Vulnerabilities / Threats
11/12/2009
01:20 PM
Connect Directly
Google+
LinkedIn
Twitter
RSS
E-Mail
50%
50%

Adobe Flash Security Model Permits Malware

Web sites that allow user-generated content could be risking attack due to Flash's permissive security model.

A security researcher at Foreground Security has identified a flaw in the way Web browsers handle Adobe Flash files that could be used to compromise Web sites that allow users to submit content.

The flaw involves a way to exploit the same-origin policy of Adobe Flash to upload and serve malicious files.

The same-origin policy restricts Flash objects so that they can only access content from their domain of origin. The vulnerability arises from the fact that Flash objects on your server will execute in the context of your domain.

"The short version of all this, of course, is that if I can convince a server to serve up a file on my behalf, I can use that file to attack the server," said Foreground Security researcher Mike Bailey in a blog post.

Bailey has posted screen shots demonstrating the vulnerability by uploading and executing Flash (SWF) files using cPanel's File Manager and the Squirrelmail Web e-mail client. In a video posted on YouTube, he demonstrates how he used the vulnerability to attack Gmail.

"[A]ctually exploiting this is extremely tricky, as there are a lot of hoops to jump through," he said in a blog post, referring to Gmail. "It required uploading the SWF to my own account, then logging the victim into that account (via CSRF), loading the SWF into the browser, logging them out, and enticing the user to log in while keeping the original page loaded (eg. in another browser tab). Not simple, and that's the simplified version, but it worked beautifully."

Google has addressed the issue by blocking the CSRF login vector and now appears to be fairly well protected against Bailey's attack. Bailey, however, has posted information about proof-of-concept attacks against Google's recently added CSRF protection measures.

"We used Gmail as an example because we know that it was fixed," said Mike Murray, CISO at Foreground, in a phone interview. "The problem with this vulnerability is that a whole lot of people are vulnerable. Millions of sites are vulnerable."

Unfortunately, there's no easy fix. Adobe could address the issue by making its content policy restrictive rather than permissive, but Murray claims the company doesn't want to do so. "Adobe is afraid of breaking all of their customers' Web sites," he said.

Adobe maintains that Flash is safe if used properly.

"By nature, Flash (SWF) content is powerful, active content and should be handled with the same care as other active content technologies, such as JavaScript, to ensure a site's design does not become vulnerable to abuse scenarios," an Adobe spokesperson said in an e-mail. "Adobe has always advised that allowing arbitrary uploads or attachments of Flash (SWF) content to trusted domains should not be performed due to potential abuse scenarios, such as the ones outlined by Mike Bailey."

Adobe has published security guidance for safely hosting Flash content, the spokesperson said.

Web site owners can mitigate the risk by serving all user-supplied content from a separate domain. According to Bailey, a number of major Web sites like Yahoo Mail, Hotmail, and Wikipedia already do this.

Bailey suggests that user-generated Flash (SWF) files can be served using a "content-Disposition: attachment" header to prevent them from executing when embedded in a Web page. This is a new addition to Flash 10.0.0.2.

"On the user side there's very little you can do, other than completely disabling Flash," said Bailey.

Update: Added Adobe comment.

How are you dealing with data-centric security? Answer our survey by Friday, Nov. 13, and be eligible to win an iPod Touch. Click here to take part.

Comment  | 
Print  | 
More Insights
Register for Dark Reading Newsletters
White Papers
Cartoon
Current Issue
Dark Reading, September 16, 2014
Malicious software is morphing to be more targeted, stealthy, and destructive. Are you prepared to stop it?
Flash Poll
Video
Slideshows
Twitter Feed
Dark Reading - Bug Report
Bug Report
Enterprise Vulnerabilities
From DHS/US-CERT's National Vulnerability Database
CVE-2014-2886
Published: 2014-09-18
GKSu 2.0.2, when sudo-mode is not enabled, uses " (double quote) characters in a gksu-run-helper argument, which allows attackers to execute arbitrary commands in certain situations involving an untrusted substring within this argument, as demonstrated by an untrusted filename encountered during ins...

CVE-2014-4352
Published: 2014-09-18
Address Book in Apple iOS before 8 relies on the hardware UID for its encryption key, which makes it easier for physically proximate attackers to obtain sensitive information by obtaining this UID.

CVE-2014-4353
Published: 2014-09-18
Race condition in iMessage in Apple iOS before 8 allows attackers to obtain sensitive information by leveraging the presence of an attachment after the deletion of its parent (1) iMessage or (2) MMS.

CVE-2014-4354
Published: 2014-09-18
Apple iOS before 8 enables Bluetooth during all upgrade actions, which makes it easier for remote attackers to bypass intended access restrictions via a Bluetooth session.

CVE-2014-4356
Published: 2014-09-18
Apple iOS before 8 does not follow the intended configuration setting for text-message preview on the lock screen, which allows physically proximate attackers to obtain sensitive information by reading this screen.

Best of the Web
Dark Reading Radio