Want CNET to notify you of price drops and the latest stories?

Was Cigital security warning too hasty?

Vulnerability experts question why the company publicized a minor security flaw in a Microsoft tool after giving the software giant only about 12 hours' notice.

Robert Lemos Staff Writer, CNET News.com
Robert Lemos
covers viruses, worms and other security threats.
Robert Lemos
4 min read
Security experts gave mixed reviews Thursday to the way in which a software-reliability company disclosed a bug in Microsoft's newest tools for building applications for its .Net framework and Windows operating system.

Late Wednesday, Dulles, Va.-based Cigital told The Wall Street Journal of a flaw in Microsoft's latest tools for creating Windows and .Net programs after giving the software giant a little more than 12 hours to respond.

Some security experts criticized the quick public announcement as irresponsible.

"There is no way that Microsoft could fix this in a day," said Al Huger, vice president of engineering for vulnerability-information company SecurityFocus. "Full disclosure has to be coupled with responsible disclosure."

The issue reopens a debate on how to responsibly disclose information about security vulnerabilities. Thoughts on disclosure range between two extremes: those who believe that every detail of a potential security threat should be publicized as soon as possible, and others who believe that no details of any security flaw should ever be published.

Mainstream security experts typically believe that the creator of a flawed piece of software should first be notified and, depending on the seriousness of the flaw, allowed a certain amount of time to create a patch to fix the problem.

On Wednesday, just hours after Microsoft announced its newest tools for creating .Net and Windows applications, security company Cigital revealed that the software giant's Visual C++.Net and Visual C++ version 7 had a flaw that effectively rendered a security feature ineffectual.

Gary McGraw, chief technology officer for Cigital, said the company followed the unwritten rules of responsible disclosure in the company's announcement.

"Our policy depends on the nature of the flaw," he said. "If it's something that's out there and leaves normal users open to a 'script kiddie' attack, much more time is required before disclosing the flaw." The security community uses the term "script kiddie" to describe online vandals who are not that technically adept.

In this case, however, McGraw said the tools were just announced, so it was more important to let developers know not to use the compromised feature.

The feature, known as the GS flag, is a software switch that can be turned on when a program is compiled. Any program built with the switch turned on has additional code that checks for a frequent security problem, known as a buffer overflow, whenever the program is running. However, because of the software bug, a malicious attacker can easily bypass the feature, McGraw said.

That means that while the problem doesn't make a program less secure, the feature promises much more than it actually delivers, he said.

"Is this a super-terrible flaw? Absolutely not," McGraw said. "It is a flaw in a feature that we are urging developers not to use because they will have a false sense of security."

That may not cut the mustard with Microsoft.

The software titan has been on the warpath about responsible disclosure since last summer. In November the company formed a yet-to-be-named organization to create a set of standards for releasing information about software vulnerabilities.

"We were just notified about this yesterday morning," a Microsoft representative said Thursday. "That raises issues about responsible reporting practices."

Coincidentally, the Redmond, Wash., company is in the middle of reviewing the entire Windows code base for security problems. The efforts come a month after company Chairman Bill Gates sent a memo to all employees urging them to put security and privacy first.

Yet, other security experts argue that in this case, Cigital is on safe ground.

While it would have been more prudent to deal with Microsoft and give the giant time to respond, notifying the public of the flaw was a reasonable solution, said Chris Wysopal, director of research and development for network protection company @Stake.

"The disclosure doesn't give the bad guys a leg up," he said. "I don't think it's putting people at risk when Cigital released this information."

Last year, Cigital had been considered as a potential reviewer to check Microsoft's .Net security technology for flaws, but it lost the competition. Some have speculated that Cigital publicized this flaw out of spite.

Cigital's McGraw took issue with the implications. "There is absolutely no truth in that whatsoever," he said. "We are very much convinced that we did the right thing and we did it in an honorable way."

In addition to the outside review of the .Net framework code, done by security company Foundstone, Microsoft spent a month in December reviewing the Visual Studio.Net tools for problems. They clearly have room for improvement, said McGraw, so developers should learn to count on only themselves to produce secure code.

"It is important for developers to really learn how to design things to be secure," he said, "not to rely on compiler magic to make security problems go away."