4 ms·

If the service is well implemented this should be very hard. Basically when you login you receive a session/ticket (usually by cookie), using that session you p

by maushu 16y ago

If the service is well implemented this should be very hard. Basically when you login you receive a session/ticket (usually by cookie), using that session you post as yourself. To post as somebody else you would need to get their session which is difficult since it expires after some time and it should very hard to guess (usually an hash with 32 characters or more). What firesheep does is steal those sessions during transaction. No guessing there.

EGreg 16y ago
fine, but if I am making a third-party authentication service, how do I prevent the users of that service from being bamboozled by simple javascript? How does facebook prevent a website from being bamboozled? Is there any technique that facebook can recommend (such as signing the uid and sending the signature along with the id on all server requests) so that the website isn't fooled?
- Swannie 16y ago
  I'm not sure I understand your question. If you mean: I have an authentication service, Auth.com, and I would like the users to be able to enter their user name and password on AnotherSite.com, how do I stop javascript on AnotherSite.com hijacking their submission to Auth.com? Then the short answer is, you can't. That is why all services redirect you to their own login page, where they control the HTML (and can audit to be sure that no one has injected code into their site). What Facebook does is usually in an iFrame, which accesses the user information relatively securely. Any interaction more complex than a "like" is generally redirected via facebook.com (e.g. any permission grants). iFrame's used to be riddled with flaws, but I believe they are considered fairly safe these days. Go educate yourself on the protection of XSS and CSRF/XSRF, and you'll be able to answer all of your own questions.
  - EGreg 16y ago
    No, that's not what I am asking. And I am very familiar with how to combat XSS and CSRF, having written a PHP framework that does it for you out of the box, among other things (http://phponpie.com http://phponpie.com) I am asking another question. If I am operating an OpenID provider, say Auth.com, and CoolStuff.com uses me to authenticate a user, then they can use that session. Great. But now let's say I operate an OAuth provider, and it releases this user's uid, first_name and last_name. What measures does facebook and other OAuth providers take to prevent the uid, first_name, last_name and other data from just being changed by the user in javascript, before being posted to the CoolStuff.com servers, after they have been obtained using Javascript ... such as FB.api('fql.query', ...)
    - kertap 16y ago
      If I understand the question correctly, you are asking what's to stop someone from changing the UID, etc when the user is redirected from auth.com to coolstuff.com? From my understanding of OAuth, when the uid, first_name and last_name are sent by auth.com it also sends a cryptographic hash of everything. So if you change the uid, you would also have to change the hash and you can't change the hash without knowing the shared secret that the auth.com and coolstuff.com have decided on prior to your request.
      - EGreg 16y ago
        so what you are saying is that coolstuff.com's servers must always look for a signature from auth.com when they are accepting user data, they cannot trust the user input. That makes sense. So does facebook connect, google, etc. have this? I assume not.
        Swannie 16y ago
        Well if you protect against CSRF you should be checking a nonce for all form submissions anyway. This doesn't appear to be all that different to me. I'm intrigued now, I hope someone can give details!
        sawyer 16y ago
        OAuth 2.0 login providers use an access token that is user specific for all subsequent interaction. For example, with Facebook, once a user has logged in and given Application X permission to access their details, FB will send the user's id, name, etc. (whatever data the user has granted access to) along with a unique access token. The next time, and every subsequent time Application X wants to access the FB API on the user's behalf it is required to send that access token. From javascript you might be able to change the token, however, Application X's next interaction with the FB API will fail if the token is invalid and there is no way to derive a token value from a FB user's id.
        16y ago

by maushu 16y ago

EGreg 16y ago

fine, but if I am making a third-party authentication service, how do I prevent the users of that service from being bamboozled by simple javascript? How does facebook prevent a website from being bamboozled? Is there any technique that facebook can recommend (such as signing the uid and sending the signature along with the id on all server requests) so that the website isn't fooled?

Swannie 16y ago
I'm not sure I understand your question. If you mean: I have an authentication service, Auth.com, and I would like the users to be able to enter their user name and password on AnotherSite.com, how do I stop javascript on AnotherSite.com hijacking their submission to Auth.com? Then the short answer is, you can't. That is why all services redirect you to their own login page, where they control the HTML (and can audit to be sure that no one has injected code into their site). What Facebook does is usually in an iFrame, which accesses the user information relatively securely. Any interaction more complex than a "like" is generally redirected via facebook.com (e.g. any permission grants). iFrame's used to be riddled with flaws, but I believe they are considered fairly safe these days. Go educate yourself on the protection of XSS and CSRF/XSRF, and you'll be able to answer all of your own questions.
- EGreg 16y ago
  No, that's not what I am asking. And I am very familiar with how to combat XSS and CSRF, having written a PHP framework that does it for you out of the box, among other things (http://phponpie.com http://phponpie.com) I am asking another question. If I am operating an OpenID provider, say Auth.com, and CoolStuff.com uses me to authenticate a user, then they can use that session. Great. But now let's say I operate an OAuth provider, and it releases this user's uid, first_name and last_name. What measures does facebook and other OAuth providers take to prevent the uid, first_name, last_name and other data from just being changed by the user in javascript, before being posted to the CoolStuff.com servers, after they have been obtained using Javascript ... such as FB.api('fql.query', ...)
  - kertap 16y ago
    If I understand the question correctly, you are asking what's to stop someone from changing the UID, etc when the user is redirected from auth.com to coolstuff.com? From my understanding of OAuth, when the uid, first_name and last_name are sent by auth.com it also sends a cryptographic hash of everything. So if you change the uid, you would also have to change the hash and you can't change the hash without knowing the shared secret that the auth.com and coolstuff.com have decided on prior to your request.
    - EGreg 16y ago
      so what you are saying is that coolstuff.com's servers must always look for a signature from auth.com when they are accepting user data, they cannot trust the user input. That makes sense. So does facebook connect, google, etc. have this? I assume not.
      - Swannie 16y ago
        Well if you protect against CSRF you should be checking a nonce for all form submissions anyway. This doesn't appear to be all that different to me. I'm intrigued now, I hope someone can give details!
        sawyer 16y ago
        OAuth 2.0 login providers use an access token that is user specific for all subsequent interaction. For example, with Facebook, once a user has logged in and given Application X permission to access their details, FB will send the user's id, name, etc. (whatever data the user has granted access to) along with a unique access token. The next time, and every subsequent time Application X wants to access the FB API on the user's behalf it is required to send that access token. From javascript you might be able to change the token, however, Application X's next interaction with the FB API will fail if the token is invalid and there is no way to derive a token value from a FB user's id.
        16y ago