I will be speaking at MITRE’s Social Media Strategy and Implementation Workshop in the Washington, DC area on September 28th. My topic is Attacking Social Networks. The goal of the talk is to show some of the darker aspects of social networking. These will be items and attack vectors that people may not be thinking about. Believe it or not some people are still oblivious to common social network attacks 
If you are in the DC area stop on by.
Social networks are shrouded in mystery. Just their very existence defies the laws of physics. If it were the late 40′s men in strange suits would be trying to dissect them at some top secret facility, but we have come so far since then 
Even though this sounds ridiculous, this is what many would have you believe about social networks. Why you ask? Because many of the people that talk about attacks and the dangers of social networks don’t even use them. They make all kinds of assumptions about soc nets that are completely false. The funny thing about assumptions when you are theorizing attacks is if your assumptions are faulty then your conclusions are faulty. Let’s cut the crap and focus on the real threats to social networks and their users.
Bruce Schneier just had a post in his blog about Social Networking Identity Theft Scams. In this blog post he refers to an article on ITworld titled Why you can’t trust ‘friends’ on Facebook as clever. This isn’t clever, this is dumb and extremely improbable. This is a perfect example of people talking about social networks that have no idea how they are used. The explanation of the scenario shows a clear lack of understanding of how social network users view and interact with their network.
I will not go in to all the specifics of what they were talking about, but it is based on the premise that you view your social network “friends” as you view your friends and family from the non-web world. Now, it’s possible and even likely that you may meet someone on a social network and actually become friends with them. This may even be part of the appeal for someone participating in a social network. The problem for an attacker is cultivating a true friendship takes time, effort, and resources. Attackers and scammers are all about effort vs reward. They are not going to take 6 months to a year of effort to try and scam someone out of 100 dollars.
Some other faulty logic they used is blurring the lines between the topic they were talking about and the Nigerian scam where they compromised peoples actual accounts. They then sent messages to their friends saying they were stuck in Nigeria and needed money. Still dumb, but this is a compromise of an already established social network presence. A far greater difference than a friend of a friend that you don’t know asking for money. You can see more information about that here. True they both ask for money, but the scenarios are far different.
If you want to talk about dangerous, during the talk Shawn Moyer and I did at Black Hat and Defcon last year and even our ShmooCon talk this year I mention a concept that involved attacking innocuous functions. On certain social networks this would allow you to semi-hijack a person’s social network identity. The concept deals with blocking communication and creating a denial of service condition for all visitors to someone’s social network profile. You could then create a new, duplicate identity with the user’s information and try to re-friend previous friends. In the message you tell them something went wrong with your account and you had to create a new one.
This is far more dangerous than the scenario that the article goes in to. It’s much easier than trying to compromise someone’s account, you are able to disrupt normal communications between friends, and you are able to potentially hijack already established trust. An attacker could then run a scam under this identity giving them a higher percentage of success.
I am the last one to say that social networks are safe, for example see here and here. I just can’t stand bad information and fear mongering. Yes, fear mongering. “The child molesters are going to get your kids on the social networks”. Yuck! In a comment on his own blog post Bruce said,
“I’ve seen some of my friends on Facebook put their address and phone number on their information page. Anyone they add can see it, and one such person I know has well over 1,000 friends. Not a good combination with videos of his two small children posted.”
Why is that not a good combination. You can’t possibly believe that 0.1% of the Facebook population are child predators?
Now it’s true that some people do put far too much information on their pages. This is due to the fact that it is not clear to them what is really sensitive.
Child predators are not trolling social networks (with any significance) trying to molest your kids. Child predators are opportunistic just like other types of attackers. They are not going to see an address on a social network and pay the house a visit. There are just too many variables for the predator to deal with. Parents, guns, neighbors, witnesses, geographic locations, and many other factors make this a prohibitive method for them to use.
Now as far as them using social networks to try and contact your kids there are many factors there as well. Social networks do monitor their network. Some networks are better at it than others, but there is the monitoring factor. Not to mention the person would have to spend quite a bit of time creating a relationship with your kids, which leaves them at risk for being found out by parents. I mean hopefully your kids don’t just go off to meet with strangers. If that is the case then you have much larger problems.
As parents you have control over the internet connection and your kids usage of the Internet. Know who they are talking to and what their activities online are. Remember your being curious not paranoid. You get paranoid over things you have no control over, these are your kids Know who they talk to and who their friends are. After all, a predator is going to try to get them alone and away from parents.
There is always the rare case that is the exception to the rule. Things happen and there are people who are just nuts and don’t think logically. People have been watching too much To Catch A Predator and think that the world is crawling with child molesters. Common sense should be your guide not a television show that is trying to get ratings. Besides in that show they had people posing as teens in an adult chat rooms, not social networks. Which just goes more to the point that I made about these individuals being opportunistic.
If you want more proof about the social network threats to kids being overblown you can read more about it from the New York Times here.
Having your address on your soc net page and then a message saying, “On vacation out of the country” seems like (and really is) a stupid thing to do. Let’s look at it closer from the viewpoint of a thief. There are many variables here as well that still wouldn’t make this feasible. What about alarms, house sitters, family, neighbors, etc. This is on top of the information gathering activities that a thief would have to do prior to targeting someone anyway.
Now what is much more likely that attacker would target someone and augment their activities with information they find on social networks. These sort of targeted, personal information gathering activities can be pretty dangerous, but still not very realistic from a thief’s perspective. Thieves are opportunistic as well. What would change the scale is if you had known assets that someone REALLY wanted. This would warrant the time put in to the information gathering activities. Even in these scenarios the information from social networks only helps, the person would most likely be targeted anyway. There are rare exceptions, but just trying to put this in to perspective.
There are aspects that make social network ripe targets for attack. They are a large collecting point for users. They are made up of mostly user generated content, many allow extensions and 3rd party applications. Any large collecting point of users is going to be looked at by an attacker. These are just the facts, but when discussing dangers and threats we need to look at them in terms of real risk. When we raise the danger flag for things that aren’t necessarily a risk we may draw attention away from things that really are a danger.
I particularly enjoy the individuals who say that they would never join a social network or communicate with people who do. As if people that use social networks somehow don’t know something that they do. I turn that around, why not use social networks? Are you socially inept and not able to communicate with your fellow man? Do you even know what social networks are used for? Of course, using social networks is a personal preference. It doesn’t have any bearing on the user’s awareness or intelligence level. However there are millions of the ugliest MySpace pages in history just waiting for you to view them
Now there are some social impacts when professionals use social networks that I may cover in another post, because these have impacts as well.
The low level of probability of these attacks is no excuse to be careless with your information. I just wanted to put some things in perspective and curb potential fear mongering. When you participate in a social network you are responsible for the information you post about yourself.
I think ultimately if you read articles or hear people theorizing about attacks on social networks and they don’t have a social network presence, be skeptical. This is especially true when they are discussing social attacks. While it’s true that social networks are just web applications sometimes the vulnerabilities come from how users interact with them. This often requires participation for understanding.
Lastly, I want to make it clear once again, I am not vouching for the safety of social networks by any means. There are many dangers on social networks. I just want to make sure that we focus on the true dangers of social networks so we can raise awareness for those issues.
The past few days there has been a bit of a stink about some bogus LinkedIn profiles. There have been plenty of news sources reporting that LinkedIn profiles are serving malware or making it seem like profiles are infected somehow. A few examples of that can be found here and here and here. At least The Register called these people falling for this fools. What the titles of these reports imply are dead wrong. LinkedIn profiles are not actively attacking users.
The issue is very simple, it is a hyperlink to another site that infects idiots with Malware. A hyperlink to another site, not getting attacked from viewing a profile. When you allow users to link to off-site content, you lose control of the request, however, this isn’t like allowing users to pull content in from other sites to display on their profiles. This typically has very little impact. This is no different than any other site, message board, or social network.
Give me a break, like Beyoncé Knowles has a LinkedIn page and is going to have a hyperlink on there to a place to view her nude pictures. That’s the issue these sites are referring to, dumb isn’t it? How does that get turned in to words like serving, harboring, or redirecting? These words imply some sort of active action on LinkedIn’s part, which doesn’t describe the situation here AT ALL. If you ran a message board and someone had a hyperlink to Goatse, does that mean you are serving, harboring, or redirecting to Goatse? Of course it doesn’t. This would just be an indication of your user base. I wonder how many people were brave enough to click the Goatse link above It’s not Goatse, promise.
Is there really no end of the Internet news stories this week to scare people with so people decided we should be scared of LinkedIn? This is basically spreading FUD. I personally don’t see why LinkedIn should take any heat from this. The feature of LinkedIn that allows you to link to your Company, personal site, or some other site should remain a part of LinkedIn’s features. I really hope they don’t go with something like MySpace did with the msplinks stuff. This would basically put a big obnoxious splash page up that states you are about to visit content off of the site. Yeah, well no crap I just clicked on the link so of course I want to visit the page. I personally don’t think that is a very effective control for these types of attacks anyway. The only time that control is effective is if it isn’t clear to the user that they are visiting content off the particular site they are on. I have seen in the past MySpace profiles that were compromised and the whole profile links to a bogus MySpace login page. In that case the user seeing the warning would be alerted that something is wrong, however, you are still going to have a large amount of people just cough up their credentials anyway. Sometimes all the controls in the world just can’t fix stupid. The same people that would fall for this are the same ones that click on spam emails claiming the same thing. It’s a mentality not a technical security issue.
Let me state this, if you are not a complete idiot then this issue will not affect you in the least bit. These profiles are not performing any active attacks on users of LinkedIn. There are much more scary things out there than this, trust me. Don’t fear using LinkedIn because of issues like this. LinkedIn really has a very limited feature set which lowers their attack surface. They have much less functionality that other social network such as MySpace, Facebook, Hi5, etc. Would you really care to see Beyoncé Knowles’ LinkedIn profile anyway? I bet she is boring and fake. Her LinkedIn profile would state, “I have never had to work for anything in my life and everything has been handed to me because dummies think I have talent. I love screwing over my friends and taking money out of their pockets”. She should apologize to the world for creating that DirecTV Upgrade song. Yuck! Wait a minute, she doesn’t write her own music… Anywhoo….
I can’t believe I had to write this blog post, but the sheer number of people talking about this and linking to these stories was too much. Just practice smart Internet browsing habits mixed with common sense and you will be fine. As always, I recommend using the Firefox web browser with the extensions NoScript and Adblock Plus. Have a good week, the end of the Internet is next week
Hello Everyone. I just wanted everyone to know that I will be speaking at ShmooCon 2009 with Shawn Moyer. Our Topic is Fail 2.0: Further Musings of Attacking Social Networks. This will be an update to our Black Hat / Defcon 16 presentation. Update as in, we will have some new material and updates to what we have previously talked about. We won’t be consistently beating an already dead horse. We felt that the topic still contains quite a bit of relevance. As companies continue to shift their focus toward social networks and social networking platforms in general, they are encountering the same security problems. Even though social networks are web applications, they do offer some unique challenges over other common web applications. We will be explorisploiting these differences
Business social network LinkedIn announced their LinkedIn Applications today. The applications directory can be viewed here There are only several applications to chose from at the moment. I am sure that number will grow soon. LinkedIn uses Google’s OpenSocial just like other social networks such as MySpace, Orkut, hi5, etc. I only spent like 5 minutes looking at a couple of things. So, the following are only my quick thoughts and impressions.
The applications are delivered though the domain lmodules.com. This makes them easy to identify and block if that’s what you would like to do.
At first glance it appears that the vetting process for LinkedIn is higher than some of the other social networks. They appear to only want known businesses to create applications for their network at this time. This would help root out some possible malicious users. A vetting process is a good first step in thwarting that type of malicious behavior. I didn’t look at the difficulty in attaining a developer account, but I am assuming it is much more difficult than other social networks like MySpace, Facebok, etc. Now, whether this vetting process will stay this stringent will remain to be seen. These procedures may be relaxed in the future due to demand.
Just because the name has changed doesn’t mean the threats have changed. As a matter of fact there may actually be more on the table. Business networks such as LinkedIn are more likely to contain real information about people vs other non-professional social networks. Not that people don’t share enough about their real self on other social networks. This means the same threats exist for the capture of information as on other social networks.
There are still technical threats from social network applications on LinkedIn as well. These are the very same issues as other social networks that we have discussed in the past and demonstrated. Malware distribution, social engineering, attacking clients, information harvesting, click fraud are just some of these threats from social network applications. Moral of the story is be careful. Don’t install apps you don’t need, even though you may do so on your iPhone
So all in all the threats are the same with LinkedIn as any other social networks that employ applications. However, with a more stringent vetting process this should reduce the possibilities for malicious by making accounts harder to get.
For those of you that care, there is a caricature of me on the cover of the November issue of CPU Magazine. In the back of the magazine there is some Q&A with me mostly about social networks. It’s probably stuff you have heard Shawn and I say before, but cool nonetheless. So if you are in your favorite book store check out the magazine and see what you think.
Password Reset: Your passport to a fuxored account.
Password Reset Methods Vulnerable? Really? Get out of here, you mean that many password reset methods are vulnerable to attack? You have to be kidding. The fact that people think vulnerable password reset is newsworthy have got to be crazy. This is something that many of us have been talking about for years. Now Sarah Palin’s email gets attacked and it is big deal. It amazes me why we always wait to get screwed by something before we fix it.
Why does everything in the security world have to be a response to something. Ok, not the security world but the business security world. They are definitely two different entities. I am truly tired of reactive security. Just think if other professions followed this reactive model, like a cop asking for a bullet proof vest after they have already been shot. Nobody can say they didn’t see this coming either. People make more of their life known through social networks, photo sharing, and blogs than ever before. The simple password reset questions just don’t hold up.
There is a lot of unnecessary fear about data from social networks being used to steal someone’s identity. Although this is mostly FUD, social networks can be a great source for password recovery data. A while back we recovered a password (with his permission of course) from my friend Brian’s Sprint account using data from his MySpace page. This is when we were first starting our research for the social network hacking project.
Let’s take a step back from social networks for a sec, would your friends, co-workers, significant other, etc. be able to recover your password with the information they know about you? If the answer to that question is yes, then you need to change something. Passwords should be something that you know, not you and a couple of other people.
The information that people put on their social network pages range from minimal to wildly over the top. Some people even go above and beyond by posting survey questions that tell a lot about their personalities. Although they want to show off the depth of their personality, all it really does is show off the shallowness of their brain.
Social networks by their default nature basically allow you to “friend” the world. The information on people’s social network page typically contains information that was previously only known to traditional friends and acquaintances. This can be a huge problem for the password reset mechanism, not to mention a person’s privacy. If it’s deep and kinda scary from a privacy standpoint then it is probably on a social network. Remember when I mentioned if your friends knew enough about you to reset your password then you are in trouble, well you just friended the world with the information from your social network profile. Beyond standard profile information there are a users actions taken on a social network site and possibly social network applications that are being used as well. All of this information can be leveraged when attacking a password reset mechanisms.
You can use an email address to look up people’s accounts on social networking sites. On the flip side, someone social network profile might directly tell you a person’s email address or you can use the search features of the social network to query owner’s of certain email addresses. There are no secrets in social networking
With password resets an email account is really the jackpot. Many password reset mechanisms, including the ones from social networks, rely on sending either the password or a temporary password to the email address of the account owner. Someone who gets their email account compromised might just find that they have every other account tied to that email account compromised as well. I mean, it wouldn’t be a far stretch to figure that out once someone had access to the email account. Just think of all the crap that sites like Amazon, eBay, MySpace, Facebook, etc. send to your email account.
Typical password recovery questions really vary in complexity from site to site. What is the problem with password recovery questions in general? Well, they are not typically made up of data that is private. Unlike a password which is supposed to be something that only you know, recovery questions may be known to many people around you.
Here are some questions from Yahoo:
Some of these questions look like questions that social networks ask when you are filling out a profile, don’t they? If not questions they ask, certainly data that people put on their social network profiles or divulge through other means on a social network.
Take a glance at someone’s profile or maybe your profile on a social network. From just this page without further probing there may be an enormous amount of information. Depending on the mechanism that is being attacked, it may be all that is needed. Here is an example of some of the things that may be found just on the profile page:
The list really goes on and on. Remember that many people are on multiple social networks. Checking out other social networks may fill in the blanks. It is easy to see why this information could be a problem and I don’t think it needs any further explanation.
Some data is not so obvious and might not be directly spelled out. This may be information that has to be aggregated or inferred from the profile data, friends list, blog, group, network, etc.
The data is really limitless, but after all isn’t that what a nice web 2.0 application is supposed to provide? On the surface some of this data may seem silly for password resets but it is really not. This not so obvious information can be really helpful when when non-standard questions are used in the password reset process. This typically happens when people are left to their own devices when creating security questions. They typically create questions that are common and familiar to them. Stupid things like pet’s names, favorite teams, favorite TV shows, etc.
Just think for a moment about tagging. People may tag photos themselves with useful information. Also, friends may tag people in photos helping better define a person’s relationships with people and activities they are involved in. The URL of the social network may lead you to potential usernames / IM information such as www.myspace.com/(username). Maybe the data is completely visual like photo data. A lot of information can be obtained by looking at pictures. Favorite places, sports teams, cars, and countless other possibilities. You name it, people like pictures with their favorite things.
The actions people take on social networks helps better define relationships, networks, group affiliations, and activities. The person may place comments on other people’s photos, profiles, walls, blogs, etc. You may see comments like “That is why you are my BFF”. You may also see that someone is a member of a political party or religious group. People may discuss on boards or blogs about certain things happening in their life. Sharing is caring right?
So what you get in the end is a clear picture of who these people are. You get their likes, dislikes, friends, and affiliations are all in a nice clean package. You may have never even met this person but you have all of the information a traditional friend may have, possibly more.
If you almost have the nail in the coffin then you can turn to other sites to complete the task. You could look for name / username collisions on other sites to gain more data. You could take their high school and age information and find out who they went to school with. The possibilities are endless.
When people are given the option to choose their own security it has historically been bad. There is nothing that seems to suggest that allowing user’s to choose their security will get any better, so some of this may be wasted breath.
When looking at sites like Google, it seems they have slightly better security questions. Questions such as your library card number, frequent flyer number, etc. I think sites like these with better security questions probably have a high amount of people that end up just choosing their own questions when this option is available. People don’t seem to understand that this isn’t a function that you are going to use everyday. It is ok and preferable to use data that you may not be able to recall without looking up.
The problem of personal data leakage isn’t going to stop until people realize the potential impacts of their data being strung out for the whole world to see. I personally don’t think this will change, in fact, I think with time it will get a lot worse. We live in this voyeuristic, virtual world where people create digital representations of how they see themselves. I think that has an appeal to many people, especially those who don’t particularly find their lives that exciting.
Don’t play by the rules when dealing with a sites password reset questions. Put blatantly wrong, hard to guess, or nonsensical information in to the answer blocks. This will make any information gathered on you useless when attempting to recover your password.
It seems that many sites want you to log in. You shouldn’t use the same password on every site. Use a trusted password safe such as KeePass to store your login credentials. KeePass is open source and multi-platform. Using a mechanism like this allows you to be in control of your password recovery along with allowing you to use different passwords for different sites. It would also be a good idea to back up the database of whatever password safe you choose to use as well. Just a thought
The biggest mistake someone can make is thinking that there is nobody out there that gives enough of a crap about them to attack their accounts. People do weird things. Anybody is capable of just about anything. This isn’t being paranoid, it’s being safe. Think of it as locking the door on your house when you leave, only instead of your valuables you are protecting your data.
I find this funny, here is an article where it talks about Facebook users leveraging developer accounts they signed up for, so they can go back to the old Facebook. When you have the developer application installed it puts a link at the top of your profile page to switch back to the old Facebook. This makes sense since developers may have to maintain functionality on the old Facebook as well as the new. The funny thing is, I have had a developer account almost as long as I have had a Facebook account. I just assumed that option was on everyone’s page. It is a nice little hack, although Facebook is going to turn that option off soon.
It also seems that the users flooded the developer message boards voicing their distaste for the new Facebook. Ah… well… I hate to break it to them, but shhhh… those message boards are for the apps developers. There are Facebook staff that hang out and such, but all you are doing is irritating the person who wrote that stupid app you put on your page and don’t use. Stop! You are distracting them from doing input validation
I don’t really understand this UI rebellion. Who cares. I mean, really you can do the same stupid things you could do previously. If you don’t like it use MySpace. The more you look at Facebook the more it looks less and less like a social network anyway. The other day on TechCrunch there was an article called Facebook Isn’t A Social Network. And Stop Trying to Make New Friends There. I agree with this viewpoint. MySpace and other social networks are much more conducive to meeting new people and finding individuals with similar interests. It all depends on what you use a social network for.
I have an idea for all of the people who don’t like the new Facebook, the best way to rebel, is to quit using Facebook. That will get their attention. I know it won’t happen, but it isn’t like there aren’t alternatives. I mean, what does Facebook give you that other social networks won’t? The answer is nothing. Most people have accounts on multiple soc nets anyway. There are some 800,000 users in the I hate the new Facebook group. If they all quit using their accounts, that would have an impact. Do it or quit complaining. The choice is yours. The reason nothing changes is because they know you won’t leave.
Recently Facebook announced their Application Verification Program in an attempt to give user’s assurances that particular applications are secure. I think the intent is good but the implementation may actually cause more harm than good. Giving users an assurance that a malicious applications are secure can cause a lot of damage. People with assurances are a lot more loose with their actions where they may normally not be with no expectation of security.
Given the way many of Facebook’s applications are written it doesn’t lend itself to a proper review. The Facebook team is going to have to do reviews of submitted code that does not run on Facebook servers. This would only be a snapshot of the code at that given time. After the verification procedures are done, the developer can make whatever changes they want. They could change the verified app to a malicious app at will. I am getting so tired of security measures that don’t address the real problems. They are a waste of time. The only thing this verification program may do is stop the idiot who just learned PHP from creating the HackMe Back of social network applications. It doesn’t address the major problem that attackers are gaining access to the API and attacking social network users.
The best way to protect against malicious applications is to control the access to the API in the first place. Don’t just let anyone access the API and only need 5 friends to publish the app. Proper vetting procedures would go a long way in curbing the amount of malicious applications that get published on Facebook and other social networks. Why don’t the major social networks have vetting procedures for API access? It completely blows my mind, but that’s social network culture for ya.
Social networks are riding a thin line with security as it is. Introducing security measures that aren’t effective only cause more confusion on the part of their users. Social networks should strive to create a balance between functionality and security for everyone’s sake. Will that happen? Only time will tell. One thing is for sure though, attacks on social networks are only going to go up. The more surface you give an attacker the more options and success they are going to have.
Quite a few people have emailed me asking me what I thought about the Facebot application that was recently released. The paper is located here. Basically a group of people created an application that they published on Facebook that did click fraud. They hijacked simple requests through an application called Photo of the Day using HTML IMG tags, you know, the same thing we did on MySpace without even having to create an application, however, we had OpenSocial applications that did the same thing, and a little worse
They said they did it to prove you could turn a social network in to a botnet, you know, the same thing that we already talked about and demonstrated at both Black Hat and Defcon this year. As a matter of fact a copy of our presentation can be obtained here: Satan_Blackhat_Defcon
The title of their paper is “Antisocial Networks: Turning a Social Network into a Botnet”. The title of our HOPE presentation that we had to back out of was “Antisocial Networking: Vulnerabilities in Social Nets”. You can see this here from back in June. I am not quite sure what to think about all this, I guess it could all be coincidence. Like I said, I don’t know.
Now on Facebook the way you would have to go about turning their users in to a botnet is by creating an application. Facebook doesn’t allow linking to offsite content the way MySpace does. So if you want to use img tags, meta tags, and iframe tags you would have to use them in an application that you created.
So, my impression is Yup. Everything we talked about at Black Hat and Defcon. It’s old news, not sure why anyone is making a big deal or even writing about it.
