Using Google To Crack MD5 Passwords

stern writes "A security researcher at Cambridge was trying to figure out the password used by somebody who had hacked his Web site. He tried running a dictionary through the encryption hash function; no dice. Then he pasted the hacker's encrypted password into Google, and voila — there was his answer. Conclusion? Use no password that any other human being has ever used, or is ever likely to use, for any purpose. I think."

Salt

No, the conclusion is you should always use salted hashes.

Re:Salt

I agree. Also, fry them in bacon fat and add pepper.

Re:Salt

And blackjack ... and hookers. In fact, forget the hashes!

Re:RTFA

You're correct. You have totally invalidated the points I brought up in my post. Good show.

Re:Salt

No, the conclusion is you should always use salted hashes.

I agree, but this isn't something the user can do. I can't register for a site and say, "I need to remember to use salt!" The site has to implement it and implement it correctly.

The guy posting was posting from the perspective of the user, not the author of the system. The conclusion from the summary is still accurate since you can't make the assumption that salt is always used. The next best defense is a crazy fucking password.

Re:Salt

That's not true. The user can generate a string with something like dd if=/dev/urandom bs=21 count=1|openssl base64 , store that string, and append it the the true password each time the log in. This has exactly the same results as the site correctly implementing salting.

So what you are saying is the best defense is to use a crazy fucking password?

Re:Salt

Because if somebody gets that file, they've got your password. This way, they'll have to hack your brain, as well as your computer, to get at your password.

Re:Salt.. .so then develop

a rad ass custom mod chip that the user injects into the cerebral cortex and obdulla loongggatta and up down undah. The user then develops Tourettes Syndrome out the ass and has shit for brains now and only has to utter some crazy fucking ass phrase to seed a crazy fucking password in the solid-state gene-erator cuz they've gone fucking goddam crazy over that motherfuckin' chip in their ass and brain.

Crazy fucking luser. Crazy fucking assword. Crazy fuckin' whirled up world.

The above is the 1.0 tourettes pack, silver. Stainless-fucking-steel adds an additional language pack...

Re:Salt

Not entirely.
That adds a "local salt" but... courtesy of possible hash collisions there is another password that may work equally well.
by having the login function add the salt a straight rainbow lookup is defeated (unless you pre-computed a rainbow with the salt). As admin he could still enter the salted MD5, find a suitable password without salt, disable salting, get in enable salting, change the password. BUT a "normal" hacker without access to the DB tools and salting function of the app, but in possession of the hash table (and even the salt to some extent) would be defeated. if the attacker had the salt and hash table then with enough time the will break you login through rainbow tables, but not before.
-nB

Re:Salt

Better solution:

http://passwordmaker.sourceforge.net/passwordmaker.html [sourceforge.net]

One password for all sites, but a unqiue, "fscking crazy" password for all of them. You're welcome.

Re:Salt

agree, but this isn't something the user can do. I can't register for a site and say, "I need to remember to use salt!" The site has to implement it and implement it correctly.

The guy posting was posting from the perspective of the user, not the author of the system. The conclusion from the summary is still accurate since you can't make the assumption that salt is always used. The next best defense is a crazy fucking password.

This is why my passwords are themselves salted hashes. The likelihood of someone else using my passwords is the same as a regular hash collision, I get to use a separate password for each place one is required, and the hashing mechanism and salt are simple enough for me to keep in my head. End result: infinite number of easily generatable and retrievable passwords that look just like a hashed password when decoded.

Re:Salt

Rainbow tables? Salting breaks it.
Precomupted dictionaries? Salting breaks it.
Brute force and compare against the whole pw list? Salting breaks it.

Salting is your friend. Long salts don't cost much, but make many attacks completely infeasible. Unix has been using salted passwords since forever. Yet nthash *still* doesn't include a salt.

Re:Salt

Ice building up on your sidewalk? Salting breaks it.

Re:Salt

Pretzels missing that unique flail? Salting solves it!
Need something else to put on those fries? Salt it!
Need to make your friend's drink taste awful? Salt is the way to go.

(Somewhere along the line we left the analogy department :P)

Re:Salt

This is slashdot, we need a bad car analogy too. :P

Your car rusting too slowly? Salt solves that! :-)

Re:Salt

When a problem comes along, you must salt it.
Before the cream sits out too long, you must salt it.
When something's going wrong, you must salt it.

Now salt it! Salt it good!

Re:Salt

You're implying that salting on UNIX makes attacking the hash infeasible, this is simply not true.
Salting doesn't make breaking hashes infeasible, but it makes the attacker work harder, and makes certain highly efficient attacks infeasible.

There are only 4096 different combinations in the salting algorithm in crypt() will use which a brute forcer can easily iterate.
And I completely agree that 12 bits of salt is insufficient in a modern world. Which is why MacOS 10.4 and up uses 32 bits of salt, most Linux implementations use 48 bits of salt, and OpenBSD uses (a rather paranoid) 128 bits. Since it doesn't require any more effort from the user, and only a tiny amount of resources, there's no reason not to use a large salt.

Salting a known algorithm is almost pointless because as I just described salted passwords can be just as easily defeated if you know the mechanism
If you have the password hashes they you have the salt too. Either way, brute forcing one password is no harder. But it means you have to work harder to do a whole list of passwords, because each password has to be attacked individually.

Salting also makes precomputation (pre-built dictionaries and rainbow tables) infeasible. Every bit of salt in essence doubles the amount of storage for your precomputation attack. This is (partly) why a fairly effective set of rainbow tables for LANMAN hashes take only 500ish MB, NTLM hashes take 8.5 GB, but even for the old Unix crypt() it would take at least 2 TB. And don't even think about trying any precomputation attacks against OpenBSD; even if the user was stupid and restricted themselves to 5 digit alphanumeric passwords, your rainbow table would consume more storage than exists. Salting makes you attack each password individually, and keeps you from doing any work ahead of time.

this is why NT doesn't include salt.
NTLM doesn't include a salt because (1) MS is trying to maintain a semblance of backwards compatibility with some ill-designed challenge response authentication mechanisms, and (2) they haven't learned the lesson that salting is a valuable strategy to make attacking hashes more difficult.

Also salt was used on UNIX only because when shadow passwords didn't exist the system had to be protected against users that had the same password and could easily read the password file to compare.
That is one reason why salts were used for old Unix crypt(). The other was to make precomputed dictionary attacks harder, which is still a valid use. Today, the best reason to use a salted hash is to avoid rainbow tables.

Really, the modern reason to use a salt is to prevent the type of attack the original poster used, and to prevent rainbow table attacks. Both of these are good attack techniques, and salting completely moots them.

Re:Salt

You are correct that salting does not prevent nor make harder a brute force attack against one password.

It *does* breaks the Google attack, a precomputed dictionary, and rainbow tables, *even* if the attacker just wants *your* password.

Of these, rainbow tables is by far the most effective. Nobody computes their own rainbow tables. If I want to attack your hashed password, I'll download or buy a set of rainbow tables. Salting prevents this, because every salt value needs its own set of rainbow tables (or you have to include the salt rainbow table entries, which is approximately the same). Either way, using a 32-bit salt implies that to be equally effective, the total set of tables has to be 4 billion times larger. A 128 bit salt; well, you just can't create a set of rainbow tables for that. It just demolishes their effectiveness.

As you imply, there is a variant on salting which even makes plain brute forcing harder: don't store all of the salt. Of course this is (1) not widely deployed, and (2) imposes a high cost for legitimate use. Anyway, using repeated hash iterations is better, since you can't parallelize it.

How about "don't use your first name As your PW"?

No, the conclusion is you should always use salted hashes.

That's good advice for application developers, but the original post was offering advice to users. Still, even that is a bit of an overreaction. From TFA:

And indeed, the MD5 hash of "Anthony" was the database entry for the attacker. I had discovered his password.

Not to diminish this admin's accomplishment (it sounds like he's quite clever), but doesn't this boil down to "don't use your name as your password"? Or better yet, "don't use any proper name as a password".

Keep in mind that this was a hash of a userid (not a password) that was captured in a google index, and it's highly unlikely that someone will choose a userid on a google-indexed site that just-so-happens to be your 10+ character password that has mixed-case and special characters. I think the same "good password advice" still applies, even in a google-world.

MD5 Lookup Site & Names

For those of you who missed it in the article, the has was:

20f1aeb7819d7858684c898d1e98c1bb

And sure enough, if you read the comments to the blog, there is a site called http://md5.rednoize.com/ [rednoize.com] that reveals that the hash is "Anthony." So although Google helped, there appears to be resources online for it (if you don't have your own Rainbow Table mega database).

He could have discovered this if he had used a database complete with names, something I don't think would have been too difficult for him.

This Google search idea is kind of moot if the user uses some very basic password construction such as what I've commented on before [slashdot.org]. Also, as the blog mentions, this discussion is worthless if WordPress used salting [wikipedia.org] which is related to nonces used in security engineering [wikipedia.org]. I think that stuff has been around for, what about five years now? Wake up WordPress!

5 years?

Try decades! The good old days of Unix even had salts (even if they were just two bytes)

Re:MD5 Lookup Site & Names

He can't be much of a "security researcher" if someone hacked his own website.

Re:MD5 Lookup Site & Names

He didn't write the WordPress software, and presumably doesn't have the time to audit every bit of code it uses.
I doubt Bruce Schneier himself audited the entire Movable Type codebase, which he uses for his blog. Does that make Schneier "not much of a security researcher"?

Re:MD5 Lookup Site & Names

Crap. From their "about" page:

Additionaly everytime when you enter a non-md5 hash string into the search field, the md5 result for that search strings gets stored in our database for future use.

Thanks for warning me. I tested to see if my password was in there... it is now!!!

Obligatory

In Soviet Amerika, MD5 passwords crack you.

Re:Obligatory

What about the flip side: Using Crack to Google MD5 passwords?

My uneducated respose would be:

What about the flip side: Using Crack to Google MD5 passwords?

2343e9f361fea282776586d7056025db

I wouldn't be too alarmed.

Most MD5 password hashes, such as those used in *nix, are salted [wikipedia.org], and hence secure from this sort of vulnerability. That Wordpress uses unsalted MD5 sums to store passwords boggles my mind. It shows that the developers know even less about cryptography than I do. That's scary.

Re:I wouldn't be too alarmed.

That is totally unfair to the wordpress developers. Just because they don't care doesn't mean they don't understand.

Re:I wouldn't be too alarmed.

You do realize that most businesses (and therefore most websites you have accounts on) just store passwords plain text because it's easier to do tech support that way. Salted hashes are better than unsalted hashes, but most don't bother hashing at all.

Re:I wouldn't be too alarmed.

That Wordpress uses unsalted MD5 sums to store passwords boggles my mind. It shows that the developers know even less about cryptography than I do. That's scary.

Oh it's even better than that. It stores your md5 password in a plain text cookie, and if it receives such a cookie, sets an $already_md5 flag to true that's then passed to wp_login() which then just compares it literally against the unsalted md5 entry.

<guinness>Brilliant!</guinness>

Dark Helmet

So the combination is 827ccb0eea8a706c4c34a16891f84e7b. (lifts mask) That's the stupidest combination I've ever heard in my life. That's the kinda thing an idiot would have on his luggage.

Let me guess

The password was hunter2? [bash.org]

Re:Let me guess

What's with all the stars in your post?

Salt

This goes to show the importance of using the technique of adding salt values to passwords before hashing. Also, your salt value shouldn't be a common word ( or something which would make a common word or phrase in combination with something people are likely to use in a password).

In itself nothing new

But if I ever need to run a hash against a password database, I'll remember this lesson and first perform a Google search. Saves a lot of time and CPU cycles.

I am already doing this for telephone calls I cannot place. If it's an institution or a person that is calling because of profession, the chances that the telephone is listed somewhere on a (search engine) accessible web page is *very* large.

just look for "cf99"

5f4dcc3b5aa765d61d8327deb882cf99 is the MD5 hash for 'password'.....

search enough systems and you're bound to see some doosh has used it.

Re:french bitch

I just hate douche bags who can't spell.

Spelling errors can make your password more secure!

Great job..

John cracks the hash in 0s and that's even faster than google does.
Do I also get a slashdot story if I crack a SHA1 hash with google?

Found my password

Damn it found mine... 286755fad04869ca523320acce0dc6a4

Sorry...

But how is this surprising, I've done that many times, it's basically just a quicker way of searching milw0rm since it's likely that at some point every hash has been seen (and therefore cached) by google

Been there. Done that.

I have personally been using Google this way for a while. This is the first thing I do when I encounter a passwd hash during a pentest. This is a technique that works very well especially for hashes produced by random apps that you have no idea what hashing algorithm they use. It works well not because the public passwd hash databases indexed by Google are large (they are not), but because they are very diverse, both in term of number of algorithms (MD5(), MD5(uppercase()), SHA1(), etc) and in terms of number of hash formats (hexadecimal value, decimal value, base64, etc).

And above all, it only takes 2 sec to perform the Google search.

on a related note...

... I wish Google would collect/show/use checksums of files in search results. It would be a great way to find identical files.* Thousands of uses:

• I found this file on my computer and I forgot where it came from.
  
• I downloaded this file but I forget where I got it. It's too big to email so I would like to send a friend a link to the original file.
• I want to see if anyone has taken this pic from my site and posted it elsewhere.
• This download is taking FOREVER. Is anyone else hosting this exact file?

and many, many more. I had this idea years ago and sent it in to them but haven't heard anything since. I don't want any credit**, just implement it and let me know when it's up and running! And the funny thing is, I'm sure Google is already checksumming every file as part of how they do all their magic. All they have to do is post the data!

* and, since collisions are possible, it would provide a nice corpus to study collisions, etc. in the real world.

** this isn't an entirely original idea. Linux distros have been posting checksums for years as a way to let users verify that their downloads were not corrupted; as a bonus, I (and I'm sure some others) have done searches of those values to find sites hosting that particular release.

Man, I need to change my password NOW.

Results 1 - 10 of about 101,000 for d41d8cd98f00b204e9800998ecf8427e. (0.04 seconds)

Google indexed already

As I write this message, this story has been posted for only about an hour. However, a google search for the hash already throws up this article as the second link. Damn they index the web fast!

Google Hash

I know of efforts in this regard that date back 3 years or so, although I'm not aware of whether these projects are still online. There are some good discussions about the idea at http://ibneko.livejournal.com/668715.html [livejournal.com] and http://www.dragoslungu.com/2007/06/22/google-md5-hash-search-engine/ [dragoslungu.com]. My interest is that I'm attempting to get Google to index such hashes at http://www.nth-dimension.org.uk/utils/ghash.php [nth-dimension.org.uk]. In my case I'm actually attempting to get Google to cache my hashes to minimise my storage costs as rainbow tables take a fair bit of disk space to store although the idea hasn't been particularly successful due to Google algorithms :(.

Credibility?

Am I the only one who thinks that a "security researcher" whose site gets hacked and is about as credible as an accountant who fails an audit?

And for his sake I really hope that he knew about rainbow tables and just decided for some indecipherable reason not to mention that they are far more effective for password cracking than Google searches.

And who submitted this story to Slashdot with the sensational summary about "any password used by anybody, ever" being vulnerable to Google searches? That's an easy enough claim to completely debunk by taking MD5 hashes of several passwords and sampling which ones come back. Let's see:

92259762923b4e79d2073ecb03217462 (hash for 'july2007') - Nothing
6e933f3054f533c63dd59479ca9f4b6f (hash for 'hello_world') - Nothing
2c6c8ab6ba8b9c98a1939450eb4089ed (hash for 'abc123') - Google found this one as an md5 example
6a51f1fe97bdebece7652842a0e2351e (hash for 'pickles') - Nothing
5eaaf94141c371ce96675aa6445003c4 (hash for 'happy') - Nothing

So basically not even common words get picked up by Google, much less "any password used by anybody else, ever".

Re:Credibility?

Your strings have newlines in them, maybe you meant:
echo -n happy | md5sum

most password fields don't accept newlines, so trying without them:
3e652df0f1332cfc9df779d49667defc - still nothing
99b1ff8f11781541f7f89f9bd41c4a17 - still nothing
e99a18c428cb38d5f260853678922e03 - abc123
fd03204cfdc557b0f0d134773ae6fff5 - obscure, it finds a flash app on a site called pickles and things
56ab24c15b72a457069c5ea42fcfc640 - happy

So it is still not that much of a problem, but at least happy is on the list.
I wonder if negative outlook words are more or less secure?

Re:Credibility?

I looked these up on google, and they directed me to some slashdot page...

No worse than Subversion

It's no worse than Subversion's insistence on storing user passwords for any protocol but SSH public keys in a local plaintext file.

Do not *EVER* allow a Subversion system to use the same passwords as the user system, and if you have access to the user's accounts, run a check of their stored Subversion passwords to make sure they didn't use their same password somewhere else as for their local user account.

My Hash

09F911029D74E35BD84156C5635688C0

new worm spreading

Google is now shutting down servers and re-routing as they try and halt the spread of the newly-detected worm that tries to do a DOS on google, by making affected machines do a google search with random strings that look like 0cfa9f600839f57e90e5559b8ee54864

But seriously, as fun as it is to look up all your hashed responses on google, I'm going back to por... work :)

You might also want to check out http://utilitymill.com/utility/Goog_Your_Hash [utilitymill.com] to see if your password is 'safe'.

No, it means that as the attacker...

No, it means that as the attacker, security through obscurity is not only advantageous but should be highly encouraged. It doesn't work when designing security, but it's quite effective in conjunction with other basic protections when used in limited cases for nefarious means. For instance, you know where to find resources on MD5, SHA-1, et cetera but do you know where to find resources about WQEWERSDF (the crappy but effective and thoroughly obfuscated hashing method i just made up)?

Consider that with the fact that the vast majority of everybody, but especially incident responders, sysadmins, et al can barely read strace output, must less reverse my backdoor, regardless of whether i pack it full of fun for the RCE or not.

Hash DBs have been around for years.

Like GData [gdataonline.com]. That has been around since the summer of 2005.

Finding your hashes on google

I suppose about anything gets logged nowadays.

I like to search for a part of my password on my computer every once in a while to see if programs store in plain text.

If my password would be "zxcvbbn.34#$" then I would search for "vbbn".

So for google; you can search for a part of your hash but still be secure.

Hold the pepper...

but can I get some salt with my password? ;)

HMAC

You can use the standard HMAC algorithm on top of MD5 or SHA1 to adequately hash a password. It's much better than simply appending or prepending garbage to your cleartext.

PHP5 has a function built-in and I'm sure most other languages have comparable implementations available. It's not fool proof by any stretch, but if you use a randomly generated fixed "key," it at least prevents someone from using Google to discover the cleartext.

Better still: Use a unique value for the account + a randomly generated key. For example:
    Key = "c,.rcph203p9h"
    UserID = 12
    HMAC_KEY = "c,.rcph203p9h::12"

That will make it computationally difficult to crack, as each password must be brute-forced individually.

Salt it

Use no password that any other human being has ever used, or is ever likely to use, for any purpose.

I'd take that advice with a pinch of salt. :-)

I pepper my hash...

Salt is nice, but pepper is better.

I want to thank you all...

... As I read through the replies to this article it occurred to me I should investigate the possibility of purchasing large quantities of metal futures. So I headed over to http://www.metalprices.com/ [metalprices.com] and realized it was a good time to buy aluminum and tin.

Thanks again!!

Make the hashing inefficient

OK, so we all agree that we should salt passwords with something. But even without that we should also make the hashing algorithm as slow as possible, perhaps running MD5 recursively 10 times, or using a slower algorithm on itself multiple times, or running different algorithms on the output of others etc. Most of the time this isn't going to have a significant effect on performance of the whole system and it also becomes infeasible to create a rainbow table because each hash will take a lot longer to create.

on salted hashes...

You don't realize how many developers have no clue what a salted hash is and how it works. I've had discussion with developers who weren't complete morons when it comes to programming that had no fscking idea what I was talking about. They fully knew what a hash was, but it was unconceivable that you could store a salt value next to a hash. They were massively deniyng that it could work, trying to make fun of me "it's impossible, that's not how cryptographic hashes work" etc.

On a related side not you'd be amazed at the number of developer that have no fscking clue about how public key cryptosystems works.

Reverse MD5 in multiple databases

The site http://md5.noisette.ch/ [noisette.ch] try to reverse MD5 hashes in looking up into multiples other existing databases. It "knows" most of the hashes posted into the comments...

Re:Don't panic!

You never have used rainbow tables have you? You're in for a rude awakening.