News for hackers. Stuff that matters.
Re:1Password Watchtower Mac App Integration by ColinDabritz on Thursday May 01, 2014 @11:17AM (#7680267)
Do you know what would be amazing?

The ability to change all vulnerable passwords, automatically as a group.

I know it's very challenging because password change/reset procedures vary widely, but that part of the process is still manual. 1Password already automates logins almost entirely successfully. Perhaps password changing can be automated as well?

If it were robustly automated, you cold easily rotate your passwords once a month, with very little fuss, for example.

This is really something that authentication schemes could benefit from supporting an automated API for.

Good work on the service. I suspect that heartbleed was the current use case, but this service will be very helpful for future breeches and issues. For example they could alert you that XYZ Service had a data breech, it's fixed, and you should rotate your password.

Great work.

Re:1Password Watchtower Mac App Integration by AGKyle on Thursday May 01, 2014 @12:32PM (#7680829)
[I work for AgileBits, makers of 1Password]

Thanks for the feedback!

Filling into sites and gathering the necessary information for this I think is far easier than determining how a password change takes place. We have an algorithm that we've developed over years that is constantly changing in subtle (and not so subtle) ways to make filling more accurate. Some of the biggest changes in a long time will be coming in the 4.2 version of the browser extension, currently in beta.

But you're right, sites and their password change processes differ greatly. Some require the old password, some require only the new password. Some only require the new password once (admittedly rare).

I think at best though we'd only be able to do this if we went to the password change page for the site and attempted to fill the data in for changing the password. There's no way we could bulk update as that would probably require we hard code each individual site rather than relying on an algorithm for filling data into the site.

It's a tricky thing, but we're always trying to come up with new ideas and ways to accomplish those ideas.

Heartbleed was the first use case for Watchtower, but you're correct in that it'll be available for other situations as well. Now that the foundation is there we can leverage it for other issues in the future.

Thanks again for the wonderful feedback!

Kyle

AgileBits

Re:1Password Watchtower Mac App Integration by masklinn on Thursday May 01, 2014 @01:12PM (#7681124)
> Filling into sites and gathering the necessary information for this I think is far easier than determining how a password change takes place.

Maybe start talking it out with browser developers and big websites stuff so there's a way to standardise an endpoint or in-page meta-information allowing for automated password reset?

Re:1Password Watchtower Mac App Integration by ColinDabritz on Thursday May 01, 2014 @11:12PM (#7683952)
This could be a real win. If there was a way of marking a standardized form with some meta data or tags to say "hey, this is safe to use automatically in the expected way" such as the typical form with old > new > repeat new or something. One little HTML change and it advertises compatibility.

That way, classy sites could enable a 'safe enough/good enough' flag that any software could use.

There is also the thought that it may be worth hard-coding for the biggest sites. The high profile sites are the most vulnerable when a vulnerability hits, there are big existing databases of user names around, and these can be exploited automatically and quickly on relatively high value targets (gmail accounts, amazon, facebook, twitter etc). Even top 10 would be helpful, but top 100 seems approachable, especially since the bigger sites are probably more consistent. The payoff here could be big next time around.

Also if the framework were around for doing this, you could target a particular site or three that just had a major breech, and push the rules for reset out with the breech notification.

I know it's a huge amount of work to do manually, but even a relatively focused effort could have payoffs, and if working with big sites can start a 'automated password reset' standard rolling, others might adopt it.

Just thinking out loud, I know you've got to juggle priorities and features with limited time.

Thanks again!

Re:1Password Watchtower Mac App Integration by threatofrain on Thursday May 01, 2014 @04:57PM (#7682645)
Are you guys ever thinking of a Linux version :(
Re:1Password Watchtower Mac App Integration by AGKyle on Thursday May 01, 2014 @06:24PM (#7683033)
Thinking about it? Sure. We've had a lot of requests for it.

It's a pretty big job to write a whole new version of the application. It's also a big question of whether we'd sell enough copies to maintain it. Traditionally Linux users haven't been big on purchasing software, particularly closed source software. I'm not sure how much this has changed, if anyone has some real solid stats on it I would love to see it. I was once a Linux user (started on Redhat 4, but fell in love with Debian) so I know what the past was like, but have been out of the Linux world for long enough to be unable to know how things have changed in this regard.

Either way though, I don't have any pull on this, but I can pass the request along. I think we (AgileBits) would all love to see 1Password on more platforms. It's just a bit of a careful process to go through. We hear the requests for Linux loud and clear. If it's something you want, please let us know. Include any thoughts for how you'd use it and what you envision it being. It's far easier to take to the powers that be ideas and thoughts than just requests.

Now though, the fun part is that the current Windows beta (version 4) will work in WINE and I do hear from users that the browser extensions will also function using the beta. Obviously this isn't an ideal solution but it might hold you over well enough.

If you'd like to try the beta, you're welcome to sign up here:

https://agilebits.com/beta_signups/winnewsletter.html

Let me know if that helps!

Kyle

AgileBits

Re:1Password Watchtower Mac App Integration by threatofrain on Saturday May 03, 2014 @01:03PM (#7691464)
I'm afraid I don't actually know any numbers on Linux users' willingness to buy software, but I am aware of the reputation that Linux users generally prefer their ecosystem of free apps.

However, I bet that many Linux users are also Windows or OSX users, and if a password manager is to be effective, it has to be omnipresent. That's why I have begrudgingly paid for LastPass, while my dual Windows / OSX license for 1Password 4 has sat unused. I tried for awhile, but I couldn't stand the gap in coverage.

So I narrowed it down to RoboForm and LastPass, and I chose the latter because it seemed more popular and I hoped that to predict company longevity. Because as nice as lifetime licenses are, once software is no longer supported, esp. for things like intelligent form-filling, chances are you won't want to use your software anymore.

Re:1Password Watchtower Mac App Integration by AGKyle on Monday May 05, 2014 @09:58AM (#7698569)
I can totally respect that decision. I'll pass the request along to the developers. I can't promise anything but hopefully it's something we can consider in the future. It's certainly not something that will pop up next week though :) For a native Linux application we would pretty much have to rewrite the entirety of the application specifically for Linux, save the extensions.

Thanks for the great feedback though! Hopefully you'll get what you want in the future.

Kyle

AgileBits

Re:1Password Watchtower Mac App Integration by kolev on Thursday May 01, 2014 @07:09PM (#7683214)
Well, maybe you should try a Kickstarter campaign and see how it goes... I'm sure you have to invest a bit to create a video and all, but I'm sure it will be a successful campaign.
Re:1Password Watchtower Mac App Integration by senorprogrammer on Thursday May 01, 2014 @11:32AM (#7680373)
I wish that affected sites would simply reset my password themselves, internally, and then inform me when I visit the site that I need to use the 'Forgot Password' function to reset my password.

As it stands, they get to say they're patched while I remain vulnerable, and now it's my fault if my account is compromised if I don't get around to resetting my password (or they're a site I visited once, years ago, and forgot all about).

Simply updating OpenSSL without fixing passwords just shifts the onus to the users.

Re:1Password Watchtower Mac App Integration by bdcravens on Thursday May 01, 2014 @02:18PM (#7681671)
That's a tough call, especially for small companies. There are very real support expenses (not every app or it's users have SaaS-like support requirements) that don't make this an easy call.
Re:1Password Watchtower Mac App Integration by jfb on Thursday May 01, 2014 @11:22AM (#7680305)
As a half-way step, they could aggregate and normalize the various stupid rules that sites impose and then automatically generate valid, maximally secure passwords by default. Eventually, 1Password (or similar) could handle e.g. changing them all on a given schedule. I already try to have no passwords older than three months, and it's a PITA.
Re:1Password Watchtower Mac App Integration by r00fus on Thursday May 01, 2014 @11:50AM (#7680497)
Problem: Aggregation and normalization might result in an empty set as contradictory rules prevent harmonization.
Re:1Password Watchtower Mac App Integration by jarrett on Thursday May 01, 2014 @12:00PM (#7680556)
I think the parent comment is saying that 1Password would record the rules per site, not maintain a global list of rules that satisfies all sites. The latter would indeed yield an empty set, as some sites require a special character and some require no special characters.
Re:1Password Watchtower Mac App Integration by jfb on Thursday May 01, 2014 @02:40PM (#7681832)
Yes, exactly; for each site that 1Pwd can recognize as distinct, record the restrictions. Even share these restrictions. And perhaps people could put attributes on their password entry elements (or header metadata) that would describe the limitations of their password systems, so it wouldn't be specific to any one password manager; then, too, they could retire their terrible, terrible, terrible client-side strength checking crap.