Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

KeeWeb: Unofficial KeePass web and desktop client

github.com

Post reply on HN
Print view View original

101–110 of 125 posts

Re: KeeWeb: Unofficial KeePass web and desktop client

#101

Post by blacksmith_tb »

dchest wrote: post #41
dangerlibrary wrote: post #38

Earlier quoted context omitted.

Just so we're clear: your assertion is that because the password generator uses a non-cryptographically secure PRNG, generating unique 20+ character alphanumeric passwords for every login is worse than people's default behavior? I mean, I get that it's worse than other keepass implementations - that's obviously a problem - but if this gets people to stop using "hunter2" or "p4$$w0rd" that's got to be worth something,…

No, it's not worse for your scenario. (That wasn't my assertion) However, this is an alternative to KeePass/KeePassX, so the typical behavior of KeePass users is to generate passwords with it, not reuse bad passwords. For example, I use KeePassX to generate strong passwords for long-term encrypted archives, and if I switched to this app, I wouldn't get the same security. I'm confused, though, should I say: "All right…

Hmm, I tend to think of KeePass as a manager for passwords for accounts, services, etc. I hadn't considered storing archive passwords in it, or are you only using it to generate them? It seems like pwgen would do that fairly well...

Re: KeeWeb: Unofficial KeePass web and desktop client

#102

Post by oneeyedpigeon »

DenisM wrote: post #95
oneeyedpigeon wrote: post #68

Earlier quoted context omitted.

The comment I replied to which suggested that strong-encryption was a differential between keepass and 1password.

As far as I can see, the comment you replied to contains no mention of these things. Can you quote the relevant part?

cmrx64: it all depends on how much you trust 1password and what your threat model is. For me, the advantage of keepass is that I don't need to upload my credentials anywhere

oneeyedpigeon: But if you want to sync your credentials across devices, you still have to upload them somewhere, right?

dorfsmay: You are uploading a file that is encrypted using very strong encryption, not plain text password

I took that to mean:

(with keepass) you are uploading a file that is encrypted ... not plain text password (as for 1password)

dorfsmay has now confirmed that was their meaning in this comment: https://news.ycombinator.com/item?id=11177045

Re: KeeWeb: Unofficial KeePass web and desktop client

#103

Post by relaxitup »

paulryanrogers wrote: post #91
relaxitup wrote: post #88

Does there exist an enterprise grade server/webui solution based on a keepass db? We are looking for an enterprise password manager solution that does not need all the ldap/ad integration bells/whistles (although we may explore ldap integration with the tool in the future). So I was thinking why not just use keepass. And by enterprise grade I guess I really mean it needs to be a multi user solution, but everyone woul…

PleasantSolutions.com may help. (Full disclosure: I sell an unrelated KeePass plug-in.)

This looks very interesting thanks!

Re: KeeWeb: Unofficial KeePass web and desktop client

#104

Post by marcosdumay »

blakesterz wrote: post #42
judofyr wrote: post #36

Earlier quoted context omitted.

Is there a practical attack for this usage? The attacker would need to have a bunch of your passwords already? EDIT: Not that I'm justifying using Math.random, I just don't see why you so strongly recommend against using this tool.

>> Is there a practical attack for this usage? This is the question I ask every single time I read about anything in security. It feels like there's just SO MANY THINGS wrong about every damn thing now the only way I can figure out what I have to really worry about or focus on is anything with a PRACTICAL attack that's easily automated and remotely exploitable.

It's (relatively) easy to verify a software against a strict security definition and decide if it meets the definition or not.

It's incredibly difficult (I'd say "impossible", but I'm being pedantic for now) to check a flaw against such strict definition and check it against all possible use cases and all possible inputs and decide if one of them will lead to bad consequences.

Thus, if there's any practical attack? How the hell would I or anybody else know. We can only know a resounding "yes" after you get owned, nobody can ever tell you a honest "no" to this question.

Re: KeeWeb: Unofficial KeePass web and desktop client

#105

Post by dorfsmay »

apawloski wrote: post #98
dorfsmay wrote: post #97

Earlier quoted context omitted.

No, but my understanding is that with 1passsword and similar service the web client sends the password in unencrypted form to the server. A rogue employee, is even the combination of a bug and a leak would expose your password. With keepassx, your password never leaves your device in unencrypted form.

This is very much untrue. 1Password syncs an encrypted vault through separate channel (e.g. Dropbox, iCloud) -- it has zero-knowledge of your passwords. It just picks up a big encrypted blob from wherever you store it. "The easiest way for us to protect your data and data about you is to not have that data in the first place. You may be noticing a theme by now: we can’t reveal or abuse data that we don’t have. We do…

Thanks for the clarification.

Re: KeeWeb: Unofficial KeePass web and desktop client

#106

Post by Nadya »

Touche wrote: post #93
dchest wrote: post #87

Earlier quoted context omitted.

Looks like you didn't understand the sadness of the issue I was reporting. The author of the software constructed his own random number generator from two primitives: Salsa20 stream generator and Math.random. The last part is what makes this PRNG theoretically insecure: because Math.random is not guaranteed to give cryptographically secure random numbers, thus, theoretically, the generated passwords can be guessed. (…

This is what I'm talking about. You, a security enthusiast, are only interested in the technical security details. You respond to a comment about tradeoffs with details on the bug. I don't care about why the bug happened or how easy it is/isn't to fix. I care about whether the existence of the bug is something I should be so concerned about as to not use the software. In order to gauge that, I need a little more info…

>I don't care about why the bug happened or how easy it is/isn't to fix. I care about whether the existence of the bug is something I should be so concerned about as to not use the software. In order to gauge that, I need a little more info about the threat level.

Threat level is 0%. No currently known attacks exist. This threat level immediately goes to 100% when a practical attack is discovered. There is no guarantee that a practical attack exists that hasn't been brought to academic or mainstream attention (e.g some cracker has a practical attack that they're keeping under wraps). By the time the threat level hits 100% the cracker may have already broken into your account(s) before you even hear about the attack.

Therefore when something is shown that "attacking it is possible" you can make one of two assumptions

1) No practical attack exists and you'll be safe until it exists

2) A practical attack already exists and it is only a matter of time until you get pwned

Rather than worry about whether or not a practical attack already or will one day exist, I'd use cryptography that hasn't been shown to be broken.

Re: KeeWeb: Unofficial KeePass web and desktop client

#107

Post by nikolay »

Sir_Cmpwn wrote: post #37
datamoshr wrote: post #34

Earlier quoted context omitted.

I've previously looked into this and stopped half way through migrating my lastpass. I have a rather naive question: For the replication across devices this tool relies on git, but where are folks setting up the origin?

I host mine on gogs.sr.ht, a private git service I run for myself and friends. I used to host it on a private GitHub repository. There are other options like a private Bitbucket repository, or even just a public repo on GitHub - since the passwords are all encrypted, you don't really have to worry about that. I'll happily give an sr.ht account to anyone who wants one for this purpose, mention HN in your application c…

Can you approve @nikolay, please?

Re: KeeWeb: Unofficial KeePass web and desktop client

#108

Post by Sephr »

The reference implementation's domain is vulnerable to MITM attacks between KeeWeb and CloudFlare until they set their CloudFlare crypto settings to "strict" and get some real certificates of their own (e.g. Let's Encrypt).

I submitted an issue here: https://github.com/antelle/keeweb/issues/111

Of course exploiting this would be very difficult, but it is possible to MITM the connection between the CloudFlare proxy and GitHub pages as long as keeweb.info continues to not use DNSSEC.

Re: KeeWeb: Unofficial KeePass web and desktop client

#109

Post by Nullabillity »

Quiark wrote: post #66
m3Lith wrote: post #63

Earlier quoted context omitted.

There's addons for Firefox/Chrome as well as apps for Android that enable autofill and other stuff you might expect from a modern password manager (making the UX pretty close to LastPass).

Some warnings: don't use Android clipboard for passwords - AFAIK apps can get notifications when clipboard changes. Don't enable remote connections to the KeepassHttp browser integration: there's a gaping hole https://github.com/pfn/keepasshttp/issues/258

KeepShare doesn't use your clipboard.

Re: KeeWeb: Unofficial KeePass web and desktop client

#110

Post by Spooky23 »

judofyr wrote: post #36
dchest wrote: post #29

Do not generate passwords with it, it uses insecure Math.random: https://github.com/antelle/kdbxweb/blob/906e927d3e3384db4dd3... https://github.com/antelle/keeweb/blob/master/app/scripts/ut... (in meme form: https://imgur.com/FcZNflQ ) Filed issue: https://github.com/antelle/kdbxweb/issues/5 (embarrassing/funny: it was me who wrote Salsa20 "user-space" generator used here ( https://github.com/antelle/kdbxweb/blob/906…

Is there a practical attack for this usage? The attacker would need to have a bunch of your passwords already? EDIT: Not that I'm justifying using Math.random, I just don't see why you so strongly recommend against using this tool.

Its all about risk/impact.

IMO, in most cases it is a non-issue. But... If you use a badly generated password as a key for a encrypted volume, then you might have a problem.

Post reply on HN