13

After browsing another question on this site, I discovered haveibeenpwned.com and thought I'd check my email. It came up with one breach which has, apparently, not been widely shared and does not include a password or much in the way of personal information:

enter image description here

I have a 12+ characters random alphanumeric password for this site which is unlikely to be broken. I have not used the site for personal messaging, and have received but not made payments through it. So this doesn't sound like much to be concerned about, aside from a possible increase in spam.

I appreciate that there are several questions which ask what action should be taken if your details are leaked online. However, I don't want to go to the effort if remedial action is unnecessary, so I'm asking to ascertain whether or not I need to bother?

WhiteWinterWolf
  • 19,142
  • 4
  • 59
  • 107
Bob Tway
  • 549
  • 1
  • 4
  • 12
  • Possible duplicate of [What can I do if I discover that my password hash has been leaked in pastebin?](http://security.stackexchange.com/questions/108191/what-can-i-do-if-i-discover-that-my-password-hash-has-been-leaked-in-pastebin) – Stephane Dec 16 '15 at 12:52
  • 11
    @Stephane: I'm not convinced this is a duplicate, since the OP assumes here that neither the password not its hash has been leaked based on haveibeenpwned.com quoted message. – WhiteWinterWolf Dec 16 '15 at 14:03
  • I think the answer to this comes back to the point that security is a spectrum and not a binary flag. Is it possible that more was breached than was revealed (including your password)? Yes. Is it likely? Probably not, but that is a judgement call for you to make – David says Reinstate Monica Dec 16 '15 at 18:22

2 Answers2

7

The Patreon data included passwords hashed using the industry best practice BCrypt algorithm (this data is available online through a search for "patreon data dump", although possession of it may be a criminal offence in some jurisdictions, so I'm not linking to it). This does not mean that they are impossible to break, but they are relatively slow to break. Therefore, you should still consider the password as potentially compromised.

Specifically, the table data available through sources such as this reveals that a table called "tblUsers" contained a field called "Password", and that another table called "tblUnverifiedUsers" did too. Given the scale of the dump, it would be irresponsible to assume that this data was not accessed. The CEO of Patreon also provided a later statement to Motherboard stating "We encrypt all tax form information with a 2048-bit RSA key. The key used to decrypt this information lives on a separate server and was not compromised. All user passwords are hashed using bcrypt with 8 or 12 passes, depending on when the user signed up."

The idea of hashing passwords is to provide that buffer for changing details in the event of a breach - any password can be broken, given sufficient time, no matter what algorithm is used. After all, there has to be a way to compare the input from the user to the data stored in the database - the worst case scenario for the attacker is that they try every possible input string through the same algorithm, with the same salt, until they get the same output. The best case for them is that they don't need to bother doing that.

Assuming you are practising good password usage, you are probably fine - you've presumably already changed your Patreon login, and it'll be different to every other site you have. There isn't anything you can do about any of the other data though.

Matthew
  • 27,263
  • 7
  • 89
  • 101
  • Can you provide a link to back up that "The Patreon data included passwords hash(s)"? Because that statement contradicts with the image posted in the question. – Mike Ounsworth Dec 16 '15 at 14:01
  • The official statement from Patreon: https://www.patreon.com/posts/important-notice-3457485 – Matthew Dec 16 '15 at 14:02
  • From that press release: "There was unauthorized access to registered names, email addresses, posts, and some shipping addresses. Additionally, some billing addresses that were added prior to 2014 were also accessed...Although accessed, all passwords, social security numbers and tax form information remain safely encrypted with a 2048-bit RSA key.". They mention bcrypt on that page as a precaution, but do not say that any hashed passwords were stolen. Unless I'm missing something, the premise of your answer is incorrect. – Mike Ounsworth Dec 16 '15 at 14:07
  • In the case of system breaches, "accessed" is equivalent to "stolen" - if the data has been accessed, a copy has potentially been taken, and should be assumed to be in the hands of attackers. There is no way to distinguish between "they looked at it and didn't take a copy" and "they took the whole lot". – Matthew Dec 16 '15 at 14:15
  • ... but there is a difference between "they took the data" and "they took an encrypted version of the data that they don't have the key for". – Mike Ounsworth Dec 16 '15 at 14:17
  • Not encrypted. The passwords were hashed: "We protect our users’ passwords with a hashing scheme called ‘bcrypt’ and randomly salt each individual password. Bcrypt is non-reversible, so passwords cannot be “decrypted.” We do not store plaintext passwords anywhere." from the same page. – Matthew Dec 16 '15 at 14:20
  • Again, "Although accessed, all passwords, social security numbers and tax form information remain safely encrypted with a 2048-bit RSA key". Do you have anything that disputes that statement? – Mike Ounsworth Dec 16 '15 at 14:22
  • I see your edit. You're making an argument based on something *other* than the press release. Fine. – Mike Ounsworth Dec 16 '15 at 14:27
  • @Matthew: Ashley's Madison's password were protected with `bcrypt`, [this did not save them](http://arstechnica.com/security/2015/09/once-seen-as-bulletproof-11-million-ashley-madison-passwords-already-cracked/)... – WhiteWinterWolf Dec 16 '15 at 14:28
  • @WhiteWinterWolf True, although that was down to implementation errors. There haven't been any claims that I've seen made of similar errors in the Patreon implementation, and the same argument applies - you change that password, whether it can be accessed or not, and don't reuse it elsewhere. if you already did, change those ones too. – Matthew Dec 16 '15 at 14:33
  • I agree with your answer, in all cases (even if the source code analysis since the source is provided with the hashes) shows a weakness, the hashes will still provide enough time for the users to change their passwords. When you say "you've presumably already changed your Patreon login", you indeed recommend to change the password? I ask because as per my understanding, the OP asks if he would be safe without changing his password at all, "I don't want to go to the effort if remedial action is unnecessary". – WhiteWinterWolf Dec 16 '15 at 19:40
  • I don't think there was a choice provided if the OP has logged back into their account - I think it was a compulsory password reset, so unless they entered the same password again, should be ok! – Matthew Dec 16 '15 at 23:22
4

From a theorical point-of-view, the list of compromised data does not include anything regarding the password in any form, therefore as long as you trust this list to be complete about all leaked data then your password is safe. The main threats caused by the listed affect privacy, phishing and spamming issues.

However, I have some doubts whether this list is really complete since Patreon co-founder himself confirmed that password database was accessed and advised users to change their password.

While they claim to have applied the best practices and used bcrypt() to protect their password database, an Ars Technica article reminds that implementation issue could cause weaknesses still allowing hackers to extract passwords.

WhiteWinterWolf
  • 19,142
  • 4
  • 59
  • 107
  • Okay, so GIVEN: encrypted bcrypt hashes were stolen, but the encryption key was definitely not available for the given means of attack; AND: no other password leakage occurred; THEN: It is not practical to extract user passwords. – Riking Dec 16 '15 at 16:23
  • 1
    @MikeOunsworth: I can confirm for sure that the password has is present in the archive (table `tblUsers`, column `Password`), they don't seem further encrypted (the bcrypt prefix `$2a$` appears in clear), and in all cases there a numerous collection of RSA and SSH private keys provided with the archive (including a "localhost.key" RSA private key accompanying the deployment package), all this along with what seems to be a complete source code (BTW I just wanted to precise that it is not me who downvoted the other posts here, I rarely down-vote anyone). – WhiteWinterWolf Dec 16 '15 at 18:26
  • I see. If the haveibeenpwned guy did the analysis, then why didn't he include Password Hash in the above image? But alright, I'm convinced. I'll retract all my comments. – Mike Ounsworth Dec 16 '15 at 18:37
  • @MikeOunsworth: In count nearly 2 millions hashes in the dump. Why he did not mention this remains a mystery for me too... – WhiteWinterWolf Dec 17 '15 at 09:36