The Grugq is a world renowned information security researcher with 15 years of industry experience. Grugq started his career at a Fortune 100 company, before transitioning to @stake, where he was forced to resign for publishing a Phrack article on anti-forensics. Since then the Grugq has presented on anti-forensics at dozens of international security conferences, as well as talks on numerous other security topics. As an independent information security consultant the Grugq has performed engagements for a wide range of customers, from startups to enterprises and the public sector. He has worked as a professional penetration tester, a developer, and a full time security researcher. The Grugq’s research has always been heavily biased towards counterintelligence aspects of information security. His research has been referenced in books, papers, magazines, and newspapers. Currently an independent researcher, the grugq is actively engaged in exploring the intersection of traditional tradecraft and the hacker skillset, learning the techniques that covert organisations use to operate clandestinely and applying them to the Internet. You can follow him on Twitter @thegrugq.
John Little: You blog and have given conference presentations on Hacker OPSEC. You started doing this before the recent NSA revelations (and the general hysteria surrounding intelligence collection) but you were already warning hackers that states had superseded them as the internet’s apex predator. In just a couple of years we’ve moved from the seeming invincibility of LulzSec, to high profile busts, and now onto serious concerns being raised about the every aspect of the internet’s architecture, security models, and tools. Rock solid OPSEC is a refuge but maintaining it for long periods of time under significant pressure is very difficult. The deck is obviously stacked against anyone trying to evade state surveillance or prosecution so where do freedom fighters and those with less noble intentions go from here?
The Grugq: You raise a number of interesting points. I’ll ramble on about them in a moment, but before that I’d like to clarify for your readers a bit about where I am coming from. Firstly, I am not a “privacy advocate”, I am an information security researcher. My career in information security has been mostly focused around denial and deception at the technical level.
Recently, however, I became aware that this “fetishizing the technology” approach is simply not effective in the real world. So I turned to studying clandestine skills used in espionage and by illicit groups, such as narcotics cartels and terrorist groups. The tradecraft of these clandestine organizations is what I am trying to extract, inject with hacker growth hormone, and then teach to those who need real security: journalists; executives traveling to adversarial environments; silly kids making stupid life altering mistakes, etc.
The media has actually expressed a lot of interest in improving their security posture, and I am engaged in helping some journalists develop good OPSEC habits. Or at least, learn what those habits would be, so they have some idea of what to aspire to. There is a strange intransigence with some who reject improved security with the line: “but we’re not criminals! Why do we need this?” Well, the only answer I have is that OPSEC is prophylactic, you might not need it now, but when you do, you can’t activate it retroactively. As I phrased it in my “The Ten Hack Commandments” — be proactively paranoid, it doesn’t work retroactively.
So, that’s how I’ve arrived at hacker tradecraft, and where I’m trying to take it. On to the issues you’ve raised about good OPSEC and living a clandestine life.
The stress of the clandestine lifestyle is something that people tend to gloss over all too easily. This is an observation that comes up frequently in the literature about terrorist groups, espionage agents, and revolutionaries. There are a lot of compound issues which combine to make this sort of “good OPSEC” lifestyle very unhealthy for the human mind:
- Compartmentation of the ego
- Paranoia related stress
Isolation provides the strongest security, and all good security involves a significant investment in maintaining a low profile, “going underground”, “off the grid”, etc. This means that the clandestine operative has reduced visibility over the social and political landscape, and their telemetry will suffer. Degraded telemetry means they will be unable to self-correct and reorient to what is happening around them. If they are part of a cell, a group of operatives in communal isolation, they will tend to self reinforce their ideology. Effectively radicalizing and distancing themselves further from the mainstream norms of society. This additional isolation can create a feedback loop.
If the operative isn’t living a completely isolated clandestine lifestyle in their Unabomber cabin, they will have to isolate parts of their individual selves to compartment the different aspects of their lives. There will be their normal public life, the one face they show to the world, and also a sharded ego with their clandestine life. Maintaining strict compartmentation of the mind is stressful, the sharded individual will be a sum less than the total of the parts.
As if that wasn’t enough, there is the constant fear of discovery, that the clandestine cover will be stripped away by the adversary. This leaves the operative constantly fretting about the small details of each clandestine operational activity. Coupled with the compartmentalization of the self, the operative also has to stress about each non-operational activity, will this seemingly innocent action be the trigger that brings it all crashing down?
Seriously, maintaining a strong security posture for prolonged periods of time is an extremely stressful and difficult act. Operatives working for the intelligence agencies have a significantly easier time of it than those on the other side of the protection of the state: e.g. their agents; hackers; terrorists, and narcos. The “legal” operatives have peers that they can confide in and unwind with thanks to the protections of the nation state. The true clandestine agents must be guarded with their peers, the public and the adversary. Any peer might be an informant, either now or in the future. Opening up and being friendly with their peers is part of what lead to the unraveling of the lulzsec hacker group.
This leaves people who need to operate clandestinely and use the internet with a real problem. How can you be on the Internet and isolated? Well, compartmentation is the only answer, but it is expensive and fragile, even a single error or mistake can destroy the whole thing. This is why I’ve advocated that people who seek to operate clandestinely combine deception, that is, multiple covers, for their compartmented activities. It is possible to embed tripwires into the cover identities and be alerted when they’re blown.
My thinking these days is that an operative must minimize the time that they are engaged in a clandestine operation. Something like the theory of special operations, the period of vulnerability only grows the longer the operation goes on. Clandestine operational activity must be compartmented, it must be planned, it must be short in duration, and it must be rehearsed (or at least, composed of habitual actions). It is possible to do, and I believe that even non-experts can pull it off, but it must be limited in scope and duration. Prolonged exposure to underground living is caustic to the soul.
John Little: There is a significant amount of paranoia circulating in hacker and activist communities right now. How much of it is justified? More importantly, how should people go about conducting a realistic personal risk assessment before they start piling on layer after layer of OPSEC? How can they strike that balance between the tedium and isolation and security that is “good enough”?
The Grugq: There is certainly a great deal of paranoia, some of it justified, some of it unjustified, and some of it misdirected. I think it is important to remember that paranoia is unhealthy, it is paralyzing, it is divisive, and it is harmful to operational effectiveness. The goal to aim for is caution. Allowing the adversary to inflict paranoia on you, or your group, gives them an easy psychological operation “win”. So lets drop the paranoia and figure out what security precautions we must take in order to operate safely and effectively.
As you bring up, the core to effective security is performing a risk assessment, deciding what information is most important to protect, and then developing mitigation strategies to safe guard that information. There are books and manuals that go into this in great depth, so I won’t spend a lot of time on the details.
A risk assessment should focus on the most high impact items first. To determine this, you list your adversaries and group them by intent and capability. So the NSA would have a very high capability, but probably has a low intent of targeting you. Then you make a list of information about your secrets, what you are trying to protect, and group that based on the negative impact it would have if it were in the hands of an opponent. The most damaging information must be protected from the likely and the most capable adversaries.
Generally speaking, if you’re engaged in a clandestine activity that you want to protect, the core information to secure is:
- Your identity
- Your clandestine activity
- Your association with the activity
So lets take the example of the Dread Pirate Roberts, who’s been in the news recently after he got arrested. His adversaries were highly capable, including a wide range of law enforcement officials from across the globe. They were highly motivated, because DPR and his site were very high profile. So you have high capability, and high intent. Not looking good so far.
The information that was most important was his personal real world identity, followed by his location. Protecting that information would require:
- Robust compartmentation
- Reducing his exposure to the most capable adversaries (e.g. leave the USA)
- A strong disinformation campaign
- Limiting his time in “the dragonworld” (to use J. Bells’ term for the underground)
For most people engaged in a clandestine activity this list is probably what they will want to follow. The exact mitigation enacted for each component in the list is case dependent. As we discussed earlier, and as you’ve said, we need to find a good balance between an aggressive security posture and living a rewarding life.
Remember, the goal is to reduce the quantity and the quality of information available to the adversary.
John Little: So a point which both of us comment on with some regularity is the fact that security is rooted in behavior rather than technology. That’s always been true to some extent but never more than now. Tools are suspect, almost across the board. And a lot of assumptions about security have to be tossed aside. But one thing is certain, hackers adapt to the adversary. Terrorists do this well too. An attacker who can successfully parse all this and adapt is going to be a very significant threat. How can states counter the advanced threats? How can they counter hackers who know how to manage OPSEC and technical security to evade detection?
The Grugq: HUMINT. More of it.
The role of SIGINT in intelligence has basically been this weird bubble, starting around WWII when the love of SIGINT started until recently, when some of the SIGINT capabilities are starting to go dark. SIGINT is much more attractive than HUMINT. Signals don’t lie. They don’t forget. They don’t show up late to meetings, or provided intelligence information that is deliberately deceptive. SIGINT is the heroin of intelligence collection. The whole world got hooked on it when they discovered it, and it has had a very good run… it will probably continue to be useful for decades more, but really… the real utility of SIGINT will start to diminish now. It has to. The amount of encryption being deployed means that many mass collection capabilities will start to go dark. I, of course, am in total favour of this. I think that the privacy and protection of the entire Internet are more important than the ability of the US government to model the “chatter” between everyone using the Internet. The reduced security that the US government has tried (and succeeded) to force on the entire world is makes all of us less safe against any adversary.
SIGINT is really the sort of intelligence collection technique that needs to lose its prominence in the pantheon of intelligence gods. It is very easy for a serious adversary to defeat: basic tradecraft from the days of Allen Dulles will work (leave the phone behind, have the meeting while taking a walk). This tradecraft technique is described by Dulles, in 50 year old KGB manuals, and by Hizbollah operatives last year. The only way to catch people who are capable of any sort of OPSEC / tradecraft is via: a) Mistakes that they make (very easy for amateurs to make mistakes), or b) Via HUMINT. Spies catch spies, as the saying goes. It might be updated to, spies catch clandestine operatives.
Historically, the value of HUMINT has been very hit and miss, but those “hits” are extremely valuable. The major successes of the Cold War were almost all the result of human beings who became spies for the opposition: Ames, Hanssen, Walker, Howard, Tolkachev, etc. There are myriad cases with terrorist groups as well, informants is the best weapon against them. Relying on SIGINT is essentially relying on the adversary (terrorist groups) having poor tradecraft and terrible counterintelligence practices. This is simply not the case, at least not with sophisticated dangerous groups.
Double down on HUMINT and scale back SIGINT. SIGINT can be evaded, but HUMINT, essentially exploiting trust relationships, will always bite you in the ass.
John Little: Hackers are going to have to evolve in the same direction though aren’t they? Technology isn’t their salvation from an OPSEC perspective, in fact it is really the weakest link in their security model, so they will have to fully embrace good old-fashioned tradecraft and deception to avoid detection. Do you see an appreciation of that in the hacking community? It seems like a lot of big name hackers are still making fairly simple OPSEC mistakes.
The Grugq: Exactly, this is really the understanding that needs to sink in: technology alone will not save you. Hacker culture, almost by definition, is technology obsessed. We fetishize technology and gadgets, and this leads us to the deep-seated belief that if we just use the right tool, our problems will be solved. This mindset is fundamentally wrong. At best, I would call it misguided, but really I believe that most of the time it is actually counter productive.
Trust is the weakest link in the security chain, it is what will get you in the most trouble. This goes double for trusting in technology (even, as Bruce Schneier says “trust the math”). Tech is not the path to security. Security comes from the way that you live your life, not the tools. The tools are simply enablers. They’re utilities. OPSEC is a practice.
Expecting the tools to provide security for you is like buying a set of weights and then sitting around waiting for your fitness to improve. The fallacy that technology will provide the solution has to be seen for what it is, a false promise. There is nothing that will protect secrets better than not telling them to people!
Good OPSEC is founded on the same basic principles that have governed clandestine activities since the dawn of time. Hackers might be new, but good hackers require the same set of skills as the second oldest profession. Good OPSEC is timeless, and it stems from the application of the principles of clandestine operation, using caution and common sense.
The “73 Rules of Spycraft” by Allen Dulles was written before the Internet, before hacker culture (even phreaker culture) existed. I believe it is one of the most valuable guides available to understanding how to implement OPSEC. (As an interesting aside, harking back to one of my previous points, Dulles recommends taking vacations to get away from the stress of “work”.)
There are a lot of very public hackers who exhibit terrible security practices. Many of them are techno fetishists rather than espionage geeks, consequently they fail to understand how limited their knowledge is. Its the Dunning–Kruger effect in full tilt. They don’t do the research on their opposition and don’t know what sort of techniques will be used against them. By the time they figure it out, they are usually just an opportunity for the rest of us to practice Lessons Learned analysis. Of course the great tragedy is that many of the hacker community suffer from hubris that prevents them from actually learning from other’s failures.
A friend of mine paraphrase Brian Snow (formerly of the NSA) “our security comes not from our expertise, but from the sufferance of our opposition”. As soon as the adversary is aware of the existence of secrets worth discovering, and has the resources available to pursue them, hackers rapidly learn how good their OPSEC is.
John Little: I’ve always been amazed at the very public profiles of some hackers, especially where conferences are concerned. Granted, most are legitimate security researchers but there are also many in the community who occupy a grey area that is guaranteed to draw attention from intelligence or law enforcement agencies. Are hackers largely underestimating the skill with which intelligence agencies can penetrate, encircle, and absorb aspects of their community? Are we in for significant changes in the relationship between IC/LE and hackers, how hackers view themselves from a security standpoint, and how hackers engage each other?
The Grugq: Yes, very much so. There is a growing awareness of the altered threat landscape, and the need for an improved security posture. For decades the hacker community has been myopically focused on SIGINT threats, the sorts of technical attacks that have technical solutions. The HUMINT threat has been misunderstood, or ignored completely. That is changing as the hacker community is starting to learn and practice counterintelligence.
It is a difficult transition though, as some core counterintelligence principles run directly counter to the hacker ethos. There are a lot of factors at play, but one of the important ones is that hacker culture is very much a research culture. There is a great deal of knowledge exchange that goes on rather freely within various segments of the community. The problem, of course, is that the trading of information, which is so central to hacker culture, is the antithesis of a strong security posture. Many hackers realize this, so they only share with trusted friends, who then only share with their trusted friends, who then… and then suddenly everyone is on lists and someone is going to jail.
Security conferences are important events for hackers where they disseminate their research and findings, and socialize. This makes these events very target rich environments for intelligence agencies looking to build dossiers on hackers. They can see who is socializing with whom, attempt to recruit people, elicit information on capabilities, install malware on computers, collect intel from computers, and so on. That hackers would expose themselves to these activities seems very counterproductive for robust security. What gives?
The hacker community has a slightly different set of moral and ethical guidelines than mainstream society, which leads to problems with the authorities. Broadly speaking, few hackers view breaking into a system as unethical or morally wrong. Damaging the system, stealing information, or otherwise abusing the system is wrong. Simply accessing it is a challenge. The police, of course, view things differently: an illegal act is an illegal act.
For hackers the secret knowledge that they discover from active research is something to be proud of, and so we’re very excited to brag about our findings, activities or capabilities. This information is treated as something that will be kept within the community, bound by the FrieNDA. Of course, this is all based on trust, which is a very dangerous foundation for any security system. As Dulles’ says, the second greatest vice is vanity, the third is drink. Security conferences are not the places to avoid those vices!
So there is certainly this dynamic of wanting to brag about our discoveries from active research, but at the same time the tension of “what will happen if this leaks?”. These days we know what will happen, over zealous law enforcement and prosecution: weev, Aaron Schwartz, Stephen Watt, Dan Cuthbert, etc. The authorities view hackers as modern day witches, something to be feared and destroyed. It is unfortunate for the hacker community in many ways. Intelligent people who could contribute to mainstream society have their lives destroyed. So the repercussions of what are generally harmless activities can be devastating and life altering. Unfortunately, the protections that hackers turn to tend to be technological, but the problem is humans.
The hacker community is easy prey for law enforcement and the intelligence community. Very few hackers are savvy enough to spot a recruitment pitch, or to understand that what they think is amusing others view as criminal. I think this is starting to change. These days there is a lot less discussion about illegal hacking of systems (whether for monetary gain or not), and more about how to protect against the massive Internet surveillance that has been made public.
In this, I think, the hacker community and the general public are finding a lot of common cause against the LE/IC. There is a lot of good that will come out of this realization that the technology of privacy is actually important and should be ubiquitous, and easy to use. The default should be secure. Of course, as we know, this won’t help that much if someone is going around making basic OPSEC errors. So strong privacy protections for everyone will make the job of the LE/IC a bit harder, but it will also make everyone safer. I think that is a fair trade off.
Similarly, I think a lot of hackers would be quite happy to help the LE/IC community with technology support and ideas. The problem is that the relationship is a difficult one to establish. The IC is a black-hole, sucking in information and returning nothing. I don’t know how there can be meaningful engagement between the two communities, which I believe is a tremendous shame. There is a lot that can be learned from both sides, and I would love for the IC to contribute back. Law enforcement doesn’t interest me that much. Personally, my interest with LE begins and ends with studying their tools techniques and procedures for counterintelligence purposes. Something, that historically at least, few other hackers actually do. That is changing.
Hackers are learning to tighten up their security posture, they are learning about the tools techniques and procedures that get used against them, and they are learning how to protect themselves. Of course, the preponderance of criminal activity is committed in places where lax enforcement of computer crime laws allows blackhats to operate inside “protected territory”. In the long term, this is an extremely dangerous situation for those guys, of course, because without an adversarial environment they won’t learn how to operate securely. When the rules change, they will be caught out, completely unprepared.
The intelligence agencies and law enforcement departments have decades of organizational history and knowledge. The individual members can display wide ranges of skill and competence, but the resources and core knowledge of the organization dwarf what any individual hacker has available. Many of the skills that a hacker needs to learn, his clandestine tradecraft and OPSEC, are the sort of skills that organizations are excellent at developing and disseminating. These are not very good skill-sets for an individual to learn through trial and error, because those errors have significant negative consequences. An organization can afford to lose people as it learns how to deal with the adversary; but individual cannot afford to make a similar sacrifice — after all, who would benefit from your negative example?
The skills that hackers do have, the highly technical capabilities they can bring to the game, are not useful against an adversary who’s primary skill is manipulating other people. Knowing how to configure a firewall, use Tor, encrypt everything, etc. isn’t going to do much good if you also attend a conference without a highly tuned functioning spook-dar and a working knowledge of anti-elicitation techniques. The hackers are hopelessly outclassed at this game. Hell, the majority of them don’t even know that they’re playing!
Times are changing though, and hackers are starting to learn: OPSEC will get you through times of no crypto better than crypto will get you through times of no OPSEC.