Microsoft Baseline Security Analyzer exploit
Posted by Steven Parker on 29 April 2002 - 12:54 · 8 comments & 153 views
About the Author
Full name: Steven Parker
Forum name: Neobond
Contact: via forum PM
Submissions: Normal · Headline
Full name: Steven Parker
Forum name: Neobond
Contact: via forum PM
Submissions: Normal · Headline
Slashdot It!
Submit to reddit
Submit to blinklist
Bookmark on del.icio.us
Add to furl
Share on Facebook
Add to Windows Live
Add to Google
After MBSA analyzes the system for security vulnerabilities, a report is created as a plain text file that includes sensitive information that can be used by hackers to attack the specific machine. MBSA was created to help users become aware of risks and available patches. However, MBSA turns the simple vulnerability of reading local files into a much more powerful vulnerability. Such a simple vulnerability allows potential hackers to find out about vulnerabilities that enable full control over the machine that is under attack. These are automatic attacks.
This means that active content (executables, scripts, ActiveX, Java, etc.) has the ability to generate a list of vulnerabilities or read a previously created list, and can then utilize these vulnerabilities to its advantage. Even if this report can be accessed only by a specific user, the active content can access it too.
Technical overview:
MBSA creates a report that includes sensitive information that can be used by hackers and save them research time. The report can be accessed by a malicious active content:
1. The report is written to a known folder, e.g., C:Documents and SettingsusernameSecurityScans. The user cannot change this location.
2. The XML report is written in plain text and can be used by hackers to find the machine's vulnerabilities.
unlike previous exploits that went unanswered this one saw a prompt response:
Vendor response:
Finjan Malicious Code Research Center had an interesting discussion with Microsoft Security Response Center. This is their response:
Hi, Thanks very much for your note and for sending this on. We really appreciate it. To understand the issue fully, it would be good to expand this somewhat. There really are two issues here: One related to the ability to mount an attack successfully, and one related to how data is stored on a system and what could happen to that in light of a successful attack. To be clear, none of the attack scenarios that you've described are mounted through MBSA itself. Also, the attack you've described does not exploit a vulnerability in any product: in a default system this attack fails. It's only when a user chooses to run code from an untrusted source and proceed despite the security warnings provided that this attack could succeed. Protecting systems against untrusted code is vitally important, and we call this out in our 10 Immutable Laws of Security as Law #1 ( http://www.microsoft.com/technet/columns/security/10imlaws.asp ), to underscore its importance. If an attacker were able to convince a user to run their code, that code would then be able to take any actions on the system that the user can take. While it is true that MBSA stores its information in a known location, storing it in an unpredictable location would not measurably change the situation. An attacker's code could just as easily search the local system for the file. Likewise, it is true that MBSA's information can be read by the user (or code running as the user). However, even if the MBSA information were not present on the system, code running as the user would be able to determine the presence or absence of patches, simply by consulting the time/date information contained in the publicly available MSSecure XML database. Again, it is a question of degree rather than feasibility. The larger issue in both cases is the presence of code running with the user's privileges. If the attacker cannot run code, it does not matter how the MBSA data is stored, because the attacker cannot access it. If the attacker can run code, he or she does not need the MBSA data, as they already have all the privileges needed to duplicate the MBSA processing. (For that matter, the attacker could simply run MBSA itself and do a "screen scrape"). That said, we are always looking to make improvements and we appreciate concerns and feedback like this. Our MBSA team is looking at these suggestions along with others that we have received and consider them as they design future versions of this tool.
Thanks again for sending this on, we really appreciate it.
Regards,
secure@microsoft.com