5 ms·

Services like 1Password are often more secure than your solution because they need to harden vaults against full leaks. In the case of 1Password, a secret key i

by manmal 4y ago

Services like 1Password are often more secure than your solution because they need to harden vaults against full leaks. In the case of 1Password, a secret key in addition to the password ensures that brute forcing is (at the moment) not feasible, even if your password is really crappy.

notesinthefield 4y ago
Keepass has Key Files as a part of the spec https://keepass.info/help/base/keys.html https://keepass.info/help/base/keys.html On my devices, keyfiles and a KP client are stored locally. The DB rests in the cloud.
- manmal 4y ago
  Is that kind of file storage secure? I thought files stored on disk can be extracted quite easily?
brandon272 4y ago
LastPass would have also led their customers to believe that "brute forcing was not possible" and that they were taking extraordinary measures to keep vaults and data safe. I think one distinction between services like KeePass and 1Password is end user perception of how easy it is for an attacker to acquire an encrypted vault to begin with. For many, they consider a KDBX database sitting in their Dropbox account to be less likely to be stolen than an encrypted vault being held by a company like 1Password, a high value target to the most sophisticated attackers including state actors.
- hn_throwaway_99 4y ago
  Doesn't necessarily matter what LastPass "would have also led their customers to believe", the mathematical reality is still that LassPass vaults are crackable in a way that 1P vaults fundamentally are not.
  - brandon272 4y ago
    Yes, according to what 1Password is telling us. But as we've seen, what these companies say and what they actually do in practice are not always aligned. And oftentimes customers are inserting a lot of their own assumptions into the mix, not only with respect to vault encryption but vault storage and operational security.
    - manmal 4y ago
      > Yes, according to what 1Password is telling us With their very comprehensive whitepaper and Charles Proxy you can verify all their claims. Their whitepaper is one of the best resources I have found on E2EE in general. With that, you should be able to write your own 1P vault parser. Then you can verify that traffic to their server is exactly what they claim it to be. In another comment you are criticizing that their product is proprietary - that's IMO not quite true. Yes, 1P is closed source, but their crypto strategy is documented extensively - they list the exact cipher algos and settings. > not only with respect to vault encryption but vault storage and operational security That's a valid argument, BUT, if you read their whitepaper, you'll likely arrive at the conclusion that even a full leak of the encrypted vault is currently not that problematic. I wouldn't post it online, but I'm not worried if they announce a leak tomorrow.
      - brandon272 4y ago
        All valid points!
    - hn_throwaway_99 4y ago
      > Yes, according to what 1Password is telling us. But as we've seen, what these companies say and what they actually do in practice are not always aligned. That's just not accurate: 1. First off, all the encryption happens client-side. It is possible for anyone so inclined to validate how 1P and LP are doing their encryption. 2. The deficiencies in LP's encryption approach were well known for years. My point it, yes, companies will spin things how ever they want, which is why you should completely ignore what they say and only evaluate what is verifiable. And 1P's and LP's approaches are verifiably different.
princevegeta89 4y ago
Same was said about LastPass many times and look at what happened, everything turns out to be a false promise.
- hn_throwaway_99 4y ago
  That's not a fair comparison. The differences in LP and 1P encryption approaches have been well known for years, and they are fundamentally different. Now, while 1P encrypted vaults are not brute-forceable the way LP's are, that doesn't mean it's impossible to hack 1P (e.g. malicious code injection in any of their apps or plugins), but I don't like the "everything turns out to be a false promise" broad-brushing when there are real and verifiable differences in how these companies secure your data.
  - fluidcruft 4y ago
    Unless you want to spend the time doing RE of crypto (so fun!), how to do you even know who works for 1Password? Are you waiting for Snowden 2.0 to explain how things work a decade later?
DavideNL 4y ago
Note that 1Password copies the "Secret Key" to iCloud... without asking.
- manmal 4y ago
  Do you mean iCloud keychain? That's arguably one of the most secure cloud storage in existence.
  - DavideNL 4y ago
    Yes. And no, because it depends on how users setup and use their AppleID and its passwords/security/devices. My only point is, if 1Passwords decides to share the private key with [whichever cloud service] the user should at least be notified, or get a choice.
    - manmal 4y ago
      > And no, because it depends on how users setup and use their AppleID and its passwords/security/devices. Can you elaborate what the issue would be? I see that the AppleID password could be a weak link, but that's mostly mitigated by 2FA. > if 1Passwords decides to share the private key I'm not aware that they could do that. Their servers have no knowledge of either the password or secret key. Authentication happens via a zero-knowledge proof.
      - manmal 4y ago
        Ah, Apple recently improved things further: https://www.apple.com/newsroom/2022/12/apple-advances-user-security-with-powerful-new-data-protections/ https://www.apple.com/newsroom/2022/12/apple-advances-user-s...
phonebucket 4y ago
But in the context of a strong master password, the additional benefit of the secret key is of neglible benefit, while the hassle and dangers of having to synchronise the secret key remain. I'd rather use an extremely high entropy master password by itself.
- manmal 4y ago
  How do you remember a 128 bit entropy password?
  - phonebucket 4y ago
    remember 10 random words
fluidcruft 4y ago
How is 1Password's secret key more secure than using keepass's key file feature?
- manmal 4y ago
  That depends where you store that file. Is there a way to store this securely in a mobile device for use on the go?
- DavideNL 4y ago
  Or, simply a password for your Keepass database with the same entropy as 1Passwords secret key. Anyways, it depends on the app you use. Some Keepass apps haven't even been audited for example, so security of those apps may be questionable.
  - manmal 4y ago
    > Or, simply a password for your Keepass database with the same entropy as 1Passwords secret key. How would you remember a password of 128bit of entropy?
    - dagw 4y ago
      9-10 words chosen at random from a list of 10k common words in you native language will get you to 128 bits. Sure not trivial to memorise, but far from impossible.
      - manmal 4y ago
        You're not wrong of course - this would come with heavy UX issues though. E.g. it's hard to type 10 words into a password field without messing up. And _really_ hard on a mobile device. 1P would allow you to have a password like "1234" and still have the same key entropy as that 10 word password alone.
        fluidcruft 4y ago

by manmal 4y ago

notesinthefield 4y ago

Keepass has Key Files as a part of the spec https://keepass.info/help/base/keys.html https://keepass.info/help/base/keys.html On my devices, keyfiles and a KP client are stored locally. The DB rests in the cloud.

manmal 4y ago
Is that kind of file storage secure? I thought files stored on disk can be extracted quite easily?

brandon272 4y ago

LastPass would have also led their customers to believe that "brute forcing was not possible" and that they were taking extraordinary measures to keep vaults and data safe. I think one distinction between services like KeePass and 1Password is end user perception of how easy it is for an attacker to acquire an encrypted vault to begin with. For many, they consider a KDBX database sitting in their Dropbox account to be less likely to be stolen than an encrypted vault being held by a company like 1Password, a high value target to the most sophisticated attackers including state actors.

hn_throwaway_99 4y ago
Doesn't necessarily matter what LastPass "would have also led their customers to believe", the mathematical reality is still that LassPass vaults are crackable in a way that 1P vaults fundamentally are not.
- brandon272 4y ago
  Yes, according to what 1Password is telling us. But as we've seen, what these companies say and what they actually do in practice are not always aligned. And oftentimes customers are inserting a lot of their own assumptions into the mix, not only with respect to vault encryption but vault storage and operational security.
  - manmal 4y ago
    > Yes, according to what 1Password is telling us With their very comprehensive whitepaper and Charles Proxy you can verify all their claims. Their whitepaper is one of the best resources I have found on E2EE in general. With that, you should be able to write your own 1P vault parser. Then you can verify that traffic to their server is exactly what they claim it to be. In another comment you are criticizing that their product is proprietary - that's IMO not quite true. Yes, 1P is closed source, but their crypto strategy is documented extensively - they list the exact cipher algos and settings. > not only with respect to vault encryption but vault storage and operational security That's a valid argument, BUT, if you read their whitepaper, you'll likely arrive at the conclusion that even a full leak of the encrypted vault is currently not that problematic. I wouldn't post it online, but I'm not worried if they announce a leak tomorrow.
    - brandon272 4y ago
      All valid points!
  - hn_throwaway_99 4y ago
    > Yes, according to what 1Password is telling us. But as we've seen, what these companies say and what they actually do in practice are not always aligned. That's just not accurate: 1. First off, all the encryption happens client-side. It is possible for anyone so inclined to validate how 1P and LP are doing their encryption. 2. The deficiencies in LP's encryption approach were well known for years. My point it, yes, companies will spin things how ever they want, which is why you should completely ignore what they say and only evaluate what is verifiable. And 1P's and LP's approaches are verifiably different.

princevegeta89 4y ago

Same was said about LastPass many times and look at what happened, everything turns out to be a false promise.

hn_throwaway_99 4y ago
That's not a fair comparison. The differences in LP and 1P encryption approaches have been well known for years, and they are fundamentally different. Now, while 1P encrypted vaults are not brute-forceable the way LP's are, that doesn't mean it's impossible to hack 1P (e.g. malicious code injection in any of their apps or plugins), but I don't like the "everything turns out to be a false promise" broad-brushing when there are real and verifiable differences in how these companies secure your data.
- fluidcruft 4y ago
  Unless you want to spend the time doing RE of crypto (so fun!), how to do you even know who works for 1Password? Are you waiting for Snowden 2.0 to explain how things work a decade later?

DavideNL 4y ago

Note that 1Password copies the "Secret Key" to iCloud... without asking.

manmal 4y ago
Do you mean iCloud keychain? That's arguably one of the most secure cloud storage in existence.
- DavideNL 4y ago
  Yes. And no, because it depends on how users setup and use their AppleID and its passwords/security/devices. My only point is, if 1Passwords decides to share the private key with [whichever cloud service] the user should at least be notified, or get a choice.
  - manmal 4y ago
    > And no, because it depends on how users setup and use their AppleID and its passwords/security/devices. Can you elaborate what the issue would be? I see that the AppleID password could be a weak link, but that's mostly mitigated by 2FA. > if 1Passwords decides to share the private key I'm not aware that they could do that. Their servers have no knowledge of either the password or secret key. Authentication happens via a zero-knowledge proof.
    - manmal 4y ago
      Ah, Apple recently improved things further: https://www.apple.com/newsroom/2022/12/apple-advances-user-security-with-powerful-new-data-protections/ https://www.apple.com/newsroom/2022/12/apple-advances-user-s...

phonebucket 4y ago

But in the context of a strong master password, the additional benefit of the secret key is of neglible benefit, while the hassle and dangers of having to synchronise the secret key remain. I'd rather use an extremely high entropy master password by itself.

manmal 4y ago
How do you remember a 128 bit entropy password?
- phonebucket 4y ago
  remember 10 random words

fluidcruft 4y ago

How is 1Password's secret key more secure than using keepass's key file feature?

manmal 4y ago
That depends where you store that file. Is there a way to store this securely in a mobile device for use on the go?
DavideNL 4y ago
Or, simply a password for your Keepass database with the same entropy as 1Passwords secret key. Anyways, it depends on the app you use. Some Keepass apps haven't even been audited for example, so security of those apps may be questionable.
- manmal 4y ago
  > Or, simply a password for your Keepass database with the same entropy as 1Passwords secret key. How would you remember a password of 128bit of entropy?
  - dagw 4y ago
    9-10 words chosen at random from a list of 10k common words in you native language will get you to 128 bits. Sure not trivial to memorise, but far from impossible.
    - manmal 4y ago
      You're not wrong of course - this would come with heavy UX issues though. E.g. it's hard to type 10 words into a password field without messing up. And _really_ hard on a mobile device. 1P would allow you to have a password like "1234" and still have the same key entropy as that 10 word password alone.
      - fluidcruft 4y ago