing instances, while the ASN controls multiple IPs. Bonus bad IP: 178.159.36[.]120. 


Detection

For the phishing websites itself, any network traffic that resolves to the IP above.

I've noticed there are countless similar PDFs from this same campaign. Due to the way these are created (likely in bulk), a simple Yara rule can be developed as follows:











The Yara rule can be found on Pastebin here or on Github Gist here.

Note: in specific instances, this rule may false-positive - so use at your own will.

The following MITRE ATT&CK techniques are relevant:



Disinfection

There isn't much to disinfect, since there's no actual malware involved.

However, if you have been affected by this phishing campaign, do the following immediately:

  • Contact your network and/or system administrator or managed services provider if you have one and wait for their response - if not;
  • Note down the phishing page/URL, then close any open phishing pages - in fact, close the whole browser;
  • Perform an antivirus scan with your installed product, and a scan with another application, for example Malwarebytes (better be safe than sorry);
  • Change your O365 password immediately;
  • Change passwords on other websites where you used the same combination;
  • Reach out to the people in your address book you were compromised and they are not to open your email(s) or at least not any attachments or links from your email(s);
  • Verify your "Sent" emails folder (or "Outbox") for any suspicious activity. If there are no Sent emails - the attacker may have deleted them, or you may have a full compromise on your hands.;
  • Verify any (newly) created rules in your mail application (in this case O365), for example, verify there are no new forwarding rules or perhaps rules that delete new incoming emails - forwarding rules and deletion rules are sometimes set up by an attacker to gather more information or as an attempt to remain hidden; and,
  • File a complaint with your CERT, local police station, or whichever authority would handle such cases. If you are unsure how to do so, have a look here for assistance.


Prevention

  • Block the IP (or whole subnet 178.159.36[.]0/24) mentioned in this report in your firewall or proxy or other appliance;
  • Use strong and preferably unique passwords (use a password manager);
  • Set up 2FA for accounts or, preferably, MFA (multi-factor authentication);
  • Enable, deploy or implement anti-spam and anti-phishing protection;
  • Enable, deploy, or implement a URL phishing filter;
  • Trust, but verify: "did this contact really need to send me a "Payment Copy"? - if needed, verify via a phone call - not via email;
  • Be generally cautious with links and attachments. Do not click on links or open attachments from unknown senders;
  • If possible, use Firefox with NoScript enabled; and,
  • If you're in an organisation: create or organise user awareness training.

Conclusion

Phishing has been around for a long time - Office 365 phishing, on the other hand, has been around since, well, Office 365 was created. Every time a new service is created, you can imagine that phishing emails targeting that service will follow - maybe one month later, perhaps a year later - but they will.

Always try to be vigilant and follow the prevention tips mentioned above to stay safe.

As a side-note, the real Office 365 page is: https://outlook.office365.com/owa

You may find more information in the Resources section below.

Resources

Blaze's Security Blog - Cybercrime Report Template
Decent Security - Easily Report Phishing and Malware
Microsoft - Anti-phishing protection in Office 365
Microsoft - Microsoft publishes guidance to boost public sector cloud security
Microsoft - Set up multi-factor authentication
Microsoft - Set up Office 365 ATP anti-phishing and anti-phishing policies

Indicators


Thursday, September 21, 2017

Malicious ad/click networks: common or forgotten threat?


Introduction

Malicious ad/click networks and ad fraud are not entirely a new phenomenon, but it is important to realize the kind of threat it may pose. Is it a common, or forgotten threat? Maybe both.

In this blog post, we'll take a look at how a seemingly innocuous click network and advertiser, is actually showing some rather malicious behavior.



The beginning

It all starts with the following redirect:

Figure 1 - .js download

A 'critical Firefox update' needs to be downloaded and run, with the resulting file having multiple layers of obfuscation. After deobfuscating 2 layers, we get the following:


Figure 2 - malicious script

The script will attempt to download an .flv file from ohchivsevmeste5[.]com, with additional parameters. While I was unable to reproduce what happened afterwards at time of writing, it would likely fetch another heavily obfuscated JavaScript, for clickjacking purposes (and this behaviour can also be deduced from Figure 1, as it persists in the browser).

Clickjacking is not an uncommon phenomenon unfortunately, and is described by Wikipedia as:

Clickjacking (User Interface redress attack, UI redress attack, UI redressing) is a malicious technique of tricking a Web user into clicking on something different from what the user perceives they are clicking on, thus potentially revealing confidential information or taking control of their computer while clicking on seemingly innocuous web pages.
Basically, what you see is not what you get. In this case, a Firefox update is anything but said update.



The persona

The domain mentioned in figure 2 is hosted on 192.129.215[.]157, which has a ton of other domains, with most of them appearing random. One of these domains is aiwohblackhatx[.]org, which is registered to a person with email address of abdelrahman.a.y.127@gmail[.]com.

This email address is linked to a Facebook page, mltaqaalwza2f2, which claims to be a jobseeker website:

Advertisements for newspapers, newspapers and websites The largest site jobs on Facebook for Jordanians only

Until recently, that page listed a contact email address as ADMIN@ULTIMATECLIXX[.]COM. This has now been changed, and unfortunately I was unable to take a screen capture at that time.

Ultimateclixx sports a flashing website, promising instant payment:

Figure 3 - Ultimateclixx website (at time of writing)

Passive DNS data reveals that the email address mentioned above, has links to other domains, and in particular to a person or persona called 'Mohammed Farajalla'. This persona has multiple email addresses set up, all pointing to click networks:

Figure 4 - Persona (click to enhance)

This initially lead me to believe that Abdelrahman and Mohammed are the same person, and is simply an alias. However, WhoIs data from another domain, aifomtomyam69[.]org, reveals a person named 'Abdelrahman Farajallah' as domain owner.

Brothers?

Figure 5 - Ultimateclix admin forum post (click to enhance)

Brothers or not, it seems that Mohammed is the 'public face' of the company, and Abdelrahman works in the background, registering domains. A well-oiled business scheme, apparently.

In Figure 5 above, you can see a specific post from 'mhmadfarajalla', hereafter referred to as Mohammed, explaining how he joined Goldenclixx from 2013 onwards, and made quite some investments. This reply was motivated by a user questioning their legitimacy:

Figure 6 - concerns in regards to the Ultimateclixx admin (click to enhance)

This was posted in 2015 on the eMoneySpace forum, which is a website created to 'promote or talk about internet money related subjects'. Basically, how to earn money online using ads, which is completely legal.

Link to topic can be found here:
https://www.emoneyspace.com/forum/index.php?topic=361872.0

I've also set up a mirror here: https://web.archive.org/web/20170918213144/https://www.emoneyspace.com/forum/index.php?topic=361872.0


It appears Abdelrahman and Mohammed have been involved in this scheme for a prolonged period of time. While they may have initially started their project or business as a legitimate way to make money, this has definitely shifted. They are likely located in Palestine. (see also his/their Twitter account, and make your own deductions.)



The infrastructure

Earlier, I mentioned that domains involved seemed random. Have a look at these domains:

aidixhurricane[.]net
aifomtomyam69[.]org
aingucareersearchings[.]com
aiteobutigim[.]net
aiwohblackhatx[.]org

Notice anything particular? If not, what about the following domains? (includes our initial example)

ohchivsevmeste5[.]com
ohighzapiska[.]org
ohlahlukoil[.]org
ohmuogoodlacksha[.]com
ohseltelists[.]org

To clarify, the 2 first characters are the same for a whole set of domains, while the rest does appear to be (at least semi-)random. Just as a visual aid, here are a few other domains:

vaewedashrecipes[.]net
vahfebankofamerica[.]net

iechow3blog[.]org
iefaxshoeboxapp[.]net
iegiwrealarcade[.]org
iehohclock-world[.]org
iengeluxauto[.]com
ieweimz16[.]org
ieyonthesubfactory[.]net

You may have noticed vahfebankofamerica[.]net in there, it is relatively newly registered (2017-09-12), by a 'Megan Quinn', with email address of qum65@binkmail[.]com. I doubt an actual Ms. Quinn would use this email address. It may try to convince users of its legitimacy, alluding it is part of Bank of America's website. However, nothing could be further from the truth.

Interestingly enough, the domain has hosted a JavaScript at least once, with a familiar pattern:
https://vahfebankofamerica[.]net/3558451080485/150526622449473/firefox-patch[.]js

No doubt this website would also prompt you to download a 'critical security update'. Do I hear redirects mixed with ad or click-fraud and clickjacking? Who doesn't love the smell of that in the morning?

You would not have guessed, but a lot of email addresses seem randomly generated, as well as their personas. Another example includes:

Figure 7 - ohchivsevmeste5[.]com WhoIs info

Everything in the WhoIs info is fake. Not to say an Allan Yates doesn't exist, but he has nothing to do with any of this - rather he was just unlucky. This may have been the result of a 'fake name generator'.

We now have several IP's hosting a bunch of semi-random domains, set up for redirects, ads and likely clickjacking. The initial research started with 192.129.215[.]157, and expanded to/had links to 192.129.215[.]155.

Of course, I decided to take a look at 192.129.215[.]156. Surprise! It turns out it's just as bad. Maybe we should just block the whole 192.129.215.0/24 subnet?

Some more takeaways from the infrastructure - most of the domains are or have:

  • WhoIs Guard Privacy Protection/Privacy Protect:DomainsByProxy;
  • Behind CloudFlare;
  • A valid SSL cert, issued by Comodo.



What else is there?

Earlier, Mohammed boasted in his forum post about adzbazar.com, one of his new websites, and absolutely 'no ponzi scheme strategy'.

Some of the related websites provided by Mohammed are:

Figure 8 - adzbazar[.]com

Figure 9 - clikerz[.]net

I think you can start seeing a pattern here.

Circling back, with our email address abdelrahman.a.y.127@gmail[.]com, I noticed another registered domain, adz2you[.]com. You could zay zomeone likez the letter Z. Either way, using PassiveTotal's host pair functionality, we can find a hostpair with gptplanet. Gptplanet claims to:

Earn money by completing simple tasks online. Everyone can join, it’s absolutely FREE!
As far as I could see, this claim is indeed legit. As far as Mohammed and all his domains go: it didn't take me too long to dig up a forum post with an on-the-point title:
SCAM LEXIADZ ADZBAZAR

Link to topic can be found here:
http://www.gptplanet.com/forum.php?topic=18119

I've also set up a mirror here:
https://web.archive.org/web/20170920001823/http://www.gptplanet.com/forum.php?topic=18119

The topic on gptplanet also references to eMoneySpace, a forum mentioned earlier, and specifically, several topics are set up about 'referral' websites our dear friend Mohammed has set up.

One user made an excellent remark:

It comes to my realize for a whole year now, is that some people are implementing some malicious scripts onto their advertised ad for a user to click on it.

Nail, head, hitting it. It seems that Mohammed isn't done yet however with both scamming people, and infecting users:

Figure 10 - Offers4all invitation

There is also mention of another person, 'Agony'. This nickname may refer to Abdelrahman.



Prevention

Prevention in this case is rather short, so here goes:


  • Install an antivirus;
  • Keep your browser up-to-date;
  • Install NoScript if you have Firefox;
  • Install a 'well-rounded' ad-blocker, for example uBlock Origin (works with most browsers).

And, where possible, browse the internet with caution.

Note that the campaign on 192.129.215[.]157 remains highly active, and as such, it is recommended to block or blacklist it, as well as the other domains and IPs, provided at the end of this blog post.



Disinfection

When you get a prompt for download or running a 'Firefox patch' similar to above, or any other pop-ups for that matter - where you not instantiated a download yourself - cancel the download or, if not possible, kill your browser's process. This can be done via Task Manager for example.

While I haven't seen any evidence of other malicious behaviour besides ad, click-fraud and clickjacking, it is recommended to:


  • Uninstall and reinstall your current browser, along with its extensions;
  • Perform a full scan with your installed antivirus product;
  • Perform a full scan with another, online, antivirus product, or with Malwarebytes;
  • Change your passwords.


Additionally, you may check in your firewall or proxy logs, if there was any connection at some point with any of the domains or IPs provided in the Indicators OCompromise section below.



Conclusion

Ad fraud, clickjacking, ad networks, .... There are tons of similar networks out there. While ad networks are usually not malicious, other possibilities exist , such as:

  • An ad network is compromised;
  • An ad is compromised;
  • The ad network is malicious in itself.

Probably, at some point, there should be better security controls for ad networks, in order to prevent an attack or campaign such as the one described in this blog post. Proper security hygiene is necessary for the ad networks, but just as well for any website that serves up ads.

Clickjacking and click-fraud is a common and a very real threat. Are you watching your logs, and acting on them?


IOCs