New Mozilla Firefox 1.0.3 Exploit

An anonymous reader writes "News sources are reporting that a 'killer' new Firefox exploit has been revealed today by FrSIRT who warn that this 0day exploit/vulnerability (as yet unpatched) should be rated as critical. Summary of the exploit: If a user clicks anywhere on a specially crafted page, this code will automatically create and execute a malicious batch/exe file. Proof of concept code supplied by FrSIRT."

Uh oh!

Hey everyone let's use IE now, because it's safer than Firefox.

Oh, wait.

Re:Uh oh!

Start your stop watches and let's see how long before a patch is forthcoming. To my mind that is the real test Then compare that time to M$' response time.

Re:Uh oh!

Start your stop watches and let's see how long before a patch is forthcoming

Might as well hit stop now. The bug isn't exploitable any more since update.mozilla.org itself has been fixed.

Re:Uh oh!

Are you telling me you expect a noob to know this? How is my grandmother supposed to know of this?

Know what? Whats wrong with your grandma, Alzheimer's?

Why doesn't the little red arrow (update icon) display yet?

Because you don't need to update anything. It was fixed on updates.mozilla.org. The site needs to be in your white list of sites that are allowed to install software to be vulnerable. I'm sure they will have a more permanent fix later at some point, but the current exploit no longer works. Go ahead and try it.

So, as far as I'm concerend -- it's not.

But you're a bit of a fool, so I'm not sure your opinion counts.

Re:Uh oh!

In Firefox, to stop this vulnerability:

Web Features->Allow web sites to install software

I'll switch to MS IE as it has no known serious vulns

Internet Explorer Long Share Name Buffer Overflow [secunia.com] Highly Critical

Yeah... whatever. I don't mind if you would rather use a browser with a known serious security problem, but saying that IE has no known serious issues is misinformed.

Re:Javascript ! Will it ever go away ?

" javascript. The language that has no purpose anymore."

Look into Firefox's chrome directory and say that again.

Has he dropped this in bugzilla as well?

Because THAT, with some documentation, would be helpfull. Still, as long as it doesn't create *nix r00tkits on the fly on my box, I'm on the safeside :)

Re:gah

Now we'll see Microsoft going "OMG DON'T USE FIREFOX YOU CAN'T EVEN CLICK ON SOMETHING SAFELY!".

You mean like the F/OSS evangelists do everytime a flaw is found in Internet Explorer?

However, I do think there is an important lesson in here - a lot of open source advocates have set an unreasonable level of expectations by proclaiming the amazing magic of open source: A fantasy world where every line is thoroughly vetted by thousands of super-experts, and if the source is available that instantly disproves the existence of malicious intent (put a trojan out, mark in GPL and make the source available, and I'd bet a lot of the converted would immediately download and install blindly. There are countless OSS projects where no one but the author ever bothers looking at the code).

Re:gah

I have to disagree. This sort of exploit is extremely worrying.

At first, Mozilla fans (me included) all said "the chances of Firefox getting 0wned by exploits is very slim, Mozilla is secure by design -- IE isn't".

By about 0.9 or 0.10 the holes started pouring in -- but it was ok: "This is simply Mozilla Foundation's bug patching contest, they are working FOR us instead of AGAINST us."

After this it wasn't only white-hat mozilla funded security experts that started showing there was holes in the code. We changed our story again and, somewhat rightly, pointed out that "these are very theoretical and it would be very hard to use this to exploit a computer like IE can".

This is a really big problem. This will get exploited like crazy as it seems exceptionally easy to do. Not only that, I expect the only fix from Mozilla will be as usual, a 5MB binary installer with the files changed. This is unacceptable on a 56k modem and people just won't bother upgrading to a secure version.

Re:Pretty serious exploit

You converted 45 % of your family to Firefox?

Yup - secure...

Maybe it's time to accept Firefox has it's fair share of exploits?

And the best part, is the patch management system in Firefox is so damn poor (ie. non-existant), getting these patches distributed to end-users is a real damn chore (assuming they are distributed at all).

Re:Yup - secure...

I'm running Firefox 1.0.2 and it displays:

1. No update notification
2. No red blob in a corner.
3. No dialog box telling something new is available.

The feature seems unreliable at best.

Re:Yup - secure...

The problem is that it:

a) Only works on Windows,
b) Makes you install the entire installer again instead of a 'diff'-style patch,
c) The installer is nearly 5MB, which means it's too big for most to download on 56k or GPRS

Another problem with the 1.0.1, 1.0.2 and 1.0.3 updates is that they all required 'staggering' based on language becuase MozFo doesn't have the sort of server infrastructure to serve millions of downloads at once.

Re:Yup - secure...

Patching is something where Firefox really needs to catch up on.

One of the advantages of IE is that when an exploit comes around you just send everyone a 300k file instead of 20MB of browser. With Firefox, you have to send them an entire browser every time 1 exploit comes out.

What Firefox needs is some sort of patching element built in to deal with patching the browser instead of forcing a complete downoad. It's not that Firefox cant do this. In fact, since most of the code is spread out across many files it should be a cakewalk to just update the affected file(s) automaticially with little to no user intervention. This would keep the file size download to a very minimum, allow it to update more frequently without waiting for a point release, and be easier to handle for people who dont know or care about security issues.

Re:Yup - secure...

This is already being worked on and should be in 1.1. ^_^ Check out ben's blog [mozillazine.org] about it.

A quote: "Darin has figured out how to get binary patching working, and is working on a system for incremental background update download."

Re:Yup - secure...

Firefox 1.1 will have support for binary patches, meaning no more full application download to fix a single bug.

Nasty

That's nasty! I'm glad that in Linux files aren't automagically executable when you give them a certain name :)

Re:Nasty

Unfortunately, the exploit could have just as easily created a file starting with #!/bin/sh, and passed 555 as the 'permissions' argument to createUnique [xulplanet.com].

Why on earth the browser thinks it's necessary to allow scripts to create executeable files is beyond me.

Re:Nasty

Well, in Windows it would only have administrator priviliges if the user was dumb enough to run Firefox as an administrator. ;)

This was reported to bugzilla some time ago!

This was reported to the mozilla bugzilla a while ago. https://bugzilla.mozilla.org/show_bug.cgi?id=29269 1 [mozilla.org]

Re:This was reported to bugzilla some time ago!

interesting - even when you go past the "cant view bugs from slashdot" stuff, it seems access to this bugreport has been denied. Yay open source!

Re:This was reported to bugzilla some time ago!

It's a severe security-related bug, so the bug report is restricted. This is meant to stop script kiddies from scanning bugzilla for unpatched exploitable bugs. Unless you're a disciple of the full disclosure persuasion, that is the correct way. The Mozilla Foundation discloses all bugs when a patch is available to the general public.

It's "Open Source", not "Sploitz4Free".

Re:This was reported to bugzilla some time ago!

And yet, when Microsoft does this, somehow it's "reprehensible".

And on the flip side - where's all the folks who defend Microsoft's practices? Shouldn't they be also standing up here and saying how responsible the Mozilla Foundation is?

Really - why try to paint this as an "open source vs. Microsoft" issue? If anything, this is the usual "full disclosure" vs. "reponsible disclosure" vs. "no disclosure" debate. The underlying development model has little to do with it.

Summery?

Exploit summery? Well, the weather is improving but I doubt that the exploit caused it.

Reported and temporarily fixed

Bugzilla bug 293302 [mozilla.org] has been filed. A temporary fix has been implemented on UMO.

Tried it on my Mac...

didn't work

FrSIRT's Post!

It looks like a hacker alias, but it really stands for French Security Incident Response Team. Exploit description cached here [64.233.161.104].

Stolen exploit

They were already working on patching this, but it was stolen before they could finish and leaked to bugtraq with LIVE material in the exploit (it's not a proof of concept, folks!) and no explanation or advisory.

Reminder: Bugzilla blocks /. referers. Copy URL and paste in new to view. (Beware Slashcode's extra spaces.)

https://bugzilla.mozilla.org/show_bug.cgi?id=29269 1 [mozilla.org] %lt; Original security bug (probably still blocked to outsiders to prevent someone stealing it before mitigation)

https://bugzilla.mozilla.org/show_bug.cgi?id=29330 2 [mozilla.org] %lt; Duplicate (reported after leak)

They are going to release a 1.0.4 shortly, I gather.

Still more timely than most of Microsoft's advisories... despite their earlier announcement. http://www.eeye.com/html/research/upcoming/index.h tml [eeye.com]

Leaked known bug

A^C^E, a Firefox security researcher, is claiming on Addict3D.org [addict3d.org] that this is a 0day duplicate of a leaked, known bug. He says, "I suspect that my server was compromised, and I am currently using my contacts to find the culprit and bring him to justice."

Also, bugzilla.mozilla.org is claiming they've been slashdotted. Go easy on em.

I keep clicking on the exe files...

...but Firefox keeps suggesting I run it with Wine. I don't get it, I'm not thirsty. I'd rather run it with a nice plate of steak and eggs.

Tried the test exploit they supplied...

...with Firefox 1.0.3 on Windows 2000, and it didn't execute anything. Anyone else try it on Windows?

Possible workaround:

Uncheck Tools > Options > Web Features > Allow web sites to install software

Are you sure?

Just curious, I downloaded the page and loaded it up on several systems:

Win XP, Firefox 1.0.3
Win 2k, Firefox 1.0.3
FreeBSD, Firefox 1.0.3

and none of them did anything. The javascript looks like it should save a file (c:\booom.bat) and run it which should echo "malicious commands here" and wait for a keypress.

Is this truly an issue with Firefox and not some other software? If so, any ideas why it doesn't work?

Re:Are you sure?

Reading the Secunia [secunia.com] explanation:

Successful exploitation requires that the site is allowed to install software (default sites are "update.mozilla.org" and "addons.mozilla.org").

So, unless you've whitelisted the exploit site (which generally would mean it's a site you trusted enough to install an XPI from), or the Mozilla website has been compromised, the exploit won't work.

Re:Are you sure?

That is incorrect. The exploit works by loading a page from a trusted site (one of the mozilla.org sites on the whitelist), then taking advantage of another Firefox bug to run some javascript in the security context of the trusted site.

Re:Are you sure?

We made some server-side changes on update.mozilla.org to mitigate the attack.

This isn't much of an "exploit"

The actual advisory page is here [frsirt.com]. The "Solutions" section says this:

Disable JavaScript, or disable the "Allow web sites to install software" option [Tools - Options - Web Features].

Why would anyone run routinely with "Allow web sites to install software" enabled ?

Re:This isn't much of an "exploit"

> Why would anyone run routinely with "Allow web sites to install
> software" enabled?

1. It's on by default
2. We naievely assumed that the whitelist of web sites allowed to install software did its damn job.

Re:This isn't much of an "exploit"

The design is flawed.

Agreed -- and even worse, the design was copied directly from Microsoft's ActiveX system!

It's a bit frustrating to see Firefox advocates continually prattle about "Security ... activex LOL", when FF does in fact have a nearly identical feature as ActiveX. And when there's a mechanism for installing program files from webpages, people will tend to find holes in the sandbox. Hopefully this quiets the "better by design" crowd.

Re:This isn't much of an "exploit"

See, the thing is, the whitelist is broken.

Firefox is only supposed to download and install from things in the whitelist. Unfortunately, it doesn't check the site correctly, and therefore can be tricked into thinking another site is mozilla.org

So even though you "secured" your system, it's still vunerable because as long as you have anything in your whitelist (especially mozilla.org or the defaults), you're vunerable.

Once the whitelist is working again properly, this won't be an issue.

New FrSIRT Vulnerability

FrSIRT Vurnerability Alert!!

FrSIRT will go down 2 minutes after the start of a brutal Slashdotting.

This shouldn't be a competition.

There's not many comments yet, but most of them have a similar theme: " Oh no, now Microsoft and Internet Explorer users can get payback for all the trash talk we've thrown at them." Then they rationalize it with, "But, MS and IE are way worse because of quantity, severity, and duration until patch."

Now think about it for a minute. Who are you really at war against? Security exploits and the people who would exploit them, or browsers other than the one you use and the people that use them?

This reminds me of the days when Mac zealots would get all freaked out every time PC's got faster. "OMG, this is bad news! Now there are 3GHz PCs for under 500 dollars!"

This really boils down to people rating the quality of Product A compared to the suckiness of Product B. Personally, I've been using Products A, B, and C for a long time. When there is a problem found with Product B, that really doesn't make Product A perform the task I use it for any better.

If you want to call yourself a truly knowledgeable computer user, then you have to acknowledge that Products A, B, and C all have their strengths and weaknesses and therefore have tasks their better suited for as well as tasks in which they're not the best solution.

If you look at it from the proper perspective, every time an exploit is found by good people before bad people have a chance to do harm with it then it is good for everyone.

This particular exploit also demonstrates how foolish it is to posture and sling insults. The whole time FF users slung insults at IE when exploits were found, this exploit was there lurking below the surface waiting to be found.

Let applications that are without exploit cast the first stone. Since that's never going to happen, argue your cause based on its merits.

Secunia: Extremely Critical

Secunia have already released an advisory explaining how the exploit works:

http://secunia.com/advisories/15292/ [secunia.com]

This is the first Firefox exploit that has received the rating 'Extremely Critical'.

--- Extract from Secunia's site ---

Description:
Two vulnerabilities have been discovered in Firefox, which can be exploited by malicious people to conduct cross-site scripting attacks and compromise a user's system.

1) The problem is that "IFRAME" JavaScript URLs are not properly protected from being executed in context of another URL in the history list. This can be exploited to execute arbitrary HTML and script code in a user's browser session in context of an arbitrary site.

2) Input passed to the "IconURL" parameter in "InstallTrigger.install()" is not properly verified before being used. This can be exploited to execute arbitrary JavaScript code with escalated privileges via a specially crafted JavaScript URL.

Successful exploitation requires that the site is allowed to install software (default sites are "update.mozilla.org" and "addons.mozilla.org").

A combination of vulnerability 1 and 2 can be exploited to execute arbitrary code.

NOTE: Exploit code is publicly available.

The vulnerabilities have been confirmed in version 1.0.3. Other versions may also be affected.

Solution:
Disable JavaScript.

Fixes for large sites

For people running Firefox in a business or school with centrally locked down settings I think a quick fix might be to add

lockpref("xpinstall.enabled","false");

xpinstall.enabled seems to be the preference changed by "Allow websites to install software"

Linux and MacOS vulnerable, too

This exploit will work on Linux and MacOS, too, if anybody bothers to write an attack for them.

The basic problem is that the Mozilla developers, in their futile attempt to create a "platform", put in a mechanism comparable to Active-X - a way to dynamically download executable programs. Of course, they tried to make sure this "feature" could not be used for purposes of evil. Like Microsoft, they failed.

Understand, this isn't subtle. The code uses built-in Mozilla JavaScript extensions to create a local file in a very straightforward way. It then calls "nsILocalFile::launch()" (which does exactly what you think it does) to launch it. Those are capabilities that shouldn't be in a browser's JavaScript engine at all.

Having designed in a potential security hole big enough to drive a semitrailer through, they tried to make it "secure" with the usual crap approaches - signatures, lists of trusted sites, and disabling for certain types of URLs. They failed. They forgot to make those checks for "favicon.ico" files (Mozilla's implementation of a Microsoft icon-in-the-toolbar gimmick.)

Plugging that hole is not the answer. The problem is more fundamental. "nsILocalFile::launch()" needs to be removed. Browsers have no business launching arbitrary executable programs. Period.

Re:Linux and MacOS vulnerable, too

this problem is equivalent to xpinstall having a buffer overflow exploit which allows code execution.

No, it's not. This isn't anything subtle like a buffer overflow. This exploit uses standard features to download an executable (which shouldn't be allowed) and then execute it (even worse). This is a designed-in hole. It passed Mozilla's code review on April 9, 2002.

Personally, i'm all for removing extensibility of firefox, dropping support for helper applications and external view source. are you really a proponent of such things?

Yes. The Netscape/Mozilla "browser as platform" thing didn't work out. That's why Firefox exists. Firefox has legacy code from the Mozilla era, and much of it needs to come out.

Is this a design issue that will breed more bugs?

In a nutshell, Firefox has the idea that some sites are privileged (namely the sites on the whitelist for installing software), it lets privileged sites have a dangerous degree of control over the user's computer, and it has at least one way for unprivileged sites to execute code in the context of a privileged site.

What are the important differences between this and Microsoft Internet Explorer? In MSIE some sites are in the Trusted Sites or Local Machine zones and therefore privileged. Such sites have a dangerous degree of control over the user's computer, and there have been many ways for unprivileged sites to execute code in the context of a privileged site.

Is Firefox doing something better than IE in its design, or are we going to see a whole class of bugs like this one in the future?

Does it affect the mozilla suite? (seamonkey)

Well that's the essential question. If it doesn't I'd rather flee to mozilla suite than IE.

Trusted Sites Only?

The security advisory doesn't explain it too well, but it it seems to imply that this only happens with sites that you've added to your list of sites trusted to install software (in which case it isn't really much of a problem).

Re:I'm sure everyone whill complain

And everyone will say ":oh no firefox is a security risk" whaaaa. well this isnt really the case and is overstating things just a bit. When it comes down to it firefox still has many quicker fixes and the bug is probably already fixed by now.

Perhaps the bug is already fixed in the dev tree, but this is irrelevant if the fix takes 3 months to deploy to users. Hopefully, the fixes to the auto-update system coming up in 1.1 (where a "security fix" does not consist simply of "re-install the whole of Firefox with this new version") will make the whole deployment aspect faster. Although I have to say, Firefox 1.0.3 seemed to follow quite quickly on the heels of 1.0.2, which is encouraging! :)