Slashdot Mirror


Facebook Fixes Bug That Allowed Users To Set Other Users' Passwords

An anonymous reader writes: Facebook has paid $15,000 (€13,600) to an independent security researcher who discovered a simple way of resetting passwords for other people's Facebook accounts, setting a new passphrase and effectively taking over profiles.

The problem was in the fact that Facebook also runs a Beta platform on beta.facebook.com. This platform's "reset password" feature did not include brute-force protection and allowed anyone to guess the six-digit verification code sent to someone's phone when resetting the password. This issue also raises another question: How many unsafe features are on Facebook's beta platform that have not been patched simultaneously with the main platform?

11 of 49 comments (clear)

  1. It's not a bug...it's a feature by evolutionary · · Score: 2

    It's not like Facebook was really private anyway...People can mark/identify others without the account owner's consent. So this is no surprise to me. Security/privacy is not exactly a priority at facebook. (the opposite actually..)

    --
    "Imagination is more important than knowledge" - Einstein
    1. Re:It's not a bug...it's a feature by Anonymous Coward · · Score: 2, Informative

      People can mark/identify others without the account owner's consent.

      Any time I'm tagged anywhere, I get notified and can force remove it if I choose.
      But it's not like you or anyone else can prevent someone from simply adding text to a picture with their name on it. But let's blame facebook for that too, because reasons.

      Security/privacy is not exactly a priority at facebook

      In relation to this article, this bug only affects people who give FB their phone number and set it to their 'account recovery' preference. I've never done either, mine works via my email, and the "code" they send is pretty damn long and includes letters/number/etc.

  2. Re:beta? by OzPeter · · Score: 4, Funny

    You'd normally expect more features in beta, even if not stable. Weird to see less protection on the beta platform

    You never saw /. beta did you?

    --
    I am Slashdot. Are you Slashdot as well?
  3. Better question: why running prod data in beta? by xxxJonBoyxxx · · Score: 4, Interesting

    I could see having a per-account switch to "allow me to use my account in beta" (default = OFF) for developers who want to play with this stuff, but why would you want to expose your production customers to untested software like this?

    >> Weird to see less protection on the beta platform

    Not if you've ever seen teams refactor code in a large codebase. When that occurs, you often lose a lot of the "history" and "memory" of a branch, which often resurfaces bugs, edge cases take care of years ago and new vulnerabilities.

    1. Re:Better question: why running prod data in beta? by KGIII · · Score: 2

      > ... your production customers ...

      There you go again. You seem confused as to who the customer is and, by extension, who gets the prioritized attention and care. (Hint: It's not the people who have 'user' accounts.)

      --
      "So long and thanks for all the fish."
    2. Re:Better question: why running prod data in beta? by Jason+Levine · · Score: 2

      Testing? We have no time for testing! We have to ship the product out now because marketing told everyone that the release date was today. It seems to work well enough when we ran it that one time so it must be fine. Besides, you need to work on these five dozen other projects since we fired half the staff and kept the workload the same.

      --
      My sci-fi novel, Ghost Thief, is now available from Amazon.com.
  4. Re:questions abound by Anonymous Coward · · Score: 4, Insightful

    Facebook is not about customer service. It's not Mickey Ds.

    FB is largely a platform for people, namely Americans, to bolster their ego without doing any real work. You just post some tit pictures and let the "likes" roll in, and if puffs your ego up. FB is one of the worst things to ever happen to the American psyche. Everyone thinks they are a bad ass with talent, whereas reality is closer to "whiny bitch that no one cares about"

  5. Re:Thank god we're all safe now by gurps_npc · · Score: 2
    Almost 90% of people do not use that passcode.

    What I don't understand is why so few people use 8068. It's a perfectly good passcode, but it's the least chosen one.

    --
    excitingthingstodo.blogspot.com
  6. only $15k? by Anonymous Coward · · Score: 4, Insightful

    fucking cheapskates.

    ___

    wtf is with capcha treating me like a nigerian prince trying to send webmail? captcha: zmnjwfm

  7. Problem is password reset itself by WaffleMonster · · Score: 3, Insightful

    Schemes for resetting passwords fundamentally lower the security of the system and almost always rely on insecure transports (Email and SMS).

    At the very least users should be given the option of not allowing any password reset or recovery features to be used in conjunction with their account.

    Rather than conceding to inevitability of forgotten passwords I would rather see sites warn users ahead of time what the consequences are including suggestion to write it down and store it in a safe place.

    --
    From origional descent devs
    http://media.revivalprod.com/O...

    1. Re:Problem is password reset itself by castionsosa · · Score: 2

      Password recovery is in itself, an art form.

      One thing I've wondered about is the concept of password recovery providers. Not a central website, since it can get compromised, but different organizations, similar to how OpenID is set up.

      When setting an account with some provider, one chooses a recovery provider or providers, and what methods will be used to get back the account. This way, if someone has their own dedicated VM or device that makes an OATH number, that can be used. Another provider sends an encrypted SMS message, and has an app that decrypts it for the user. Still another provider sends out a physical card via registered mail with a bunch of scratch-out blanks. Another provider has a database of recovery questions (similar to how PGP Server used to have a way to recover keys), and someone uses x out of y questions (where are the bodies buried) to get a valid recovery code.

      This would provide a lot of flexibility, but still have solid security. For example, someone might have a basic E-mail account, and for them, just a "click here to get a recovery code" message is good enough. Other people might want a physical device, similar to a SecurID "calculator" that is offline and airgapped, requires a PIN to get a recovery code. Still other people just want to have a scratch-off card with passwords in one lock-box, and SMS messages on their phone for general use.

      I've also wondered about a device similar to the SecurID fob, except with no battery (plug it into a USB port to power it, but it doesn't use the plug for data), it would keep sync with time via a 3G connection, but would function as an offline device (punch button, get recovery code.) This could be tossed in the desk drawer, safe, or other spots, for something to recover an account, should all else fail. If it used a standard TKIP protocol, it would be simple, and decently secure.