Thursday, September 22, 2016

Hopefully a sign of new things to come...

It's been a busy year this year.  In my day job working in food security, I've been to Taiwan, Arizona and South Korea.  I've met a lot of people who want to help solve some really really big problems that literally affects billions of people.  It's been rewarding to see this year unfold, if a little challenging.

In my spare time, I've also had some rewards.  As most people know, I've had nearly two decades of challenges with one of my banks.  Well, something interesting happened.

Traditionally, the customer/bank relationship looks like this:

It's not a productive loop, and it's prone to issues.  For instance, I've experienced "We're looking into your problem" when I've not actually stated what the problem is yet.  (In programming, we call this a "race condition").  Another problem is that if you are the one reporting, there's the sensation that things disappear into a black hole as you never get feedback.

But, that was all I had for 18 years.

About a month ago, something I saw emanating from the CIBC twitter team that was obviously "incorrect from a technical standpoint" annoyed me.  On that day, I was going to be downtown, so I thought I might as well just break the "Report" -> "Thanks" cycle and walk into the bank with the solution to the issue.  Long story short, my "let's just cut to the chase" style wasn't met with the same enthusiasm.

I have no idea what switch flipped that day within the bank, but after 18 years, we changed to this method of communication:

Now, instead of the uncertainty of whether technical messages are getting through, or are distorted in a game of "broken telephone", the bank was asking what I needed? 

That's a very simple question for me - If someone gives me a room full of technical and policy people that I can speak unimpeded and natively to in order to explain a) what I can see and b) where I can see it, then I can get effective feedback (anyone that's ever worked with me knows I don't like red-tape or unnecessary delays) as to whether what I'm saying is even being understood, as well as the added bonus that you can hold a Q&A session to clear up any loose ends. I think that in 30 minutes, I offloaded more information about customer optics and technical issues than I have since the "triple-Interac-debit whilst saying money was not debited" debacle of 2013.  

Another good thing that came out of this is the customer service person I normally deal with (and probably frustrate to no end) was also in on the conversation.  In an age where KYC ("Know Your Customer") is a buzzword in banks and large organisations and doesn't actually mean that they "know" who you are, I think it was good that the frontline customer service person who normally has to deal with me could see me in my native habitat.  Instead of being the keyboard warrior on Twitter that they know and recognise, they could the bigger picture when this wound-up-spring was finally let loose.

We didn't agree on everything (for instance, 9:59am to 1:49pm is still basically 4 hours in my book), but I think it was a productive use of my time and their time.  I was told new things that nobody had explained before, and they were told things that I see as problematic.  

Time will tell if things actually get cleaned up security wise, but my guess is it will because now people know the extent of what can be inferred outside the bank, and most crucially, know that people outside the bank know.

Thursday, August 25, 2016

Securing the news...

I read an interesting article today, about how the news outlets are getting targeted by hackers.

This doesn't surprise me for a number of reasons:
  1. Most news today is not an unbiased account of fact, but rather it takes sides and therefore it becomes a polarizing factor that can raise ire amongst some people.
  2. News media outlets today are not used to securing news, so something that is unprotected can be adulterated and changed into something it was never intended to be.
In Canada, we have two major corporations that control most of the news and media in general.
  • Bell Media 
  • Rogers Media
Regular readers of this blog will realise something:  This is Bell Canada and Rogers, the two monolithic cellphone and internet providers in Canada.  These two organisations are traditionally not good at security with their own customer facing systems, so you can imagine what type of security I have noted has been imposed on their media divisions.  

I did a quick 2 minute scan to see what I could find security wise....

What I found was this:  The media divisions are open to the exact same methods of being compromised as each parent company is currently known to be.  This percolates down through each media property such as a radio/tv station website, or newspaper site.  So, in short, it's possible to adulterate the news in Canada.  

We all know the phrase about those who control the message control the people - and that's what makes the media a target for hackers.

Wednesday, July 6, 2016

Moving Low and Slow In The Security Theatre.

As most of my regular readers know, I spent a bit of time over recent years fussing over my banks and Canada's financial institutions in general.  

For those that don't know what's going on, a quick recap...

Things started off with me gradually losing faith in my primary bank's ability to maintain a secure banking experience because of a series of events spanning a few years that highlighted to me that something was awfully awry, and things degraded slowly into more problems that were also spotted in my secondary bank.  Later, I found similar failures across many banks in Canada and eventually found the government at risk, too.  This culminated in the Spring of 2016, when I coordinated with one bank on the issue, and then successfully raised the alarm that basically most of the entire country of Canada was at risk and the RCMP leapt into action.   

Whilst all this was going on, I had to make sure to never do anything that amounted to, or could be construed as hacking.  This is actually very easy - and here's how it happened.

This is a Bentley convertible.  

(Click for bigger image)

I've never been in this particular car shown in the picture.  I can probably guess accurately that neither have you.  However, both you and I can probably agree that the roof is down on this car, and that if (hypothetically) we were in this car with this roof position and it started raining, we'd get wet.  The reason we know this without ever entering the car is simply because we understand what we're looking at.

Same applies to the banks and the Canadian government.  I can look at CIBC or ScotiaBank and without even logging into them, plainly see how they can be compromised because I understand what I'm looking at.  Same thing at the Canadian Government...

Often hackers are caught because having breached a system, they bang about inside, tripping monitors as they scan ports, probe and push systems trying to fumble about looking for the proverbial pot of gold.  

We see banks respond on social media about this type of threat, such as shown here:

The problem with this, as you can guess, is these measures only apply to hackers that break into a bank.  In my case, there was no hacking into any banks, no entry to any bank systems, and yet everyone at the law enforcement level is onboard with me because they understand what they're looking at.

Understanding technology like this is a variant of the "Low and Slow" method of hacking.  I say "variant" because whilst it shares all the traits of the "Low and Slow" method of hacking, there is no "hacking" here.

Additionally, it has to be pointed out that operating outside of a bank or government in this manner shows up something else; It's not security. It's "theatre". If you watch the show long enough, you start to see the props and the set moving about.  That needs to change.

I'll leave you with one last thought:  I'm just one guy who only wants his bank to not put him at risk, and with limited time on my hands, I figured out something that affects the entire country.  There are likely cadres of criminals out there figuring this out on a daily basis and, logically, they must go undetected as the banks cannot see them.

Monday, June 27, 2016

Verification versus Authentication

On my Twitter feed, I see there is a continuous trickle of tweets asking for CIBC bank to adopt two-factor authentication ("2FA").  It looks like this:

You normally see about 2 or 3 of these a week - and they all originate at the same site about 2FA.  As you can see, after the request to consider that the bank adopts 2FA, there's a canned response that goes like this "We take security very seriously, so we have two step verification".

Of course, that irks me, and here's why:  The customer is talking about authentication (i.e. making sure the person accessing their bank account is the correct authorized person who should be accessing the account), and the bank is responding on the subject of verification. In the case of a bank sending a code to a phone number on file, all the bank is verifying is that regardless of whether they're authorized or not, the person trying to access that account also has the phone belonging to the person who's account is trying to be accessed. 

That's a fundamental flaw in security.

If you don't know the difference, two step verification is where you supply a password, and the bank sends you a code and you type this in as well.  So, imagine your other-half has your phone and you're in the middle of a messy breakup, and they know your password, the bank sends a code to your phone in their hands, and voila!  

There's a really obvious problem here, and anyone with an ounce of security savvy will tell you, physical access is 9/10ths of the problem.  This is why people are asking for 2FA.

With 2FA, you have to supply something in addition to the password.  This usually is two items out of this list of three:
  • Something you have (eg password)
  • Something you know (eg maternal grandmothers maiden name)
  • Something you are:  (eg biometric, location, etc).
It doesn't have to be those three, but they are the most common.

As you can see, the response from the banks totally undermines any confidence that they even understand what's being asked, because in the situation pointed out above, the bank is providing the tools to the attacker to complete the compromisation of the customer.

Of course, the access agreement is written with a totally one-sided assumption that the customer is the only person who could ever put the bank or the customer into jeopardy.

(click for bigger)

The part that says "Without limiting the generality of the first sentence in this Section 9," makes me shake my head, because of course, the first sentence says that you are on the hook for "any losses" whilst ignoring the logical reasoning where as often happens, the bank has set the customer for failure in the first place.

In a nutshell, the security situation can is analogous to going into the sea to scuba dive and the dive master says "There are sharks here, and we take your security seriously, so here's some fresh raw beef steaks to hit the Sharks with", and then having setup the divers with the tools to be eaten, has them also agree to an agreement which is totally one-sided and places on blame on the customers.

And people wonder why I banned the CIBC mobile apps from my house?

Tuesday, June 7, 2016

An Overview of Canadian Bank Credential Phishing

In Canada, we have our fair share of bank account phishing.  Primarily, these scams mostly originate from two distinct teams, and each team has it's own trademark way of operating.

In one corner, we have Team Asia.

  • They register a proper domain.  
  • They set up their own DNS and make the site look like a proper clone of the real bank site.
  • They are sometimes able to operate for a few months before they get taken down.
  • They send invites from the SMS code 7000 (formatted as 700-0).

In the opposite corner, we have Team Russia.

  • They don't register a proper site, preferring to hang off the back of an existing site.
  • They don't set up their own DNS, preferring to use the service.
  • They get taken down quickly, so are much more proliferate.
  • They send text messages from full phone numbers, usually in Alberta, Ontario or British Columbia.

There are a few stragglers that I haven't assigned to one group or the other, but one particular code base does show up in a number of these, which means they're either the same person/group, or they're buying templates from the same source.

To give you an example of Team Russia:

Here's the SMS from area code 250.

As you can see, it's rather sloppy in comparison to Team Asia.  The final link hangs off a Brazilian site, seen here:

URL aside, the site looks real, until you try to navigate, at which point you run into this:

And for anyone interested in the data, here that is (click for bigger version).

As you can see, this isn't complicated at all, and that's a good thing because most of Canada's banks have online banking with holes that are not secure enough to guard against anything much more aggressive than this.

So, there you go.  The state of Canadian bank phishing in one quick post. 

Wednesday, June 1, 2016

Online Banking and Hosts File

Over the years, I've had my fair share of runaway data.  This is usually caused by Bell Canada, as they resell your data to third parties (if you want privacy, you have to pay Bell an extra fee of $2 per month), and once that data has left Bell, it's going to run and run as it passes from marketing company to aggregator to directory service to marketing again.

As a result of Bell Canada and the three year battle to get a data noose around them, we have a strange win-win situation; Bell Canada sells my data over and over to the marketing people so they get their money, and the buyers (marketing people, directory services, etc) then scrub my details from the incoming data.  That stopped the runaway data in it's tracks. As a bonus, I left my old residential data that Bell leaked some years ago online, so now it optically looks like Bell puts out stale data about me.  It's a wonderful system and works really well.

For this post, I'm going to cover how I tidied up a similar problem a while back with online banking.  When I log in to my two banks, I want to be dealing with the bank and the bank alone, not sharing my purchasing habits with the bank through a third party, or even know that any third party is tracking me at the bank and then reporting that to some computer hardware store 30 minutes later.

Both my banks use Omniture/Adobe Analytics.  Right there, you got the holy trinity of data sharing going on.  Between the two banks and Apple (all the iTunes and Apple store run through the same system), the amount of back and forth of data would be astonishing.  So, a while ago, I did something about it as a result of trying to work out why a password issue (unrelated) was giving me so much hassle.

I ended up just driving all the junk requests for tracking and analytics, and marketing to localhost ( Yes, I could opt out of some of this stuff (the banks don't offer you the ability to opt out, but if you track where the banks send this stuff, you eventually end up at Adobe and THEY have a link that allows you to opt out), but that just sets a cookie in your browser, and as a developer, I'm resetting my browser cookies and environment more frequently than the average person, and that would opt-me-in again.  

So, my solution was just hack this stuff off at the knees by permanently editing my hosts file.  If anyone sends a page asking my browser to go talk to Adobe Analytics so it can generate another damn survey or indirect piece of targeted marketing, it will now go into a dark hole and never be seen again.

NOTE:  Only change your hosts file if you know what you're doing. I'm not going to tell you how to change the file, as I don't want to be responsible for what you might break. I'm just telling you what I did. YMMV.

So, what entries are in my hosts file? 

The Scotiabank entries look like this:

#ScotiaBank Changes
#Callback with Trusteer DMG.
#Omniture Profiling & Tracking
#Oracle Maxymiser, marketing & optimisation

The CIBC entries look like this:

##CIBC Related Changes
#Omniture Profiling & Tracking
#General Adobe Hidden Profiling
#Oracle/TAG merchandising, marketing and live chat
#More Marketing, targetting & feedback.
#Even More marketing.

And there you go.  

It works for me.  Your mileage might vary.

Thursday, May 26, 2016

Anomaly with City of Toronto site security...

I ran across an anomaly on the City of Toronto website today....

They go to pains to point out that you're on a secure site in a weird "in-ya-face" way.  This is normally a red flag for me. People make a point of pointing out something when they believe your mind needs to be changed.  For instance, I'm told my burgers are 100% beef because someone in marketing thinks we have ideas that maybe it's not.  I'm told the latest cars are much more fuel efficient, because someone thinks I might have other opinions on this.  Now I'm being told in big message boxes that something is secure.  Mental red flag.

OK, let's suspend disbelief for a second and imagine that it actually is secure - where is the browser padlock?  

(Click for bigger version)

I clicked the "Close" button and then got to the next screen, where again it points out that you are in a secure site on the first line.  Again, there's still no browser padlock.  I also noted the usual bank-trade trick of proclaiming things are secure and placing the onus on the user to make sure they've done their bit for security.... even though the padlock is still missing.

(Click for larger version)

So, I took a quick look at the certificate. The first thing you notice other than the City of Toronto is apparently in Macau, is the verification error that this is a self-signed certificate - so this is no more valid that if I generated one here on my computer and delivered it on a USB stick to City Hall. I've highlighted both items in red.

Jasons-2013-MacBook-Pro:~ jasoncoulls$ echo ^d | openssl s_client -connect
depth=3 C = US, O = "VeriSign, Inc.", OU = Class 3 Public Primary Certification Authority
verify error:num=19:self signed certificate in certificate chain
Certificate chain
 0 s:/C=CA/ST=Ontario/L=Toronto/O=City of Toronto/OU=Technology Infrastructure Services - macau-w1/
   i:/C=US/O=VeriSign, Inc./OU=VeriSign Trust Network/OU=Terms of use at (c)10/CN=VeriSign Class 3 Secure Server CA - G3
 1 s:/C=US/O=VeriSign, Inc./OU=VeriSign Trust Network/OU=Terms of use at (c)10/CN=VeriSign Class 3 Secure Server CA - G3
   i:/C=US/O=VeriSign, Inc./OU=VeriSign Trust Network/OU=(c) 2006 VeriSign, Inc. - For authorized use only/CN=VeriSign Class 3 Public Primary Certification Authority - G5
 2 s:/C=US/O=VeriSign, Inc./OU=VeriSign Trust Network/OU=(c) 2006 VeriSign, Inc. - For authorized use only/CN=VeriSign Class 3 Public Primary Certification Authority - G5
   i:/C=US/O=VeriSign, Inc./OU=Class 3 Public Primary Certification Authority
 3 s:/C=US/O=VeriSign, Inc./OU=Class 3 Public Primary Certification Authority
   i:/C=US/O=VeriSign, Inc./OU=Class 3 Public Primary Certification Authority
Server certificate
subject=/C=CA/ST=Ontario/L=Toronto/O=City of Toronto/OU=Technology Infrastructure Services - macau-w1/
issuer=/C=US/O=VeriSign, Inc./OU=VeriSign Trust Network/OU=Terms of use at (c)10/CN=VeriSign Class 3 Secure Server CA - G3
No client certificate CA names sent
SSL handshake has read 4853 bytes and written 636 bytes
New, TLSv1/SSLv3, Cipher is AES128-GCM-SHA256
Server public key is 2048 bit
Secure Renegotiation IS supported
Compression: NONE
Expansion: NONE
No ALPN negotiated
    Protocol  : TLSv1.2
    Cipher    : AES128-GCM-SHA256
    Session-ID: 00000D2AD210CD02429618321FAA1D60105219435858585857471DBB000016D0
    Master-Key: DDEF69EA999B9A7FA1AAA0AE94D5F3A00F5F9FF5F47D854BAC1621B018CC81751C17BCE855B6FE497AD99445EAE5830C
    Key-Arg   : None
    PSK identity: None
    PSK identity hint: None
    SRP username: None
    Start Time: 1464278459
    Timeout   : 300 (sec)
    Verify return code: 19 (self signed certificate in certificate chain)


So, there you have it.  Whilst the city tells you it's secure, both the browser and the openssl tool verification gives errors or refuses to show the padlock.