Quirks in Tech

F-Secure has a good, quick, overview of the recent attacks against Facebook, Twitter, and (presumably) other mobile developers. Significantly, we’re seeing an uptick in attacks against developers rather than just against platform manufacturers. The significance? Even though the phone OS may be ‘secure’, the applications you’re loading onto those devices may have been compromised at inception. 

Smartphones: the source of anxiety and worry for IT managers that keeps on going.

The Australian Department of Defence, Intelligence and Security division, has produced a particularly good walkthrough for hardening the iOS environment (.pdf). I’d recommend it to the curious and for system administrators who are interested in evaluating/contrasting their own iOS deployments. 

First, the good news: 1Password has released a new version of their product on iOS. The company outlines a whole pile of reasons for supposedly delaying security upgrades - some of which include the updates will slow the speed at which users can access their encrypted data - but fail to identify what I suspect is a key motive behind the upgrade. If you recall, I wrote a while ago about key failures in mobile password managers. 1Password was amongst those who had flawed security implementations.

To be clear: security, especially good security, is damn hard to engineer. 1Password didn’t have the gaping flaw that others did - i.e. storing passwords in plaintext!! - but it was flawed. In the security community this (ideally) is resolved when someone critiques your secured infrastructure. In today’s world you should also credit the security researcher(s) who identified the flaw.

Unfortunately, this isn’t what 1Password has done. As far as I can tell, there is no formal recognition from the company that they have had flaws in their mobile security model pointed out by a third-party. This is a shame, given that a key factor that builds genuine trust in security is transparency. It seems like 1Password is willing to address problems - they’re not dwelling in a security by obscurity paradigm, to be sure! - but not credit others with finding those problems in the first place.

Update: My very, very bad. I missed an earlier piece from 1Password, where they note the research. That is available here. It would have been ideal to see a reference to this in their update but, admittedly, credit had previously been given. 

Symantec is warning that the next generation of smartphone viruses has come:

Researchers from security vendor Symantec Corp. have identified a new premium-rate SMS Android Trojan horse that modifies its code every time it gets downloaded in order to bypass antivirus detection.

This technique is known as server-side polymorphism and has already existed in the world of desktop malware for many years, but mobile malware creators have only now begun to adopt it.

A special mechanism that runs on the distribution server modifies certain parts of the Trojan in order to ensure that every malicious app that gets downloaded is unique. This is different from local polymorphism where the malware modifies its own code every time it gets executed.

This is a clever means to avoid the rudimentary analysis systems that the major vendors use to ID malware. It’s also (another) indication of how important antivirus is going to become for the mobile marketplaces. I suspect that, by the end of the year, a lot of users (on iOS, Android, and the rest) are going to wish that the post-Steve Jobs smartphones on the market today met Jobs’ initial thoughts regarding smartphones when Apple released the iPhone. Specifically, he held that:

He didn’t want outsiders to create applications for the iPhone that could mess it up, infect it with viruses, or pollute its integrity

While our pocket computers are better now that apps are available, I can’t help but think that Jobs’ earliest worries are now looming at today’s potential nightmares.

I’m sorry, but what Path did is (in some jurisdictions, such as my own) arguably a criminal offence. Want to know what they’ve been up to?

 

When developer Arun Thampi started looking for a way to port photo and journaling software Path to Mac OS X, he noticed some curious data being sent from the Path iPhone app to the company’s servers. Looking closer, he realized that the app was actually collecting his entire address book — including full names, email addresses, and phone numbers — and uploading it to the central Path service. What’s more, the app hadn’t notified him that it would be collecting the information.

Path CEO Dave Morin responded quickly with an apology, saying that “we upload the address book to our servers in order to help the user find and connect to their friends and family on Path quickly and efficiently as well as to notify them when friends and family join Path. Nothing more.” He also said that the lack of opt-in was an iOS-specific problem that would be fixed by the end of the week. [emphasis added]

No: this isn’t an ‘iOS-specific problem’ it’s an ‘iOS lacks an appropriate security model and so we chose to abuse it problem’. I cannot, for the life of me, believe that Apple is willing to let developers access the contact book - with all of its attendant private data - without ever notifying the end user. Path should be tarred, feathered, and legally punished. This wasn’t an ‘accident’ but a deliberate decision, and there should be severe consequences for it.

Also: while the Verge author writes:

Thampi doesn’t think Path is doing anything untoward with the data, and many users don’t have a problem with Path keeping some record of address book contacts.

I think that this misses a broader point. You should not be able to disclose mass amounts of other people’s personal information without their consent. When I provide key contact information it is for an individual’s usage, not for them to share my information with a series of corporate actors to do whatever those actors want with it. The notion that a corporation would be so bold as to steal this personal information to use for their own purposes is absolutely, inexcusably, wrong.