Submitted via IRC for Bytram
Websites Can Exploit Browser Extensions to Steal User Data
While web applications are bound by the Same Origin Policy (SOP) and cannot access data from other web applications unless mechanisms such as Cross-Origin Resource Sharing (CORS) are implemented into both, browser extensions are not subject to the same rule, meaning they can read and write data on web applications.
The extensions also have access to a broad range of sensitive user information, including browsing history, bookmarks, credentials (cookies) and list of installed extensions, and can download files and store them on the user’s device.
Browser extensions and web applications are executed in separate contexts, but they can interact by exchanging messages, regardless of the browser. This allows web applications to exploit extension privileged capabilities and steal sensitive user information, Dolière Francis Somé from the Université Côte d'Azur, Inria, France, says in a research paper (PDF).
The researcher analyzed the communication interfaces exposed to web applications by Chrome, Firefox, and Opera browser extensions and discovered that many of them can be exploited for access to privileged capabilities.
“Through extensions’ APIs, web applications can bypass SOP and access user data on any other web application,” Somé explains.
“Our results demonstrate that the communications between browser extensions and web applications pose serious security and privacy threats to browsers, web applications and more importantly to users,” the researcher continues.
(Score: 2) by NotSanguine on Tuesday January 22 2019, @10:14AM
A good point. I did some cursory poking around and discovered that determining the total number of extensions available for each browser (not counting extensions that aren't available through the relevant cetralized add-ons/extensions sites for each browser -- HTTPS Everywhere for example), would be a time consuming task. As such, it's likely that the number of vulnerable extensions is higher (perhaps significantly higher) than the number of vulnerable extensions found in the study.
Interestingly, Figure 3 [inria.fr] details the distribution of users per extension. More than half of the extensions tested have fewer than 1000 users.
It's rather annoying that the specific extensions found to be vulnerable were not detailed. It's unclear why that is. I could hazard a few guesses, but I'll refrain for the moment.
That was the primary point I was making, given Google's (non)response.
No, no, you're not thinking; you're just being logical. --Niels Bohr