If you're using off-the-shelf "hidden" encrypted volume schemes, you're not going to be evading state-level adversaries; if you can find these projects and conveniently use them, state vendors can and will write scanners that find them. They're paid to do it; new detections are how they get to charge for maintenance and new versions.
Then you're down to two issues:
(1) Concealing an encrypted volume jacks suspicion way up; whatever pickle you were going to be in if you just kept an encrypted DMG on your desktop, you're trebly in now.
(2) Your actual security comes down to the strength of the encrypted volume, and this is 20-year-old encryption. He can't use a memory-hard KDF because of his BOM, which, like, fair enough, but that doesn't change the fact that this KDF is probably ~50x faster than standard bcrypt hardness and on realistic human passwords is probably crackable in minutes-to-hours on a dedicated rig.
I winced particularly at the observation that the hardness was set where it is because of how long it takes to open the encrypted volume on this hardware, because whatever scheme is being used to run the KDF now, an attacker won't bother; they'll just copy the bits and attack them on serious hardware.
Point (1) isn't meaningful if your adversaries aren't states. It might make sense to have a hidden encrypted volume for the same reason you'd want an encrypted safe. But point (2) applies to everybody.
>If you're using off-the-shelf "hidden" encrypted volume schemes, you're not going to be evading state-level adversaries;
This would evade 99% of them.
Most "state-level" adversaries are just high school dropouts working border jobs told to "investigate suspicious things".
I once had a police officer literally raid my room for drugs during a massive party we were throwing in university, for example.
They took everything. They found everything. Everything except for the $1,000 piece of drug paraphernalia I had sitting right out in open view in the middle of the apartment... the Volcano vaporizer.
It didn't look like anything they had seen before so they didn't think anything of it.
Same would be true with this USB drive in 99% of border crossing privacy invasions... plug in USB, ask for password, read it/copy it/whatever... they're not taking it apart and performing open heart surgery on a $5 USB drive at the border.
They might throw it out... or keep it or something... sure... but in general... this is a hell of a lot better than a regular USB drive imho.
This is self-soothing, not a real security plan. It doesn't matter how competent a security agency is, because there's a whole ecosystem of vendors selling into that space; knowledge of how to attack encrypted disks like this is outsourced, and, importantly, those firms have incentives to mop up even random stuff like this, because vendors will be selected in part based on lists of how many different circumvention and privacy tools they can detect.
This isn't stand-up-to-state-actors security. This is keep-state-actors-from-noticing security. They know you're guilty, they'll get it. The XKCD version of password cracking works no matter how secure your hardware.
But while states realistically can crack most anything it doesn't mean they can crack *every*thing. There are simply too many flash drives crossing the border every day.
But there is a big flaw I see: 8gb. You're carting around an 8gb flash drive these days??
We're talking about exactly the same threat. What I'm pointing out is that state adversaries won't have to "notice" this; all they'll have to do is plug the device into a standard commercial forensics scanner product --- and we're stipulating that they're plugging the device in already (else what does it matter what bits are on it).
They pay other people to notice this stuff for them!
I think part of the entire design is that it's quite hard to detect.
1: Reads/Writes are just routed to a COTS SD-card.
2: Unless the (correct?) password is detected in the write-data that starts the disconnect procedure.
The only way to detect it from what I can see is to profile writes then append a "password:" string multiple times to measure the write-delay, only works if the CPU cost is large enough to overtake the SD-card measurably and some constant-time optimizations should make it more or less undetectable.
It's a clever design and I think it should be possible to optimize to become more or less unrecognizable.
> The only way to detect it from what I can see is to profile writes then append a "password:" string multiple times to measure the write-delay,
- Have the first check be a simple 8bit hash that filters out most passwords in microseconds, or use a customizable prefix instead of “password:”.
- have your password checking thread run in background at idle priority
- when you get an async password match, force usb disconnect and reconnect and the system will rescan the bus and mount your real drive.
Short of adversary dumping drive firmware (or them figuring out your hn account t, having a LLM scan the messages and finding this conversation) it’s not really easily detectable..
If I was building a black box to detect hidden data on a USB stick, I'd include a feature whereby it measures power consumption and flags USB drives that don't consume the expected power for that type of drive.
What an unreasonably complex device to imply you’re going to make 50,000 copies of and ship out to every border agent in the nation after telling them every SINGLE flash drive crossing the border MUST be scanned. That’s so much wasted time, energy, and money. Are you building a similar device for SD cards? HDDs? SSDs? External drives? You’re going to build a database of exact charge draws for every device in existence which happens to have some form of storage on it?
Fucking up every single import and traveler on the off-chance one in seven billion scans detects one of these devices and, further, carries some data the US cares enough about to chase down to this degree?
No. Nope. Nuh uh. Insanely unreasonable. You will get ONE at a station, which you only bother to use if you can convince your $12/hr employee to even give a shit. Literally an insane proposition.
> They pay other people to notice this stuff for them!
Money, even at state level, is not some infinite resource (and neither is time).
States put effort into what they consider worth investigating - no state strip searches every incoming traveller, and goes through every item in their possession - it's possible for them to do, and if they did they'd find a hang of a lot more contraband, but it's costly, time consuming, and politically risky.
Try to explain how the hidden content would be found? It effectively doesn’t exist until the password is in place. It isn’t hidden, it isn’t connected at all.
He deleted everything ultimately. He printed it out in ultra tiny font in order to comply, but it was nearly impossible for law enforcement to use it. The FBI never decrypted the files.
It's hard to tell if you're being serious here. At the end of the day, Lavabit's security came down to... the FBI's inability to read an ultra tiny font?
That's not the actual story (like, that happened, but all it did was provoke the DOJ), but it's remarkable to see someone cite that as a success for Levison.
> It didn't look like anything they had seen before so they didn't think anything of it.
You will miss your government's incompetence when they start rectifying these failures. I know I do.
They are the tyrants who oppress us with a clean conscience. They never rest and have infinite money. Their occasional episodes of bumbling incompetence is the only respite, but those won't last. Especially now with AI to think for them.
Ya. Especially when were at the point they know you know about hidden encrypted disks because they have your full internet history for the last 10 years.
I guess it's human nature to build the Torment Nexus.
> Most "state-level" adversaries are just high school dropouts working border jobs told to "investigate suspicious things".
The state-level threat isn't just about the direct employees' competence. They also have the funding to purchase tools made by people who really know what they are doing. Think of things like Pegasus spyware or Darksword iPhone hacking.
I can believe that a keystone cop wouldn't know the difference between a bong and a flower vase, but border crossing guards will know a USB drive when they see one and pass a copy of it to someone who knows what they are doing.
That's kinda the defining characteristic of a state-level threat. They have all the money they could ever need. The reason Pegasus can charge that much is because there are people able and willing to pay that much.
Anyone using forensic software can dump an image of your entire drive including all partitions, byte for byte copy. Just because you got away with being careless once doesn't mean this will work every time. If you don't want your files to be copied in encrypted or plain unencrypted form, don't bring them with you. You're better off re-downloading them over the web at that point.
Trying to keep data on a USB drive hidden long term is a losing game is the thing.
If you're intercepted with anything of value what you want is to rapidly destroy it.
This was the big problem I found when I was trying to figure out a secure way use Yubikeys: they're both long term unreliable, and short term impossible to wipe or destroy fast enough.
Literally a printout of 1 time passwords has better security properties in an adversarial situation.
For a storage device what you want isn't a secure store, what you want is a pin pull or switch which wipes it instantly.
> If you're using off-the-shelf "hidden" encrypted volume schemes, you're not going to be evading state-level adversaries;
I kinda assumed this was a solved problem now, up to the information-theoretic limit. Just create a FAT32 fs (for example), put some boring decoy files on it, and then unmount it. Traverse the FAT to map which blocks are in-use vs not-in-use, and use the not-in-use ones to store your encrypted data in some arbitrary header-free indistinguishable-from-random format. The result should look just like someone had set up a FAT filesystem on top of a securely erased drive.
If anyone ever mounts and subsequently writes to the FAT volume they'll clobber your data - oh well.
The remaining issue is that you need some special software on the host to mount the hidden volume, which itself raises suspicions. This could be solved by bundling it with OSes by default.
Instead of FAT32, use the flash drive mapping. Flash drive contain mapping from sectors to blocks. This is used to deal with flash erasure, wear leveling, and failures. Custom firmware could map the real filesystem to bad blocks.
Does the forensic software read the whole physical flash drive or just the logical one? They might if they want to get deleted files. Can they detect that there are lots of failures?
You need the wear levelling system for actual wear leveling, otherwise the drive won't last very long. Competent forensic analysis can check the firmware and the physical flash contents, although most forensic analysis is not especially competent.
Doesn't FAT have the concept of "badblocks" to handle the old days when sectors could randomly fail on drives and they had to work around it? It would only be slightly more suspicious but it would allow the drive to be mounted and used as a regular drive to avoid the even more suspicious unmounted volume issue. The data in the sectors would obviously be encrypted so it looks like random noise if someone does start poking around looking at it.
It does, but it's extremely unusual for the feature to be in use on modern hardware.
I thought of a much simpler approach though. Personally, it's completely normal for my USB drives to contain some OS installer dd'd to the first few GB, and to be wiped with urandom between uses. The installer .iso is typically immutable and will never expand beyond its initial size. Rather than mess around with block mappings, I can just use last n GB as a hidden storage volume. dm-crypt "plain" mode is headerless and can be configured to start from some offset within the underlying block device.
The only giveaway here might be forensic examination of the drive's internal wear-leveling data.
>This could be solved by bundling it with OSes by default.
Apple and Microsoft? I trust apple a bit more and MS none at all.
Linux is a bit different because the kernel is the OS, but only a small part of all its different packages. Also you'll have to watch it like a hawk because nation states will constantly try to compromise it.
I often see comments like this on encryption related project. Depending on the reader you make what might be an essential point - but I also don't really know how useful it is to say it here. The author makes no claims about what adversaries this should be sufficient against and is completely open about their constraints and decisions (allowing you to make the critique in the first place).
Basically evading a state-level actor is going to be hard. People do get lucky but to evade them in a straight-up contest you need to very clever (if it's possible at all). Given that, I just dunno how useful it is to emblazen warnings on AES-based encryption schemes that it won't fool the NSA. Either the person reading it is not technically equipped to evaluate the claims and may be scared off of a solution that protects against some adversaries while searching for an impossible zero-config state-proof solution. Or the person already knew as soon as they saw 'AES' that this isn't the project they need.
I don't mean to denigrate the inclination to highlight weaknesses like this - good communication around this is genuinely challenging.
I'm generally pretty fatalistic about avoiding state-level adversaries. So, in a sense, it doesn't really matter how one thinks a security countermeasure is going to stack up against an IC attacker. The important thing is that people understand how hard this is to do, and take that into consideration before adopting tools like these. You can easily make things worse for yourself.
I think this post is a fun technical case study on its own. It only takes on urgency because it literally markets itself as a tool to slip past state adversaries. I think it's important that people understand that it isn't the state itself that's going to take the time to detect something like this; it's some commercial forensics vendor they use automatically.
My first take is that it is terrible terrible opsec to talk about this sort of thing if you actually want to take it across a border or have any kind of "elevated" encounter with law enforcement or intelligence.
Like if they have good tools for profiling you from your posts they'll totally expect the OP to have some interesting hardware and stories to tell. Contrast that to just being some boring guy who has a cover story that happens to be true or the colorful character who doesn't really have anything to hide but who can keep spooks talking for hours.
A while ago I participated in security design for an open source foundation. (Security is not a specialty of mine, but I aspire to be competent enough for my role as a developer working in web dev and open source.) We attempted to design our processes to assure some level of defense against state-level threat actors despite assuming we could not keep them out of our network. The main goal was to ensure an audit trail in the event that our source repositories or release archives were compromised and changed.
> If you're using off-the-shelf "hidden" encrypted volume schemes, you're not going to be evading state-level adversaries; if you can find these projects and conveniently use them, state vendors can and will write scanners that find them. They're paid to do it; new detections are how they get to charge for maintenance and new versions.
How would those scanners work? If the device presents itself as a bog-standard usb drive, with all the same identification strings, etc., and the hidden volume is completely inaccessible until triggered (e.g. like swapping on sd card for another), it seems like such a scanner would have to rely on things like precise timing and performance characteristics vs what the drive presents as, and there are a lot of legitimate USB drives out there to characterize.
Edit: and if that's the case, maybe pick a lower-performance drive to mimic (and use its case). Then have some kind of characterization step to add custom delays to match its timing and performance characteristics.
> (2) Your actual security comes down to the strength of the encrypted volume, and this is 20-year-old encryption. He can't use a memory-hard KDF because of his BOM, which, like, fair enough, but that doesn't change the fact that this KDF is probably ~50x faster than standard bcrypt hardness and on realistic human passwords is probably crackable in minutes-to-hours on a dedicated rig.
I agree. What's the point of doing on-device encryption? Seems like this device should only do the volume-swapping bit. Once it "it unmounts [the unhidden section] and remounts the second hidden section, you should have to decrypt that hidden section with soemthing like VeraCrypt.
I'll also note that anyone plugging this into a forensics machine is going to have a bit of a dilemma: Perform write tests, or not? Because if they probe it with password:... attempts to check on timing leakage, that's potentially damaging evidence.
You can eliminate the problem entirely by using a random full-allowed-character-set password. The whole point of KDFs is that the password/passphrase is predictable (usually: by being generated from a seed dictionary).
Random is important, character set not so much. And yes when I say 2 characters I mean random ones. Even 26x26 is plenty to compensate. Or you can add one random word.
We're both saying the same thing: you can make the KDF irrelevant if you (as you just specified) use an AES key as an input. The entire point of KDF is that humans don't do that.
No, what I'm saying is: If your password was good enough for a quality KDF, then improving it to use a 50x worse KDF is very easy. If the worse one requires you "use an AES key", then the good one pretty much also required you to "use an AES key". If the good one didn't require that, then the 50x worse one also doesn't require that. Generate a single letter and slap it on the end, or human-pick two letters.
50x is not the decider between allowing a good password and allowing a bad password. It's a tiny little nudge.
Edit: To put "little nudge" another way, if you have a 1-10 scale of password quality, most steps in that scale are going to be more than 5.6 bits apart.
This is what GNU shred does. It has a high speed PRNG meant to avoid performance issues from directly using /dev/[u]random. Point it at your preferred block device or partition and let it fill your drive with entropy. With modern flash devices there is still the issue of hidden reserved space for block replacement but you will cover most of the drive. Reduce iterations to 1 since the default is meant for legacy hard drives and will needlessly add erase cycles to flash.
Be careful in countries like UK which have laws requiring the decryption of any data they find. There are people sitting in jail because they couldn't decrypt the "random noise" on their computers.
This is exactly the point. If storing random data was the norm (formatting utilities would offer it) then it would be hard to make a case against it.
And it actually _does_ make sense for formatting utilities to offer it, as this is a good way to ensure that data is really gone after formatting (writing all 0s or all 1s could leave information at the physical layer; and any intermediate layer could not compress random data much so the data must be overwritten).
There literally are not. Anyone competent enough to be an expert witness will be able to plainly explain to everyone else how statistical analysis obviously delinates the difference between truly random noise and an encrypted volume.
I heard one argue that "zooming" on an image by adding pixels with an algorithm is normal, and he's had such "evidence" used in court before. I was a little horrified as someone who has worked on digital forensic software at a former employer. Adding pixels to an image is not evidence, you're fabricating something out of thin air, especially if its drastically smaller, its even more likely to be BS.
Even if that is true, I have better things to do with my time and money than pay a expert witness to explain: see, I wasn't breaking the law, I just wanted the cops to think I did. That just seems like a really bad idea.
Depends on what you mean, but my subtext was that a lot of steganographic designs have turned out to be detectable; it's a design-level concern, where most consumer encryption problems are implementation issues.
One possible trick: round down the normal volume size to the nearest 1Gb and encrypt it (who doesn't use encryption today?!?) and use it for non-secret data with a password that you're willing to divulge if pressed. This is innocuous enough.
Then use the rest of the drive (~1Gb) for a hidden TrueCrypt volume with a secure key.
> Concealing an encrypted volume jacks suspicion way up; whatever pickle you were going to be in if you just kept an encrypted DMG on your desktop, you're trebly in now.
Time to bust out the old "Can't tell if this MP3 is a steganographically important encrypted file or just a Nurse With Wound EP" trick then.
I want something like this in a keyboard: a completely hidden drive that the OS cannot see until I type in a certain string of characters (password) to activate it with a separate string of characters to deactivate it.
Plausible deniability requires people to believe that you don't know about any so called "hidden" volumes. That goes out the window once you buy Hidden Drive from Hidden Drive Company.
There are other issues with AES-CTR - an attacker can flip bits at chosen offsets in the plaintext without knowing the key. For example, if they know (or can guess/infer) what offset the relevant portion of the `sudo` binary is at, they could flip the branch instruction responsible for password checking. This can be done in one step, as opposed to the multi-step plaintext recovery attack described in the article.
With XTS, the attacker is forced to completely garble a whole block at a time.
If you can do without Windows, I would say formatting a USB with a first partition as an 8 gig DOS partition and a 2nd partition as Linux LUKS may work just a good.
But a few people may know about Linux. So if on OpenBSD, you can create a second OpenBSD encrypted partition. I think that will probably look like garbage to 99.999% of the people you may be hiding from.
Be aware, if dealing with a sophisticated Gov agency, all bets are off.
>Many places don’t respect privacy laws, in certain situations you may be forced to unencrypt your media, or worse, assumed to be guilty. A Veracrypt hidden volume is useful in the former situation, but not the latter.
> ...
>Due to AI demand the cost of eMMC memory is unusually high, so I chose to go with an SD card for memory. Someone will find your SD card if they tear it apart, but of course everything is encrypted.
So which one is it? It sounds like this is trading off between "having veracrypt installed on your computer" vs "having a custom usb drive". In other words, hiding something in software vs hardware. I'm not really convinced hardware is better. The hardware device is going to get discovered if the x-ray it, not least because of the super suspicious embedded sd card reader. Even if we concede that hiding in hardware is better, surely you can just use whatever trick to hide the software (eg. putting in a secret password mounts a second drive with veracrypt)? That way you don't run into the crypto issues that the parent poster mentioned.
If you're using off-the-shelf "hidden" encrypted volume schemes, you're not going to be evading state-level adversaries; if you can find these projects and conveniently use them, state vendors can and will write scanners that find them. They're paid to do it; new detections are how they get to charge for maintenance and new versions.
Then you're down to two issues:
(1) Concealing an encrypted volume jacks suspicion way up; whatever pickle you were going to be in if you just kept an encrypted DMG on your desktop, you're trebly in now.
(2) Your actual security comes down to the strength of the encrypted volume, and this is 20-year-old encryption. He can't use a memory-hard KDF because of his BOM, which, like, fair enough, but that doesn't change the fact that this KDF is probably ~50x faster than standard bcrypt hardness and on realistic human passwords is probably crackable in minutes-to-hours on a dedicated rig.
I winced particularly at the observation that the hardness was set where it is because of how long it takes to open the encrypted volume on this hardware, because whatever scheme is being used to run the KDF now, an attacker won't bother; they'll just copy the bits and attack them on serious hardware.
Point (1) isn't meaningful if your adversaries aren't states. It might make sense to have a hidden encrypted volume for the same reason you'd want an encrypted safe. But point (2) applies to everybody.
>If you're using off-the-shelf "hidden" encrypted volume schemes, you're not going to be evading state-level adversaries;
This would evade 99% of them.
Most "state-level" adversaries are just high school dropouts working border jobs told to "investigate suspicious things".
I once had a police officer literally raid my room for drugs during a massive party we were throwing in university, for example.
They took everything. They found everything. Everything except for the $1,000 piece of drug paraphernalia I had sitting right out in open view in the middle of the apartment... the Volcano vaporizer.
It didn't look like anything they had seen before so they didn't think anything of it.
Same would be true with this USB drive in 99% of border crossing privacy invasions... plug in USB, ask for password, read it/copy it/whatever... they're not taking it apart and performing open heart surgery on a $5 USB drive at the border.
They might throw it out... or keep it or something... sure... but in general... this is a hell of a lot better than a regular USB drive imho.
This is self-soothing, not a real security plan. It doesn't matter how competent a security agency is, because there's a whole ecosystem of vendors selling into that space; knowledge of how to attack encrypted disks like this is outsourced, and, importantly, those firms have incentives to mop up even random stuff like this, because vendors will be selected in part based on lists of how many different circumvention and privacy tools they can detect.
This isn't stand-up-to-state-actors security. This is keep-state-actors-from-noticing security. They know you're guilty, they'll get it. The XKCD version of password cracking works no matter how secure your hardware.
But while states realistically can crack most anything it doesn't mean they can crack *every*thing. There are simply too many flash drives crossing the border every day.
But there is a big flaw I see: 8gb. You're carting around an 8gb flash drive these days??
We're talking about exactly the same threat. What I'm pointing out is that state adversaries won't have to "notice" this; all they'll have to do is plug the device into a standard commercial forensics scanner product --- and we're stipulating that they're plugging the device in already (else what does it matter what bits are on it).
They pay other people to notice this stuff for them!
I think part of the entire design is that it's quite hard to detect.
1: Reads/Writes are just routed to a COTS SD-card. 2: Unless the (correct?) password is detected in the write-data that starts the disconnect procedure.
The only way to detect it from what I can see is to profile writes then append a "password:" string multiple times to measure the write-delay, only works if the CPU cost is large enough to overtake the SD-card measurably and some constant-time optimizations should make it more or less undetectable.
It's a clever design and I think it should be possible to optimize to become more or less unrecognizable.
> The only way to detect it from what I can see is to profile writes then append a "password:" string multiple times to measure the write-delay,
- Have the first check be a simple 8bit hash that filters out most passwords in microseconds, or use a customizable prefix instead of “password:”.
- have your password checking thread run in background at idle priority
- when you get an async password match, force usb disconnect and reconnect and the system will rescan the bus and mount your real drive.
Short of adversary dumping drive firmware (or them figuring out your hn account t, having a LLM scan the messages and finding this conversation) it’s not really easily detectable..
If I was building a black box to detect hidden data on a USB stick, I'd include a feature whereby it measures power consumption and flags USB drives that don't consume the expected power for that type of drive.
What an unreasonably complex device to imply you’re going to make 50,000 copies of and ship out to every border agent in the nation after telling them every SINGLE flash drive crossing the border MUST be scanned. That’s so much wasted time, energy, and money. Are you building a similar device for SD cards? HDDs? SSDs? External drives? You’re going to build a database of exact charge draws for every device in existence which happens to have some form of storage on it?
Fucking up every single import and traveler on the off-chance one in seven billion scans detects one of these devices and, further, carries some data the US cares enough about to chase down to this degree?
No. Nope. Nuh uh. Insanely unreasonable. You will get ONE at a station, which you only bother to use if you can convince your $12/hr employee to even give a shit. Literally an insane proposition.
Can you flesh that out? If I've missed something and oversimplified the design here, I'd want to clarify that!
> They pay other people to notice this stuff for them!
Money, even at state level, is not some infinite resource (and neither is time).
States put effort into what they consider worth investigating - no state strip searches every incoming traveller, and goes through every item in their possession - it's possible for them to do, and if they did they'd find a hang of a lot more contraband, but it's costly, time consuming, and politically risky.
He's saying they just plug your USB stick into some gizmo sold to the state, and it's going to find enough to escalate it.
If the idea is that there's no point in encrypting anything on a USB stick, and you just hope they don't look at it at all, sure.
But there is no threat model here that includes "check the USB stick" and does not automatically lead to finding and breaking "hidden" content.
Try to explain how the hidden content would be found? It effectively doesn’t exist until the password is in place. It isn’t hidden, it isn’t connected at all.
I’m no expert but I can imagine a scenario where an x-ray scan can notice an SD card embedded into the controller and flag that for follow-up
SD card the author says was added to offset the cost of eMMC but would presumably stand out as anomalous with whatever AI is classifying the scan
> The XKCD version of password cracking works no matter how secure your hardware.
Lavabit would like to enter the conversation.
The xkcd approach worked in that case; the wrench in that case was legal, not physical, but boy would it have hurt nonetheless.
He deleted everything ultimately. He printed it out in ultra tiny font in order to comply, but it was nearly impossible for law enforcement to use it. The FBI never decrypted the files.
What about the other Lavabit users whose data he had handed over previously, before deciding to say no?
It's hard to tell if you're being serious here. At the end of the day, Lavabit's security came down to... the FBI's inability to read an ultra tiny font?
That's not the actual story (like, that happened, but all it did was provoke the DOJ), but it's remarkable to see someone cite that as a success for Levison.
I was arrested once and interrogated in great detail about how to access the data on a bluetooth usb dongle.
Just to say, I know what you mean.
> It didn't look like anything they had seen before so they didn't think anything of it.
You will miss your government's incompetence when they start rectifying these failures. I know I do.
They are the tyrants who oppress us with a clean conscience. They never rest and have infinite money. Their occasional episodes of bumbling incompetence is the only respite, but those won't last. Especially now with AI to think for them.
Ya. Especially when were at the point they know you know about hidden encrypted disks because they have your full internet history for the last 10 years.
I guess it's human nature to build the Torment Nexus.
You are conflating „state level actor” with „police level action”.
There is a world of difference if you get searched by your local police and a country level agency.
There are military units dealing with forensic investigations they have 100x or 1000x more budget and much smarter people.
But they don’t care about some drug dealers. They don’t waste time on silly stuff.
> Most "state-level" adversaries are just high school dropouts working border jobs told to "investigate suspicious things".
The state-level threat isn't just about the direct employees' competence. They also have the funding to purchase tools made by people who really know what they are doing. Think of things like Pegasus spyware or Darksword iPhone hacking.
I can believe that a keystone cop wouldn't know the difference between a bong and a flower vase, but border crossing guards will know a USB drive when they see one and pass a copy of it to someone who knows what they are doing.
A copy of this USB drive would show nothing from the encrypted portion.
Pegasus spyware operations cost something like $500k a pop, last I checked.
That's kinda the defining characteristic of a state-level threat. They have all the money they could ever need. The reason Pegasus can charge that much is because there are people able and willing to pay that much.
Can you check again and share where you're getting that number?
It may have gotten cheaper
https://prodefence.io/news/pegasus-spyware-operating-costs-c...
Thanks! Per-seat licensing here checks out.
Anyone using forensic software can dump an image of your entire drive including all partitions, byte for byte copy. Just because you got away with being careless once doesn't mean this will work every time. If you don't want your files to be copied in encrypted or plain unencrypted form, don't bring them with you. You're better off re-downloading them over the web at that point.
Unless they disassemble this drive, a forensic dump will only show the open partition.
Trying to keep data on a USB drive hidden long term is a losing game is the thing.
If you're intercepted with anything of value what you want is to rapidly destroy it.
This was the big problem I found when I was trying to figure out a secure way use Yubikeys: they're both long term unreliable, and short term impossible to wipe or destroy fast enough.
Literally a printout of 1 time passwords has better security properties in an adversarial situation.
For a storage device what you want isn't a secure store, what you want is a pin pull or switch which wipes it instantly.
> If you're using off-the-shelf "hidden" encrypted volume schemes, you're not going to be evading state-level adversaries;
I kinda assumed this was a solved problem now, up to the information-theoretic limit. Just create a FAT32 fs (for example), put some boring decoy files on it, and then unmount it. Traverse the FAT to map which blocks are in-use vs not-in-use, and use the not-in-use ones to store your encrypted data in some arbitrary header-free indistinguishable-from-random format. The result should look just like someone had set up a FAT filesystem on top of a securely erased drive.
If anyone ever mounts and subsequently writes to the FAT volume they'll clobber your data - oh well.
The remaining issue is that you need some special software on the host to mount the hidden volume, which itself raises suspicions. This could be solved by bundling it with OSes by default.
Instead of FAT32, use the flash drive mapping. Flash drive contain mapping from sectors to blocks. This is used to deal with flash erasure, wear leveling, and failures. Custom firmware could map the real filesystem to bad blocks.
Does the forensic software read the whole physical flash drive or just the logical one? They might if they want to get deleted files. Can they detect that there are lots of failures?
You need the wear levelling system for actual wear leveling, otherwise the drive won't last very long. Competent forensic analysis can check the firmware and the physical flash contents, although most forensic analysis is not especially competent.
Doesn't FAT have the concept of "badblocks" to handle the old days when sectors could randomly fail on drives and they had to work around it? It would only be slightly more suspicious but it would allow the drive to be mounted and used as a regular drive to avoid the even more suspicious unmounted volume issue. The data in the sectors would obviously be encrypted so it looks like random noise if someone does start poking around looking at it.
It does, but it's extremely unusual for the feature to be in use on modern hardware.
I thought of a much simpler approach though. Personally, it's completely normal for my USB drives to contain some OS installer dd'd to the first few GB, and to be wiped with urandom between uses. The installer .iso is typically immutable and will never expand beyond its initial size. Rather than mess around with block mappings, I can just use last n GB as a hidden storage volume. dm-crypt "plain" mode is headerless and can be configured to start from some offset within the underlying block device.
The only giveaway here might be forensic examination of the drive's internal wear-leveling data.
>This could be solved by bundling it with OSes by default.
Apple and Microsoft? I trust apple a bit more and MS none at all.
Linux is a bit different because the kernel is the OS, but only a small part of all its different packages. Also you'll have to watch it like a hawk because nation states will constantly try to compromise it.
I often see comments like this on encryption related project. Depending on the reader you make what might be an essential point - but I also don't really know how useful it is to say it here. The author makes no claims about what adversaries this should be sufficient against and is completely open about their constraints and decisions (allowing you to make the critique in the first place).
Basically evading a state-level actor is going to be hard. People do get lucky but to evade them in a straight-up contest you need to very clever (if it's possible at all). Given that, I just dunno how useful it is to emblazen warnings on AES-based encryption schemes that it won't fool the NSA. Either the person reading it is not technically equipped to evaluate the claims and may be scared off of a solution that protects against some adversaries while searching for an impossible zero-config state-proof solution. Or the person already knew as soon as they saw 'AES' that this isn't the project they need.
I don't mean to denigrate the inclination to highlight weaknesses like this - good communication around this is genuinely challenging.
I'm generally pretty fatalistic about avoiding state-level adversaries. So, in a sense, it doesn't really matter how one thinks a security countermeasure is going to stack up against an IC attacker. The important thing is that people understand how hard this is to do, and take that into consideration before adopting tools like these. You can easily make things worse for yourself.
I think this post is a fun technical case study on its own. It only takes on urgency because it literally markets itself as a tool to slip past state adversaries. I think it's important that people understand that it isn't the state itself that's going to take the time to detect something like this; it's some commercial forensics vendor they use automatically.
My first take is that it is terrible terrible opsec to talk about this sort of thing if you actually want to take it across a border or have any kind of "elevated" encounter with law enforcement or intelligence.
Like if they have good tools for profiling you from your posts they'll totally expect the OP to have some interesting hardware and stories to tell. Contrast that to just being some boring guy who has a cover story that happens to be true or the colorful character who doesn't really have anything to hide but who can keep spooks talking for hours.
A while ago I participated in security design for an open source foundation. (Security is not a specialty of mine, but I aspire to be competent enough for my role as a developer working in web dev and open source.) We attempted to design our processes to assure some level of defense against state-level threat actors despite assuming we could not keep them out of our network. The main goal was to ensure an audit trail in the event that our source repositories or release archives were compromised and changed.
> If you're using off-the-shelf "hidden" encrypted volume schemes, you're not going to be evading state-level adversaries; if you can find these projects and conveniently use them, state vendors can and will write scanners that find them. They're paid to do it; new detections are how they get to charge for maintenance and new versions.
How would those scanners work? If the device presents itself as a bog-standard usb drive, with all the same identification strings, etc., and the hidden volume is completely inaccessible until triggered (e.g. like swapping on sd card for another), it seems like such a scanner would have to rely on things like precise timing and performance characteristics vs what the drive presents as, and there are a lot of legitimate USB drives out there to characterize.
Edit: and if that's the case, maybe pick a lower-performance drive to mimic (and use its case). Then have some kind of characterization step to add custom delays to match its timing and performance characteristics.
> (2) Your actual security comes down to the strength of the encrypted volume, and this is 20-year-old encryption. He can't use a memory-hard KDF because of his BOM, which, like, fair enough, but that doesn't change the fact that this KDF is probably ~50x faster than standard bcrypt hardness and on realistic human passwords is probably crackable in minutes-to-hours on a dedicated rig.
I agree. What's the point of doing on-device encryption? Seems like this device should only do the volume-swapping bit. Once it "it unmounts [the unhidden section] and remounts the second hidden section, you should have to decrypt that hidden section with soemthing like VeraCrypt.
Given that the name defaults to "Phantomdrive" I have a hunch that it is not as fully nondescript as you are suggesting it ought to be.
This is probably fine for evading local cops.
I'll also note that anyone plugging this into a forensics machine is going to have a bit of a dilemma: Perform write tests, or not? Because if they probe it with password:... attempts to check on timing leakage, that's potentially damaging evidence.
You can compensate for not just 50x but 500x by adding 2 characters to your password. If that's the speedup factor, it's not itself a big worry.
You can eliminate the problem entirely by using a random full-allowed-character-set password. The whole point of KDFs is that the password/passphrase is predictable (usually: by being generated from a seed dictionary).
Random is important, character set not so much. And yes when I say 2 characters I mean random ones. Even 26x26 is plenty to compensate. Or you can add one random word.
We're both saying the same thing: you can make the KDF irrelevant if you (as you just specified) use an AES key as an input. The entire point of KDF is that humans don't do that.
No, what I'm saying is: If your password was good enough for a quality KDF, then improving it to use a 50x worse KDF is very easy. If the worse one requires you "use an AES key", then the good one pretty much also required you to "use an AES key". If the good one didn't require that, then the 50x worse one also doesn't require that. Generate a single letter and slap it on the end, or human-pick two letters.
50x is not the decider between allowing a good password and allowing a bad password. It's a tiny little nudge.
Edit: To put "little nudge" another way, if you have a 1-10 scale of password quality, most steps in that scale are going to be more than 5.6 bits apart.
Someone should write a disk formatting utility that randomizes the bits on your drive.
Now you can claim that any encrypted drive (which just looks like pure noise if done correctly) is a drive that was just formatted.
This is what GNU shred does. It has a high speed PRNG meant to avoid performance issues from directly using /dev/[u]random. Point it at your preferred block device or partition and let it fill your drive with entropy. With modern flash devices there is still the issue of hidden reserved space for block replacement but you will cover most of the drive. Reduce iterations to 1 since the default is meant for legacy hard drives and will needlessly add erase cycles to flash.
Be careful in countries like UK which have laws requiring the decryption of any data they find. There are people sitting in jail because they couldn't decrypt the "random noise" on their computers.
This is exactly the point. If storing random data was the norm (formatting utilities would offer it) then it would be hard to make a case against it.
And it actually _does_ make sense for formatting utilities to offer it, as this is a good way to ensure that data is really gone after formatting (writing all 0s or all 1s could leave information at the physical layer; and any intermediate layer could not compress random data much so the data must be overwritten).
Fair warning; any links to specific cases (not the law itself)?
This lists two unnamed individuals.
https://edri.org/our-work/edri-gramnumber7-16decrypt-data-ca...
> There are people sitting in jail because they couldn't decrypt the "random noise" on their computers.
No, there isn't.
There literally are not. Anyone competent enough to be an expert witness will be able to plainly explain to everyone else how statistical analysis obviously delinates the difference between truly random noise and an encrypted volume.
If there's a difference visible, that's a weakness in the encryption algorithm.
Encrypted data should be statistically indistinct from random noise.
It's an information theory thing. Everything in the ciphertext should be sitting at maximum entropy.
Loll, you have a lot of faith in the legal system that I do not. Expert witnesses are commonly full of shit.
I heard one argue that "zooming" on an image by adding pixels with an algorithm is normal, and he's had such "evidence" used in court before. I was a little horrified as someone who has worked on digital forensic software at a former employer. Adding pixels to an image is not evidence, you're fabricating something out of thin air, especially if its drastically smaller, its even more likely to be BS.
Are you saying any photos I take with phone that had digital zoom is automatically I admissible? Or if not, isn't that consistent with what they said?
but... how _else_ would you Enhance the image?
Even if that is true, I have better things to do with my time and money than pay a expert witness to explain: see, I wasn't breaking the law, I just wanted the cops to think I did. That just seems like a really bad idea.
> the difference between truly random noise and an encrypted volume
A LUKS volume with a detached header is indistinguishable from random data.
Sounds a lot like stenography.
https://github.com/dishather/steganodisk
The security track record of steganography is not great.
Isn't the same true of consumer grade encryption? In a perfect world, they should be used together.
Depends on what you mean, but my subtext was that a lot of steganographic designs have turned out to be detectable; it's a design-level concern, where most consumer encryption problems are implementation issues.
That would just be dd if=/dev/urandom of=/dev/path/to/device
One possible trick: round down the normal volume size to the nearest 1Gb and encrypt it (who doesn't use encryption today?!?) and use it for non-secret data with a password that you're willing to divulge if pressed. This is innocuous enough.
Then use the rest of the drive (~1Gb) for a hidden TrueCrypt volume with a secure key.
> Concealing an encrypted volume jacks suspicion way up; whatever pickle you were going to be in if you just kept an encrypted DMG on your desktop, you're trebly in now.
Time to bust out the old "Can't tell if this MP3 is a steganographically important encrypted file or just a Nurse With Wound EP" trick then.
I want something like this in a keyboard: a completely hidden drive that the OS cannot see until I type in a certain string of characters (password) to activate it with a separate string of characters to deactivate it.
Plausible deniability requires people to believe that you don't know about any so called "hidden" volumes. That goes out the window once you buy Hidden Drive from Hidden Drive Company.
Probably better just to use a veracrypt hidden volume, and preferably not a SSD.
Might as well be rot13.
If you rented 8 MI300X's or the nvidia equivalent, I don't even think an unreasonably long password would matter.
It should finish quickly enough that you would be upset with all of the money you have now wasted by having to commit to a month of utilization
If you need 8 MI300X's for a month to do ROT13, then you're doing it wrong.
Just a heads-up, your blog post has a broken link for
> As for the security of the device, I’ve verified things with functional tests like this
Otherwise great project!
There are other issues with AES-CTR - an attacker can flip bits at chosen offsets in the plaintext without knowing the key. For example, if they know (or can guess/infer) what offset the relevant portion of the `sudo` binary is at, they could flip the branch instruction responsible for password checking. This can be done in one step, as opposed to the multi-step plaintext recovery attack described in the article.
With XTS, the attacker is forced to completely garble a whole block at a time.
Can this be done on boards with the same layouts as known commercial USB keys i.e will imaging just show this to be what it is?
Could use an Opal or Pyrite drive connected to a USB caddy and sedutil-cli to manage it.
Good stuff! If I remember correctly there was company at Defcon selling memory cards that had this feature built into them a couple years ago?
I miss TrueCrypts alt boot option. You enter a specific password it boots an alt OS so you could have an automatic alternative thing if under duress.
Rule no. 1: you don't announce your security by obscurity methods :)
>Let’s now assume our attacker has a really powerful machine, has gotten your salt and now will compute their own “table”.
You should be using a TPM so such an attack is impossible. The TPM should be what is rate limiting the guesses. You don't need a KDF here.
A very interesting solution!
If you can do without Windows, I would say formatting a USB with a first partition as an 8 gig DOS partition and a 2nd partition as Linux LUKS may work just a good.
But a few people may know about Linux. So if on OpenBSD, you can create a second OpenBSD encrypted partition. I think that will probably look like garbage to 99.999% of the people you may be hiding from.
Be aware, if dealing with a sophisticated Gov agency, all bets are off.
But this partition would be visible.
The OP had a version that is invisible.
Hmm, a partition of any sort has mapped the territory.
Dude. Veracrypt. Hidden volumes. This is an old and well-solved problem set.
Hidden volumes aren’t that hidden if Veracrypt can find them.
Encrypted USB drives, there are definite legal uses, but like others have said, it raises suspicion.
The whole point is you are unable to detect that the drive is encrypted because it is only visible to USB once the password is written.
The whole point is asking what is so sensitive.
In this case you can just crack it open and see that the SD card inside is larger than the 8 GB block device exposed.
>Many places don’t respect privacy laws, in certain situations you may be forced to unencrypt your media, or worse, assumed to be guilty. A Veracrypt hidden volume is useful in the former situation, but not the latter.
> ...
>Due to AI demand the cost of eMMC memory is unusually high, so I chose to go with an SD card for memory. Someone will find your SD card if they tear it apart, but of course everything is encrypted.
So which one is it? It sounds like this is trading off between "having veracrypt installed on your computer" vs "having a custom usb drive". In other words, hiding something in software vs hardware. I'm not really convinced hardware is better. The hardware device is going to get discovered if the x-ray it, not least because of the super suspicious embedded sd card reader. Even if we concede that hiding in hardware is better, surely you can just use whatever trick to hide the software (eg. putting in a secret password mounts a second drive with veracrypt)? That way you don't run into the crypto issues that the parent poster mentioned.
Could've at least sealed the SD card in an epoxy blob.