7 ms·

“End-to-End incompatible with Chrome Update functionality”

28 points by hodgesmr 12y ago

McDiesel 12y ago
Since when does the google store update an extension not installed through the google store?
- malka 12y ago
  well. Probably as soon as the NSA requests it.
opendais 12y ago
This is the most likely attack vector for any automatically updating program on your computer, not just End-to-End. What would stop a maliciously updated Chrome from recording all your keystrokes in the browser [all of your password; passphrases] as well as copying where you were at the time? Etc. This is one of those things where if you think the Government is going to silence you for being a dissenting voice and/or steal your info because of it...you grab an open source project, you compile it, you use that. You don't grab closed source software that automagically updates. I don't think its reasonable to expect Google to protect the user from -every- potential attack vector.
- simias 12y ago
  I think the reason many people (including myself) have a certain reticence to this google made crypto-in-the-browser is that nowadays I feel the #1 reason to use such an extension would be to protect ourselves from eavesdropping from Google or the government. So basically you're trusting Google for giving you the tools to escape Google's scrutiny, knowing full well that if they really want to access your messages they can easily backdoor your crypto extension. So what's the point? To me it's just homeopathic crypto.
  - raldi 12y ago
    You have access to all the code and can read it and compile it yourself, and then turn off automatic updating. And if you're not tech-savvy enough to do that, you can hire someone else to. And if even that's not enough to please you, then tell me, what could possibly satisfy you?
    - simias 12y ago
      Use a third party tool instead of putting all your trust into a single point of failure. There are plenty to choose from. What you're proposing is pretty complex and time consuming, you can't expect the average user of this extension to do that (I know I wouldn't). Or alternatively decide that you're fine with trusting google with whatever info is in your emails because you believe they have to much to lose if they're caught eavesdropping on some random Joe like me. That's what I do. And if tomorrow I feel I have a mail sensitive enough that I don't want to risk Google or some government intercepting it I certainly won't do it from my web browser's using Google's own extension.
      - raldi 12y ago
        How does using a third-party tool solve anything? Its code probably won't have as many skilled eyeballs reviewing it.
        simias 12y ago
        Maybe, maybe not, GnuPG has been here a long time already for instance. But that's not the point anyway, the point is that Google can update your browser and the extension without you being aware of it very easily. You can turn it off of course, but who does that? Running an outdated browser is not exactly the best security measure...
raldi 12y ago
It doesn't seem like they're ignoring it. They just don't have a quick fix for this hard problem. If you've got an easy answer, please post it on the bug!
comboy 12y ago
It's open source and you don't have to install it through the google store.
danielweber 12y ago
What's the precedent for a company ever being required to ship backdoored products to their customers by government legal order, NSL'd or otherwise? "The FBI once lobbied for the government to give them that power!" or "look at these service companies that had evidence in their possession they were required to turn over!" are non-answers.
- hodgesmr 12y ago
  http://en.wikipedia.org/wiki/RSA_BSAFE http://en.wikipedia.org/wiki/RSA_BSAFE
  - vezzy-fnord 12y ago
    An older example. One that attracted a lot of attention in the 90s, but was soon forgotten and is now largely a historical curiosity, even though it should have caused people to be aware of the pitfalls of unsupervised intelligence way before Snowden: https://en.wikipedia.org/wiki/Clipper_chip https://en.wikipedia.org/wiki/Clipper_chip
    - danielweber 12y ago
      Who was forced to use the Clipper chip? As your link points out, all the government could do was try to encourage manufacturers to go along, but even in the crypto wars of the 90's they couldn't make anyone get on board. EDIT Did anyone use the Clipper chip in any shipping product, ever? Note that a positive answer to this still wouldn't answer the question of a company being forced to do it by government order.
  - danielweber 12y ago
    This is absolutely not the government requiring someone to do something by court order. it was also revealed that NSA had secretly paid RSA Security $10 million to use Dual_EC_DRBG This instance is unsettling -- especially to RSA customers -- but it is very much a different beast from a court or legal order requiring installation of a backdoor.
- 12y ago
Foxboron 12y ago
Couldnt this be countered by writing your own script checking the hash of the files X times a day?
- danielweber 12y ago
  Once you assume the browser you are using for secrets is actively hostile, it's a pretty Sisyphian task to try and guarantee anything.
- Igglyboo 12y ago
  Your average user is going to know how to do that or even want to do that. If they have to jump through a ton of hoops they just won't use it and be worse off.
willvarfar 12y ago
This is true if you have any program or OS upgrade going, whether manual or automatic. "Ignore" is just a linkbait title :(
- dang 12y ago
  Yes, and it was also badly editorialized. We've changed the title to be (approximately) that of the bug report. Submitters: the HN guidelines ask you not to change titles except when they are linkbait or misleading. Changing them to be more linkbaity and misleading is right out! Edit: Oops, I forgot: the submitted title was "Google ignores the simplest and most likely attack vector with End-to-End".
DanBlake 12y ago
Maybe the solution is to have the automatic downloads of chrome be anonymous and build the system in such a way that changing it would not be possible. Basically, change the chrome automatic updater to not send any identifying information when it requests a update. That way, you can be sure that Google couldn't just target 'you' with a update. Then, you just need to rely on the fact that people would be watching the chromium code for any changes which would negate the above anonymity. The real challenge would be for Google to develop a way where they could not still identify people from their other data (IP, cookies, etc..) when they were requesting a update. Maybe have a third party host/store chrome update binaries? Something like amazon S3 or something which would not data share with Google.
- AnthonyMouse 12y ago
  > Maybe have a third party host/store chrome update binaries? Something like amazon S3 or something which would not data share with Google. Obvious problem there is that whoever is putting pressure on Google could put the same pressure on Amazon. Solution is to use P2P. If your device has uploaded the new code to two other devices (and so on) before installing it, targeting specific devices becomes much more difficult.
  - DanBlake 12y ago
    using tor to download it should work pretty well, yes.
    - AnthonyMouse 12y ago
      Just using tor doesn't necessarily fix it. If you're still getting the update directly from Google then there is still the possibility that they can fingerprint your device -- this is one of the hard problems with using tor and one of the reasons you have to be so careful when using it for anything important. Also, if large companies start using tor for software updates it's liable to fall over from the load. The advantage of using P2P (meaning something in the nature of BitTorrent) is that you can't limit the scope of the update. If you upload it to the target, the target will give it to everybody else.
Zigurd 12y ago
It is a hard problem, and it's a positive sign that Google acknowledges the problem. Providing end-to-end security also goes against the trend of expanding the non-open parts of their Android app suite. In addition to the reported bug, this plugin is handing cleartext back to Google-controlled code. Web apps and good security are still miles apart. But this is still a significant change from a year ago when we heard internet portal CEO kvetching about the NSA and not even mentioning and-to-end encryption. There is still a VERY long way to go before this counts as democratizing end-to-end security. Any portal that has real time communication tools and a social graph could also provide tools for automating Web-of-trust and key exchange. All journeys etc.
antmldr 12y ago
This isn't a bug / complaint / observation of a vulnerability of End-to-End per se, you could argue Microsoft could be NSL'd to do the same to a user's operating system. To counter this you'd need a secure, distributed way to release updates in Chrome. I don't think that's quite in scope of what this project is trying to accomplish.
- AnthonyMouse 12y ago
  > To counter this you'd need a secure, distributed way to release updates in Chrome. I don't think that's quite in scope of what this project is trying to accomplish. It's not like that would require a large amount of novel code. Use BitTorrent + sign the update, problem solved.
stcredzero 12y ago
You know, if we had DRM infrastructure we could actually trust, this wouldn't be a problem. Granted, having DRM we can trust may well be itself an insurmountable problem.
- AnthonyMouse 12y ago
  Your comment doesn't parse due to an ambiguity. Who is we? If "we" are supposed to be all of us as owners of computing devices, I'm not sure how DRM is supposed to help. Maybe the ambiguity is in what you think DRM is. If "we" is "they" as the overlords who manufacture our stuff, having a DRM infrastructure they can trust would be a primary source of the problem, because it would consequently be one that we can't.
  - stcredzero 12y ago
    If "we" are supposed to be all of us as owners of computing devices, I'm not sure how DRM is supposed to help. Maybe the ambiguity is in what you think DRM is. This is exactly the knee-jerk populist fluff we need people to actually think through. If we was everyone, then we could have agents that use our private key to sign and encrypt things on our behalf, but be a bit more confident that it wouldn't fall into the wrong hands. (Actually, we'd best spawn agents with a time-limited signed ephemeral private key.) "We as in the overlords" is so predominant as a model, people have forgotten to think about the implications of such tools for themselves. Most DRM "discussion" gets reduced to "two legs baaad, four legs goood."
    - AnthonyMouse 12y ago
      > If we was everyone, then we could have agents that use our private key to sign and encrypt things on our behalf, but be a bit more confident that it wouldn't fall into the wrong hands. (Actually, we'd best spawn agents with a time-limited signed ephemeral private key.) What are we supposed to gain from "agents" here? You have your private key, nobody else has it, you sign things with it. If somebody else has it then it isn't your private key. For DRM to even pretend to work for any purpose that anyone thinks of when people say "DRM" (i.e. preventing the copying of bits given to untrusted parties), it has to be a centralized trust model. Some party that "everyone" can trust has to certify that the device or software you're giving the secret bits to is not going to divulge them. The problem is that there is nobody that everybody can trust. Trust is not a global hierarchy, it's a relativistic relationship between peers. If (as it seems) you're trying to achieve something separate that may work using a distributed trust model, why are you calling it DRM?
rlx0x 12y ago
I think its a moot point, if the government is the attacker, who tries to spy on you specifically, there is absolutely NOTHING you can do to prevent that from happening. The US government has (literally) secret laws, that grants themselves the right to go to another country and kill someone without due process or trial or any kind of repercussion if they 'accidentally' kill innocent bystanders. Its sort of laughable to talk about end to end encryption and possible NSL when you really think about it.
- fulafel 12y ago
  Here you make the mistake of assuming security is binary. Making attacks more expensive and likely to be exposed are effective deterrents. There are many things that people in us government would like to do but cannot afford it.

28 points by hodgesmr 12y ago

McDiesel 12y ago

Since when does the google store update an extension not installed through the google store?

malka 12y ago
well. Probably as soon as the NSA requests it.

opendais 12y ago

This is the most likely attack vector for any automatically updating program on your computer, not just End-to-End. What would stop a maliciously updated Chrome from recording all your keystrokes in the browser [all of your password; passphrases] as well as copying where you were at the time? Etc. This is one of those things where if you think the Government is going to silence you for being a dissenting voice and/or steal your info because of it...you grab an open source project, you compile it, you use that. You don't grab closed source software that automagically updates. I don't think its reasonable to expect Google to protect the user from -every- potential attack vector.

simias 12y ago
I think the reason many people (including myself) have a certain reticence to this google made crypto-in-the-browser is that nowadays I feel the #1 reason to use such an extension would be to protect ourselves from eavesdropping from Google or the government. So basically you're trusting Google for giving you the tools to escape Google's scrutiny, knowing full well that if they really want to access your messages they can easily backdoor your crypto extension. So what's the point? To me it's just homeopathic crypto.
- raldi 12y ago
  You have access to all the code and can read it and compile it yourself, and then turn off automatic updating. And if you're not tech-savvy enough to do that, you can hire someone else to. And if even that's not enough to please you, then tell me, what could possibly satisfy you?
  - simias 12y ago
    Use a third party tool instead of putting all your trust into a single point of failure. There are plenty to choose from. What you're proposing is pretty complex and time consuming, you can't expect the average user of this extension to do that (I know I wouldn't). Or alternatively decide that you're fine with trusting google with whatever info is in your emails because you believe they have to much to lose if they're caught eavesdropping on some random Joe like me. That's what I do. And if tomorrow I feel I have a mail sensitive enough that I don't want to risk Google or some government intercepting it I certainly won't do it from my web browser's using Google's own extension.
    - raldi 12y ago
      How does using a third-party tool solve anything? Its code probably won't have as many skilled eyeballs reviewing it.
      - simias 12y ago
        Maybe, maybe not, GnuPG has been here a long time already for instance. But that's not the point anyway, the point is that Google can update your browser and the extension without you being aware of it very easily. You can turn it off of course, but who does that? Running an outdated browser is not exactly the best security measure...

raldi 12y ago

It doesn't seem like they're ignoring it. They just don't have a quick fix for this hard problem. If you've got an easy answer, please post it on the bug!

comboy 12y ago

It's open source and you don't have to install it through the google store.

danielweber 12y ago

What's the precedent for a company ever being required to ship backdoored products to their customers by government legal order, NSL'd or otherwise? "The FBI once lobbied for the government to give them that power!" or "look at these service companies that had evidence in their possession they were required to turn over!" are non-answers.

hodgesmr 12y ago
http://en.wikipedia.org/wiki/RSA_BSAFE http://en.wikipedia.org/wiki/RSA_BSAFE
- vezzy-fnord 12y ago
  An older example. One that attracted a lot of attention in the 90s, but was soon forgotten and is now largely a historical curiosity, even though it should have caused people to be aware of the pitfalls of unsupervised intelligence way before Snowden: https://en.wikipedia.org/wiki/Clipper_chip https://en.wikipedia.org/wiki/Clipper_chip
  - danielweber 12y ago
    Who was forced to use the Clipper chip? As your link points out, all the government could do was try to encourage manufacturers to go along, but even in the crypto wars of the 90's they couldn't make anyone get on board. EDIT Did anyone use the Clipper chip in any shipping product, ever? Note that a positive answer to this still wouldn't answer the question of a company being forced to do it by government order.
- danielweber 12y ago
  This is absolutely not the government requiring someone to do something by court order. it was also revealed that NSA had secretly paid RSA Security $10 million to use Dual_EC_DRBG This instance is unsettling -- especially to RSA customers -- but it is very much a different beast from a court or legal order requiring installation of a backdoor.
12y ago

Foxboron 12y ago

Couldnt this be countered by writing your own script checking the hash of the files X times a day?

danielweber 12y ago
Once you assume the browser you are using for secrets is actively hostile, it's a pretty Sisyphian task to try and guarantee anything.
Igglyboo 12y ago
Your average user is going to know how to do that or even want to do that. If they have to jump through a ton of hoops they just won't use it and be worse off.

willvarfar 12y ago

This is true if you have any program or OS upgrade going, whether manual or automatic. "Ignore" is just a linkbait title :(

dang 12y ago
Yes, and it was also badly editorialized. We've changed the title to be (approximately) that of the bug report. Submitters: the HN guidelines ask you not to change titles except when they are linkbait or misleading. Changing them to be more linkbaity and misleading is right out! Edit: Oops, I forgot: the submitted title was "Google ignores the simplest and most likely attack vector with End-to-End".

DanBlake 12y ago

Maybe the solution is to have the automatic downloads of chrome be anonymous and build the system in such a way that changing it would not be possible. Basically, change the chrome automatic updater to not send any identifying information when it requests a update. That way, you can be sure that Google couldn't just target 'you' with a update. Then, you just need to rely on the fact that people would be watching the chromium code for any changes which would negate the above anonymity. The real challenge would be for Google to develop a way where they could not still identify people from their other data (IP, cookies, etc..) when they were requesting a update. Maybe have a third party host/store chrome update binaries? Something like amazon S3 or something which would not data share with Google.

AnthonyMouse 12y ago
> Maybe have a third party host/store chrome update binaries? Something like amazon S3 or something which would not data share with Google. Obvious problem there is that whoever is putting pressure on Google could put the same pressure on Amazon. Solution is to use P2P. If your device has uploaded the new code to two other devices (and so on) before installing it, targeting specific devices becomes much more difficult.
- DanBlake 12y ago
  using tor to download it should work pretty well, yes.
  - AnthonyMouse 12y ago
    Just using tor doesn't necessarily fix it. If you're still getting the update directly from Google then there is still the possibility that they can fingerprint your device -- this is one of the hard problems with using tor and one of the reasons you have to be so careful when using it for anything important. Also, if large companies start using tor for software updates it's liable to fall over from the load. The advantage of using P2P (meaning something in the nature of BitTorrent) is that you can't limit the scope of the update. If you upload it to the target, the target will give it to everybody else.

Zigurd 12y ago

It is a hard problem, and it's a positive sign that Google acknowledges the problem. Providing end-to-end security also goes against the trend of expanding the non-open parts of their Android app suite. In addition to the reported bug, this plugin is handing cleartext back to Google-controlled code. Web apps and good security are still miles apart. But this is still a significant change from a year ago when we heard internet portal CEO kvetching about the NSA and not even mentioning and-to-end encryption. There is still a VERY long way to go before this counts as democratizing end-to-end security. Any portal that has real time communication tools and a social graph could also provide tools for automating Web-of-trust and key exchange. All journeys etc.

antmldr 12y ago

This isn't a bug / complaint / observation of a vulnerability of End-to-End per se, you could argue Microsoft could be NSL'd to do the same to a user's operating system. To counter this you'd need a secure, distributed way to release updates in Chrome. I don't think that's quite in scope of what this project is trying to accomplish.

AnthonyMouse 12y ago
> To counter this you'd need a secure, distributed way to release updates in Chrome. I don't think that's quite in scope of what this project is trying to accomplish. It's not like that would require a large amount of novel code. Use BitTorrent + sign the update, problem solved.

stcredzero 12y ago

You know, if we had DRM infrastructure we could actually trust, this wouldn't be a problem. Granted, having DRM we can trust may well be itself an insurmountable problem.

AnthonyMouse 12y ago
Your comment doesn't parse due to an ambiguity. Who is we? If "we" are supposed to be all of us as owners of computing devices, I'm not sure how DRM is supposed to help. Maybe the ambiguity is in what you think DRM is. If "we" is "they" as the overlords who manufacture our stuff, having a DRM infrastructure they can trust would be a primary source of the problem, because it would consequently be one that we can't.
- stcredzero 12y ago
  If "we" are supposed to be all of us as owners of computing devices, I'm not sure how DRM is supposed to help. Maybe the ambiguity is in what you think DRM is. This is exactly the knee-jerk populist fluff we need people to actually think through. If we was everyone, then we could have agents that use our private key to sign and encrypt things on our behalf, but be a bit more confident that it wouldn't fall into the wrong hands. (Actually, we'd best spawn agents with a time-limited signed ephemeral private key.) "We as in the overlords" is so predominant as a model, people have forgotten to think about the implications of such tools for themselves. Most DRM "discussion" gets reduced to "two legs baaad, four legs goood."
  - AnthonyMouse 12y ago
    > If we was everyone, then we could have agents that use our private key to sign and encrypt things on our behalf, but be a bit more confident that it wouldn't fall into the wrong hands. (Actually, we'd best spawn agents with a time-limited signed ephemeral private key.) What are we supposed to gain from "agents" here? You have your private key, nobody else has it, you sign things with it. If somebody else has it then it isn't your private key. For DRM to even pretend to work for any purpose that anyone thinks of when people say "DRM" (i.e. preventing the copying of bits given to untrusted parties), it has to be a centralized trust model. Some party that "everyone" can trust has to certify that the device or software you're giving the secret bits to is not going to divulge them. The problem is that there is nobody that everybody can trust. Trust is not a global hierarchy, it's a relativistic relationship between peers. If (as it seems) you're trying to achieve something separate that may work using a distributed trust model, why are you calling it DRM?

rlx0x 12y ago

I think its a moot point, if the government is the attacker, who tries to spy on you specifically, there is absolutely NOTHING you can do to prevent that from happening. The US government has (literally) secret laws, that grants themselves the right to go to another country and kill someone without due process or trial or any kind of repercussion if they 'accidentally' kill innocent bystanders. Its sort of laughable to talk about end to end encryption and possible NSL when you really think about it.

fulafel 12y ago
Here you make the mistake of assuming security is binary. Making attacks more expensive and likely to be exposed are effective deterrents. There are many things that people in us government would like to do but cannot afford it.