Slashdot Mirror
Group Releases Anti-Disclosure Plan
dki writes "SecurityFocus reports that the Organization for Internet Safety (OIS), a group of 11 of the largest software and security companies, has released a public draft of a proposed bug disclosure standard. The document outlines a process for reporting and disclosing bugs that aims to eliminate releasing exploits to the general public. Not surprisingly, the OIS was founded out of a Microsoft-hosted security conference. Comments on the draft will be accepted until July 4th; the final copy will be released at the Black Hat Conference in Las Vegas."
"As discussed in paragraph 7.3.7 above, the Finder and Vendor should act in concert to release their respective advisories nearly simultaneously, and only after a remedy is available."
It's all in that last phrase "only after a remedy is available"
1337 Reveals limited disclosure plan
Eight hacking groups join together to set an official standard for limiting disclosure of software security holes.
#1337, Efnet â" Eight computer hacking groups rounded up a three-day Exploits in Computing Forum on Thrusday by formally announcing a coalition against full disclosure of vulnerability information, ending a week of intense speculation, and immediately sparking controversy.
WanSan, TCPuke, NetLoft, HeavySak, BitEvil, SYNergy, HPLat, and DownScope joined together to declare they would immediately begin following a policy of limited public disclosure of security vulnerability information. Members of the coalition who discover new vulnerabilities will omit from their initial public advisories any details about how a hole might be exploited in an attack, and will not include code that demonstrates the bug. Thirty days after the first advisory, a more detailed notice can be released under the rules. Full disclosure of the vulnerabilities will be shared only among the members for âtestingâ(TM) purposes.
"We felt that as responsible industry leaders, we, as a voluntary organization, are going to follow a set of reasonable standards," said DXNo, manager of 1337â(TM)s intrusion exploitation, in an interview.
1337 will also draft a proposed international standard for notifying vendors and the public about newly-discovered software security bugs, following the group's limited disclosure ethic. The organization will admit new members, under an as-yet unwritten set of bylaws. The initial draft of the limited disclosure ethic will limit the disclosure to the home pages of the vulnerable sites.
A chief objective of the group is to discourage 'full disclosure,' the common practice of revealing complete details about security holes, even if publication might aide attackers in exploiting them. The group believes that any type of full disclosure would assist software vendors into patching various vulnerabilities before they can be widely exploited.
Publishing complete information, and sometimes "exploit" code that demonstrates a vulnerability, is de rigueur among many computer security professionals, who argue that malicious hackers can acquire the same information themselves, and that network administrators and security gurus often need technical details to properly defend themselves from attack.
But Culp criticized the practice in an essay published on a Microsoft Web site last month, and blamed "information anarchy" for the epidemic of malicious worms that have struck the Internet in the last year. "It's high time the security community stopped providing blueprints for building these weapons," Culp wrote.
Under the plan, member groups would share detailed information during the 30-day grace period with âoeother communities in which enforceable frameworks exist to deter onward uncontrolled distribution.â The last category would allow member groups to share details with one another. "They're not going to ban it among themselves," says Levy. "They might be willing to limit the public access to this information, but I highly doubt that they'll limit it among each other."
"People have to do it Microsoft's way or they'll have this group telling them that they're acting irresponsibly," says Maiffret. "It's going to drive people into the underground, and could lead to more people breaking into computers." The majority of members in 1337 agree with Maiffretâ(TM)s assessment.
"We are not trying to form a secret society of exploiters," says CKLawz "We are just creating a standard... This represents one of the first process standards between security companies and vendors."
wyZopa1 estimate that it will take one or two months to produce drafts of the proposed RFCs. He emphasizes that the standards would not just limit vulnerability disclosure, but would also spur vendors to be more responsive to security vulnerability reports. "My goal in the RFC is to have equally stringent standards for vendors and exploiters," says wyZopa1. âoeWe worked hard to discover these vulnerabilities, the developers should work just as hard to fix them. Providing them with all our tools without compensation is not what software development is about.â
Karma: SELECT `karma` FROM `users` WHERE `userid`=138474;
Face it. Anyone can do security. All you need is the will, the drive, the talent, and the know-how. Which means, me, the average geek, ex-blackhat hacker, technologist, futurist dude.. can play with stuff, learn stuff, share with others, and build my CAREER off it. Not after this.. This would keep all the knowledge I gain from other professionals, from bugtraq, etc, out of my hands, forcing me to subscribe to these people's services, philosophies, etc... no. I will not. This will not stop hackers, this will only stop the admins who are trying to keep everything together. This will be a tax on the admin who manages 1500 machines, and has no time to read simple stuff.. Hackers, don't need bugtraq. Its useful, its a resource, etc, but if they want in bad enough. Fuck em all.
Why don't they ask themselves why is their product so weak and vulnerable, instead? How much of those vulnerabilities were exposed and fixed *because* they were exposed? It's a GOOD thing. There was a bug, people found it, tried reporting it, got told to fuck off, published it - and the bug got fixed.
Well, i don't think obscurity can really improve security. Most hackers find the information they need anyway. This is just an attempt of the software companies to cover up their lousy code. If they would implement secure code in the first place, none of this would be nessecary. Way to go closed source!
- Never underestimate the power of human stupidity.
Why would anyone follow these guidelines? It might piss off these companies, but anyone who really cares about security would realize that giving the vendors the exclusive right to disclose flaws (regardless how much time has passed or how many systems have been compromised) prevents people from making an informed decision to yank these programs until a solution is identified.
Mapped to the real world, it's like some idiotic Police Chief knowing damn good and well that several pizza delivery drivers are mugged every night when they go into a four-block area... but refusing to say anything - not even warning these drivers to avoid the area for a while - until after the muggers have been convicted, sentenced, and in prison for a month.
For every complex problem there is an answer that is clear, simple, and wrong. -- H L Mencken
can't hurt me.
rehab, captain ahab, you're chasing the wrong fish!
To me, it sounds like this is only relevant to organizations that maintain closed-source code trees, since the impetus on fixing the bug lies with their in-house development staff rather than with a decentralized team of programmers. It's to "protect" the company's bottom line from being affected by the propagation of bug exploits for which they have not done their part in fixing.
On the flip side, this mentality rarely exists in open-source code projects due to the lack of a need to safeguard against decreased shareholder value. Also, IMHO, open-source developers continue to maintain a sense of pride in their work that is for the most part missing in closed-source projects due to the direct recognition (and therefore, culpability) that is possible by putting your name to a piece of software that others can peruse and make comments and criticisms about. With closed-source projects, the company's reputation is primarily at stake while their programmers are largely hidden behind the web of corporate logos and NDA's, and therefore may not have reasons as strong as those of OSS developers to fix and release bugs in a timely manner.
So in other words, it's to protect lazy software companies from being held accountable for their own poor coding practices and lacadaisical maintenance policies.
Rule #1 -- Politics always trumps technology.
The document mentions only Finder, Vendor etc. What about the user? Suppose the Finder tells the vendor, "Hey, there's this bug in Passport Password Reset". Now the vendor (Microsoft) works in conjuction (collusion?) with the finder, and says, Look - this is a trade secret. Wait for a few months, and we'll watch if anyone's using this bug except us.
Poor Joe ServicePack is the one affected, and he figures nowhere in this scheme of things.
If you keep throwing chairs, one day you'll break windows....
Living in the D.C. Metro area, I was very upset when hearing that the D.C. Police Chief had been against revealing the make of the snipers' car when they finally found it out. Once this information was released, the snipers were caught in 2 hours or so, IIRC.
I agree with the parent poster - this seems like an apt analogy. At least if a non-negligible number of bugs, patches, fixes, or workarounds, even if just temporary, come from unexpected sources outside of the vendors or finders.
I welcome the day when we no longer have security bugs.
Look how this group defines a security vulnerability. According to this one, there are no security vulnerabilities.
For the purposes of this process, a security vulnerability is a flaw within a software system that can cause it to work contrary to its documented design and could be exploited to cause the system to violate its documented security policy.
There are a few publicly available systems with documented design and security policies, but it's explicitly stated that the security assumptions only hold when the system has cooperative users and is connect to a trusted network (better yet, no network at all).
Full disclosure only stimulates vendors to come up with a patch quickly. It's their fault in the first place there was an exploitable bug. We don't need a law to regulate bug disclosure. This is security through obscurity and it doesn't work. Information will leak and a limited number of people will take advantage of it anyway, while I, the use won't know to shut that buggy service down.
Let's call a spade a spade: companies like Microsoft should be liable for negligence when they sell buggy software for lots of money, at least for the amount of money the user paid for the software. Now, instead, not only do they want to be able to release buggy software with impunity, they also want to get free testing and bug fixes from users, and all that without the embarrassment and risk of having their bugs revealed.
There are precedents where doing this could be plain wrong, like in the last WebDAV vulnerability in IIS, that was discovered after being actively exploited by black hats. Hiding the facts will not stop crackers to exploit something they know first, and will only make victims unaware of what could or have happened with their sites.
The only time most of those guys release patches in a reasonable timeframe is when they're looking at a PR or a lawsuit debacle. If I find a vulnerability, I'll shoot the vendor an email and see if they hop to it or not. If they hop to it, great. If they blow it off, then it's posted publically before they patch.
That's how it'll always be handled by some people that find vulnerabilities. That's the way it's been handled in the past, and the world hasn't ended. It's just occasionally uncomfortable for a big player that sat on their ass for too long. Tough toenails. They need to put the same energy into just fixing reported problems that they do into trying to put this complicated mechanism in place to cover up problems that they fail to addres
obscurity will eventually always fail as a security protocol.
Being the cost-minimizing/profit-maximizing organizations that they are, they will always wait forever to incur new costs (like fixing bugs) until somebody "lights a fire under them." That's why quick public disclosures is important, because the bug *will* be discovered by the bad guys sooner or later, so it is better to fix it soon than wait for later.
Two problems :
:
... :-(
1) Exposing a vunerability will make any company that does not try to "repair" them liable -> costs money
2) Having to "repair" the vunerability because of the liability-problem will cost money too
Now look at the proposed "solution"
a) Don't tell anyone you know about he problem, and you can use "plausable deniability" (Just let your opponent *proove* you knew about it). Result : Nobody can effectivily sue you, and the threat of loosing money is gone.
b) The liability-threat (aka : losing money) is gone, and so is the incentive to develop a "repair" to the problem.
Guys & girls, this is a win-win situation !
For the companies that is
I disagree with you. What are you talking about here is the responsibily of the sys admin, not the vendor. The opensource community is responsive to security bugs and usually have a patch to correct the problem within a short time. However, it is the responsibility of the sys admin to apply the patch and it is his/her responsibility to monitor the security issues and respond accordingly.
The open source community is no better at getting patches deployed than the commercial vendors and is often worse.
HUH? You are comparing apples to oranges. So how many commercial vendors do you know that deploy and install patches on your systems?
Deployment responsibility is the end users wether using open sourse or closed source. The scope of this article and the parent poster are about bug fix acknowledgement and patch creation by software creators. How these patches get installed on your specific systems is another topic.
Bad boys rape our young girls but Violet gives willingly.
This was a tough call. Sure, the snipers were caught quickly afterwards at the truck stop: but the police didn't know that they were homeless vagrants who wouldn't know they'd been tagged. That was the call the police had to make, and it was not an easy one by any count.
I for one thought that while releasing the information proved to be the fortunate decision, it was not the smart one.
It's a lot easier to audit one C/C++ implementation, even if it is something gigantic like a compiler, than it is to audit an entire system's worth of code. In any case, it's the translation stage that provides safety -- when the language can guarantee safety, it translates directly to C/C++; when it cannot, it inserts runtime checks in the translation. The only way this would fail is if the translation stage was too aggressive in optimizing and failed to insert a runtime check where it should have. This doesn't happen very often though, and is in any case much better than C/C++, where the only runtime checks are those inserted by the programmer (who in large systems almost always leaves out checks that were necessary).
The other argument is often "well it's bad coders' fault, not C/C++'s fault." If that's the case, then all major software, including nearly all Linux security and server software, is coded by bad coders. There's been overflows in: Apache, Sendmail, BIND, Samba, OpenSSH, the Linux kernel, ProFTPd, WuFTPd, XFree86, and many more.
So basically, if you use C/C++, it's going to be insecure, no matter how much effort you put into it. Even OpenBSD, which spent an inordinate amount of effort auditing their code, has had a remote root exploit in their default install (and many more in stuff not installed by default but commonly used).
My preferred solution would be to use either a high-level safe language where feasible (SML, Java, Scheme, etc.), or to use a low-level safe language where not (Cyclone is probably the best of these).
10 PRINT CHR$(205.5+RND(1)); : GOTO 10