Slashdot Mirror


How Web Apps Can Turn Browser Extensions Into Backdoors (threatpost.com)

"Threatpost has a link to some recent research about ways web pages can exploit browser extensions to steal information or write files," writes Slashdot reader jbmartin6. "Did we need another reason to be deeply suspicious of any browser extension? Not only do they spy on us for their makers, now other people can use them to spy on us as well. The academic paper is titled 'Empowering Web Applications with Browser Extensions' (PDF)." From the report: "An attacker [uses] a script that is present in a web application currently running in the user browser. The script either belongs to the web application or to a third party. The goal of the attacker is to interact with installed extensions, in order to access user sensitive information. It relies on extensions whose privileged capabilities can be exploited via an exchange of messages with scripts in the web application," researchers wrote. They added, "Even though content scripts, background pages and web applications run in separate execution contexts, they can establish communication channels to exchange messages with one another... APIs [are used] for sending and receiving (listening for) messages between the content scripts, background pages and web applications."

The researcher behind the paper focused on a specific class of web extension called "WebExtensions API," a cross-browser extensions system compatible with major browsers including Chrome, Firefox, Opera and Microsoft Edge. After analyzing 78,315 extensions that used the specific WebExtension API, it found 3,996 that were suspicious. While it seems voluminous, they noted that research found a small number of vulnerable extensions overall, and that concern should be measured. However, "browser vendors need to review extensions more rigorously, in particular take into consideration the use of message passing interfaces in extensions."

12 of 34 comments (clear)

  1. So by Anonymous Coward · · Score: 1

    So basically the whole "we're dropping XUL for webextensions. Because...uh...security!" thing from mozilla which crippled all my favourite extensions was pointless.

    1. Re:So by mukinrestak · · Score: 5, Insightful

      Yep. Anyone with a drop of decent cynicism knows that the goal is to cripple ad blocking and privacy protection and that "security" is just the excuse.

    2. Re:So by macraig · · Score: 1

      The goal for us is to cripple the cripplers. Having a browser-independent updated HTTPS-ready version of Proxomitron would really help.

    3. Re:So by h33t+l4x0r · · Score: 1

      So you think your free ad blocking extension is above suspicion? That's the dumbest thing I ever heard.

    4. Re:So by gl4ss · · Score: 1

      isn't it quite obvious that a web page can communicate with a browser extension though? like really fucking obvious? further than that you could just make the extension exploit the computer directly as long as you're coding it...

      --
      world was created 5 seconds before this post as it is.
    5. Re:So by Anubis+IV · · Score: 1

      And mine. The timing of this story seems awfully coincidental after Google's bad PR yesterday with the proposed Chromium changes.

    6. Re:So by thomn8r · · Score: 1

      Yep. Anyone with a drop of decent cynicism knows that the goal is to cripple ad blocking and privacy protection and that "security" is just the excuse.

      But think of the children!

  2. What's up with sandboxing? by fustakrakich · · Score: 1

    Everything is supposed to run that way anyway, right? This shouldn't be an issue.

    --
    “He’s not deformed, he’s just drunk!”
  3. Re:Free != Open Source by h33t+l4x0r · · Score: 1

    open source projects never get hacked or bought by malware companies, so you're right. don't be suspicious at all, ever.

  4. Strawman. He never said that. by Anonymous Coward · · Score: 1

    He said: "Trust but verify."

    In the latter case, open source is above suspicion, and in any case, it's more trustworthy than closed source, where you cannot even tell without a massive effort.

    QED. You're a moron and an asshole.

    1. Re:Strawman. He never said that. by DeVilla · · Score: 1

      OK. Case study students.

      First example, Microsoft Windows 10. Everybody knows it is snooping on users in ways that many do not approve of. It has been doing so since it's release and very little has changed. No one has come up with a reliable fix.

      Second example. Cisco routers have been found to have hard coded back door accounts built into them. This has happened several times. Pretty much any one with even a modest understanding of system security knows this to be poor practice. No one outside of Cisco has any way of knowing if there are any more back door accounts in Cisco's firmware nor do they have a way to know if a new back door gets introduced.

      Third example. Mozilla has in several instances added features to Firefox that snooped on users in ways that many did not approve of. People went into the source code and found it. People and system distributors made forks of Firefox that disabled or removed these features. Mozilla ended up providing ways to disable these features in every case and in many cases removed or disabled these features by default or provided an affirmative opt in.

      Final example for today. Canonical added a feature to Ubuntu search interface that snooped on users in ways that many did not approve of. People went into the source code and found it. Canonical immediately documented how to opt-out and disable the feature for themselves. Forks and downstream distributions of Ubuntu removed the feature.

      The lesson class is not that Open Source software is invulnerable. The lesson is that it is harder to hide and force undesirable and unsafe behavior into Open Source software compared to non-Open Source software. Open Source software is not perfect. It's merely better by distributing the possibility of finding and fixing problems to a wider group than just those who would benefit from not fixing.

  5. Software freedom is best defense against malware by jbn-o · · Score: 1

    The anonymous post that included "Trust but verify." was right; only software freedom gives us the best known defense against malware. I don't agree with post moderation but if a discussion forum will have such censorious distractions, posts like that deserve far more than 0 points.

    Your post, on the other hand, in which you claim that software freedom is "bullshit" ironically highlights how valuable software freedom is: separating functionality into components solves nothing if those components are implemented with non-free (user-subjugating, proprietary) software. Whatever value the separation purports to grant users is entirely lost by not respecting a user's software freedom. In fact we know that proprietary software is often malware.