Startup Replacing Passwords With Patterns (Shayype)

By: Brian Rhodes, Published on Jun 28, 2017

This startup, Shayype [link no longer available], aims to eliminate passwords, replacing them with patterns.

Problems with passwords are clear, as simple passwords, re-used passwords, brute force attacks on passwords, etc. create significant problems.

But can 'patterns' fix this problem?

Inside this report, we share our test findings on Shayype's approach.

**** *******, ******* [**** ** longer *********], **** ** eliminate passwords, ********* **** **** patterns.

******** **** ********* *** clear, ** ****** *********, re-used *********, ***** ***** attacks ** *********, ***. create *********** ********.

*** *** '********' *** this *******?

****** **** ******, ** share *** **** ******** on *******'* ********.

[***************]

Uses ********, *** *******

****** **** * ***** password ** ***,*********** * ****** ** changing ******* **** ********** to ***** ** *** display. ** ***** *********, the **** ***** ** the ******** ******* ********* the ***** ** * sequence ******* **** **** previously *******.

*** ********* ***** ***** example ********* *** ***** of *** ******** ***** composing * ******* ********:

*** *** ** *** pattern's ********, ** **** future ********* **** *** entirely ********* *******. *** ********* refreshes / ****** ******* every **** ********* **** the ****** ** *****:

******* ****** **** *** matrix ****** '*.* ******* possible ******** ** *** 6x6 ****** (** *** use *** **** *****)', and **** ** '**** if ******* ***** *** same *******, **** *********** of ******* *** ****-**** will ** *********', ********* in * ****** ********, single *** '****' ******* of * ********.

********** * *******, ** forgetting * *** *** that ******, ******* ******* or ******** ******* ** reset **. ** *** case ** *******'* **** website, **** **** * reset **** ***** ***** chose * *** *******.

Overview *****

** *** ***** *****, we **** *** ******* ***** and *** ** ** ********* than ******** ***** **************:

Wistia video player

Claimed **********

******* ****** **** ******* one **** ***** ****** be ****** *** ********** patterns ** ******* *** not **** ** *****, their ****** ** **** secure. ** ******** ** the ******* ******** *** password, ******* ****** ******* other ********** **** ******** passwords, *********:

** ****** '*****': **** *******, ***** do *** **** ** touch *** ******, ** fixed ******* ******* ** enter * ****, *** only *** *** ******* with ***** **** *** enter ** ***** * separate ******. **** ** different **** *********** ******* where **** ****, ********* use ** **** ******** buttons *** ****** ****** or ****, '*******' ********* threats ***** ******* *** used *** ***** ******. Shayype's *** ****** *** a **** ** ******* buttons ** *** **** that ** *** **** someone ******** ** **** of *** ******* ***** used.
****-**-*****: ** *** ********* values *** ******* * - *, **** ***** is ******** ******** ***** on *** ******, ******* complicating *******. ** ******* sees * **** ***** entered, ** ***** *********** be ******* ********* ********.
******* ************: **** ** ***** prefer ****** ********, **** 'L- ******' ** '**** corners', *** ***** *** pattern ** ******* **** also ** *****, ** moving **** *****-**-****, ** leaving * ***, ** pressing *** **** **** twice **** ** *** obscurity ** * ******** roster ** ******. ********, key ******* ** ****** passwords *** ******* ** themselves, ** *** ****** they ******* ** **** symbolic ** * ******** pattern *** ********* *********.

Online ****

*** ******* ***** *'***** ** *******' ******* ***** *******. **** users ****** ***** *********, Shayype ******* ** ******* page ********** *** *** *** how ******** *** ********** *** *** *** system ***** **********.

***********

*** *** ******* ********, Shayype's ****** **** **** drawbacks, ********** *** *** physical ******** ******. *** top **** *******:

** ******* ***: ******* ** ***** an **** **** * very ***** ***-****** **********, and *** *** *** been ******* ** ******** in ** *** ********. While * ********** ****, the ******** ***** *** history *** *** ******* lacks ********** ** ********** use *** **********.
******** ***** ********: ***** **** ********* to ******* **** ****, Shayype ******** *** ***** be ********* ** ******** out *** ***** ** unauthorized *****. **** ***** that ************* ******** ** codes *** ** *******, but ********* ********** ***** can ***** ***** ***** codes.
*****-***** ****: ******* ** ***** developed ** *** **-********** from ******* ******** ******** or *************, *** **** experience ** ********* *** communications [**** ** ****** available] *** *** ***** a******** ********* ******** **** Philips, *****, ***, *** T-Mobile. *******, ******* *** have *********** ********** ******** new ******** ** ****** or ********** *** ******** through ******* ******.

Vote / ****

Versus ************

*** ******** ******** ************, Shayype ***** ** ** interesting *** *** ******** access ******* ** ******* for *** ** ****** readers. ****** **** ******* stagnant **** **** *** quickly ** ********* ** observed ** ************ *****, Shayype ***** ******* '*********** security' ** ******* ** using * ********* ****** of ********** ***** *****.

**** '**** ********' ******* that ********* ***** ******* are ****** ******* **** still **** ****** ***** stolen **** ********, ** the ***** ** *** code **** *** ****** nor **** *** ********* the ******* ******. ****** forms ** ************ **** like *****, ** ***** in *** ***** ******* ****: ******* ****:

******* *********** ******** *** security ** **** ** not ***** ***** ********* at ***, ******* ************* a *********** **** **** of ****** **********.

Still ***** ** ******

** *** ********* *** Shayype's *******-***** ****, *** the ****** *** ******* needs ** *********** ** can ** **** **** effectively *** ********, ***** the **** ** ***** record *** **** ********.

Comments (16)

Undisclosed End User #1

07/06/17 03:01pm

I know that a canadian company Cryptocard had something similar to this. I thought it was patented. Now Cryptocard was sold to Safenet, which is now part of Gemalto. I wonder if these people will get sued. :)

Brian Rhodes

IPVMU Certified | In reply to Undisclosed End User #1 | 07/06/17 04:22pm

That's a good find!

Here is a video of Gemalto's GrIDsure:

The similarities here are strong. I'll ask Shayype to comment on where/why/if their system is different.

Brian Rhodes

IPVMU Certified | In reply to Undisclosed End User #1 | 07/07/17 05:58pm

I emailed Jonathan Craymer, one of the developers of Shayype, to ask about the similiarities between it and Gemalto's GrIDsure. His response:

"I was the originator of Gridsure (Wikipedia https://en.wikipedia.org/wiki/GrIDsure ) with Stephen Howes, who I hired in as a jobbing technician to do some programming.

However Shayype is a far more developed version, offering scalability for millions of users and the addition of a high security version.

You’ll also notice from the Wiki text that a mathematician Mike Bond criticised Gridsure because in theory a hacker who records the grid display and the characters typed in say 2-3 times could perhaps work out the user’s pattern.

For this reason we’ve created a high security version (Shayype HSS) which doesn’t use the device’s keyboard, preventing a hacker knowing which characters have been typed in. This means in our view that an attacker would have to record/film the login operation several thousand times – which of course they’re unlikely to do."

Essentially, Craymer says that Shayype's separate keypad/ not matrix screen touchpad and obscured displayed inputs (*, not a number) make observation more difficult than the Gemalto version.

Undisclosed #2

IPVMU Certified | In reply to Brian Rhodes | 07/07/17 06:14pm

For this reason we’ve created a high security version (Shayype HSS) which doesn’t use the device’s keyboard, preventing a hacker knowing which characters have been typed in.

Since the public information regarding Shayype HSS is apparently limited to this one line,

its hard to evaluate it.

However, in regards to the viability of an access control application;

if it indeed forgoes the embedded device input for one using a mobile device, I would wonder what the point is anyway, since many other less cumbersome options exist already, if a smartphone is required.

Undisclosed #2

IPVMU Certified | 07/07/17 05:01am

While more difficult to explain than PINs, Shayype patterns can still be explained or sketched out and given to unauthorized users.

1 6 31 36 15 22 may not need much explaining :)

Undisclosed #3

07/07/17 05:34am

This makes it more challenging for a casual shoulder surfer to catch your pin, but a determined person might catch video of the session, and the users pattern can then be determined as long as the session pattern and the entered pin are captured.

Maybe it's unlikely to be able to get this on video thanks to a polorized filter limiting the angle of view and a well-obscured key pad in which case I guess it's pretty good at preventing stolen credentials.

But it's only marginally more difficult to share the pattern with a friend as a pin. Their demonstration video shows just how to do it. A quick scribble on a sticky note and the credentials are shared.

And any modern ACS should support resetting the password via an email link. Every WordPress blog out there does it, and it's free, so that doesn't really excite me. It should be standard.

If you want to prevent sharing of access control credentials, use dual factor authentication, anti pass back, and strong company policy against sharing.

The multifactor authentication itself will strongly limit the possibility of stolen access credentials, and scheduled access along with anti pass back will pretty much limit unauthorized access to the folks who are going to get in no matter what electronic measures are in place.

Afaik, all of this is already available on the market so as innovative as shayypes system appears to be, imo it is not adding significantly more security than what is already available in the market.

Just me?

Undisclosed #2

IPVMU Certified | In reply to Undisclosed #3 | 07/07/17 06:54am

but a determined person might catch video of the session, and the users pattern can then be determined as long as the session pattern and the entered pin are captured.

No, that would not be enough due to the duplicate digits in the grid. For instance, Brian's 411022 one-time password yields many thousand compatible patterns, not just his valid one of four corners and two center, so you would not be able to try them all before lock-up.

On the other hand, if you were able to secretly catch video of several logins by the same person, each with their own key grid, you could eventually determine the pattern, by deduction.

Maybe a quant could work the math correctly, but it would go something like each numeric code might be compatible with 50,000 patterns, ~6^6. Each attempt would yield another 50,000. Any patterns not in the patterns from the previous attempts would be discarded, since the correct pattern would have to be present in all keys.

Rinse and repeat till there is only one left. Maybe you could do it with just a handful of tries, or maybe they do something that would make it much harder, but one login definitely is not enough.

Unless you're really lucky :)

Undisclosed #3

In reply to Undisclosed #2 | 07/07/17 07:01am

Gah, you're right. My mistake! Still, that is pretty much the only attack this addresses and it seems to me it can be reasonably addressed with existing technology. But if you can add this to the list of available security features, why not I guess.

Paul Shah

07/08/17 12:26am

Can I Phish it by creating a fake image of the box to capture the pattern? It seems that all I need the pattern and it doesn't matter what the ransom numbers are.

The attack just changed from capturing the the random numbers to capturing a static pattern.

Brian Rhodes

IPVMU Certified | In reply to Paul Shah | 07/09/17 02:57am

Each number appears multiple times, so even a phising attempt would require multiple observations to understand a pattern.

I'll ask Shayype for specifics, but a single phising attempt will not likely divulge the pattern.

Paul Shah

In reply to Brian Rhodes | 07/09/17 07:31am

It's phishing the pattern, not the random numbers

Undisclosed #2

IPVMU Certified | In reply to Paul Shah | 07/09/17 06:33pm

But they don't draw the pattern, they only give you numbers back that correspond to many patterns, only one of which is the one that the user set up.

If you were to make every number on the your displayed grid unique your method would work, but since there are 36 squares normally, only using the digits 0-5, such a grid would clearly be a forgery.

Paul Shah

07/09/17 07:50pm

So if I were to present you a fake page that had the image and layout of the the matrix layout with arbitrary numbers, you were to click on the on the fake matrix and then i would be able to know your pattern and order.

I wouldn't be able to login at that moment, but I accomplished my goal.

Once I have your pattern and order, I can go to the real site, click on the real squares, grab the real random numbers to login to the real site

Then there is the MITM attack.

Undisclosed #2

IPVMU Certified | In reply to Paul Shah | 07/09/17 08:35pm

So if I were to present you a fake page that had the image and layout of the the matrix layout with arbitrary numbers, you were to click on the on the fake matrix....

There is no clicking on any matrix to log in, look at Brian's video above.

The only time you click on the matrix is when you setup the pattern for the first time, and then there are no numbers.

Conceivably, you could phish them to change their pattern, but that doesn't help you know their previous, still valid pattern.

However, it can be done, you just need several attempted logins with different grids to deduce, as I wrote above.

Paul Shah

In reply to Undisclosed #2 | 07/09/17 08:41pm

Yes, That makes sense

Brian Rhodes

IPVMU Certified | In reply to Undisclosed #2 | 07/10/17 01:52am

This is correct, Shayype users do not touch a pattern on the matrix. They visualize the pattern but never physically divulge it.

Read this IPVM report for free.

This article is part of IPVM's 6,445 reports, 867 tests and is only available to members. To get a one-time preview of our work, enter your work email to access the full article.