Hacker News Reader

Comment by hiimnate | Hacker News Reader

>security holes

It's an arrow that points in one of two directions. I can not think of a single way this could be made insecure. Even if someone was able to influence the direction the arrow was pointing, everyone is still going through the same ineffective screening process.

View on HN

One way it could be made insecure would be if it didn't use a good source of randomness, so the arrow direction could be predicted. A worst case example would be if the app just cycled through a fixed table of "random" choices like "L R L L R L R R", then the attacker could just watch the pattern and position himself in the line appropriately to make sure that his buddy Joe is the one to screen him. repositioning is easy, just dig through your bag and say to the guy behind you "I can't find my passport, you can go in front of me".

That problem would be solved by making the app very unstable and require constant restarts ;)

And all of this trouble to achieve.....what exactly?

To achieve a "screening" carried out by a co-conspirator, through which weapons, etc. could be passed.

1. The TSA is ineffective at preventing weapons from getting past their security checkpoints.

2. Everything required to make a bomb can be bought past the TSA checkpoint.

3. It would stand to reason that if a terrorist organization can get a conspirator to infiltrate the TSA, they could get, I don't know, two?

4. There's more than one person involved in the screening process.

The fewer people you need to get into the right places at the right time, the more likely your plot is to succeed.

It's not like TSA agents haven't already been caught letting people smuggle contraband past them:

http://www.foxnews.com/us/2015/12/19/tsa-agent-accused-smugg...

A worst case example would be if the app just cycled through a fixed table of "random" choices

Not even a newbie programmer is going to do this. Every language out there has a Random() function.

Seriously, all these worst case scenarios and we're still well within the land of what even a middling iOS developer could crank out in under an hour.

> Not even a newbie programmer is going to do this.

I wouldn't bet on that. I've seen all sort of horrendous monstrosities that students and new grads have come up with.

You'd be surprised what programmers will do. Sometimes due to incompetence, sometimes it's an "optimization", sometimes it's a simple mistake, yet there is some very bad code out there and it's not discovered until it causes a problem.

For example, even a newbie wouldn't trust a remote client to tell you the size of a string they are sending you so you can echo it back to them, yet we still had Heartbleed and it sat there for a long time, undiscovered.

Which usually isn't cryptographically secure. So now you need to audit the random number generator and make sure that it has no known holes, including obscure edge cases. And then you need to confirm that any implementation is correct and audited.

That quickly takes you down a rabbit hole, which may be why the contract was so expensive.

This is iOS we're talking about. The CSPRNG provided as part of Swift or Obj-C's standard library has likely already been audited for use on other govt. projects.

http://stackoverflow.com/questions/9234686/generating-random...

Great point. Thanks for the link!

Oh you need a cryptographically secure RNG? Assuming the app is written for android just replace Random with SecureRandom. Here's your $100,000 invoice.

What if the device with the arrow leaks EM radiation which can be used to reconstruct the random seed, thereby allowing an attacker to predict which way the arrow will point after n presses? Then, you can just time your place in the queue to get the result you want, as long as you can discreetely measure the device from a distance.

I can't tell if you are joking or being serious. I hope this is a joke.

I am "joking" in the sense that I don't expect even this 300K design to have taken that sort of vulnerability into account. I actually agree that most of it probably went into meaningless overhead. I am not joking in that I think that if you are designing a formally specified security-sensitive cyber-physical system, there are reasons why even very simple things are hard to get right.

I don't actually much care about the security of TSA hardware (my view is that it is all expensive security theater in any case). However, if you told me you spent 300K designing a very simple control interface for say, a critical component of the electrical grid, and the argument you gave me for the cost is proper security engineering, I would buy that. I certainly would prefer it to a $30 USD solution developed as a HackerRank project, even if the nominal functionality is the same.

They could hack into the iPad to make it display this: http://www.thecleverest.com/countdown.swf

High chance somebody gets shot.

That's a lot of hacking to display Flash on iPad :)

Reply on news.ycombinator.com