funsec mailing list archives

Re: Autoupdaters are the best security tool since Diffie-Hellman...


From: Dan Kaminsky <dan () doxpara com>
Date: Mon, 24 Dec 2012 05:36:43 -0800

On Mon, Dec 24, 2012 at 3:54 AM, Jeffrey Walton <noloader () gmail com> wrote:

On Mon, Dec 24, 2012 at 5:49 AM, Dan Kaminsky <dan () doxpara com> wrote:
Remarkably tricky to do well, though.
Do it like Apple: perform your updates over HTTP. Make it a feature so
an organization trying to manage an non-organizational MacBook can
provide DNS and the Update Service. And don't sign the catalogs (TAR
balls fetched before the signed update). No problems ;)

What I can't understand: when it was applied against in-App purchases
(StoreKit), Apple cried foul.

http://z6mag.com/technology/apple/free-apps-for-ipad-iphone-security-flaw-in-ios-goes-unfixed-by-apple-1612248.html

It would be funny if it wasn't true: "Apple has now added a 'unique
identifier' field to receipts, and given developers tools so they
could verify digital receipts on their own server. However, this only
works if the developer runs the receipt through their server first.
Apps that connect directly to the Apple App Store server are still
vulnerable to the hack."

Instead of taking advantage of the pre-exisiting relationship between
the StoreKit API and Apple Servers by pinning the certificate (similar
to SSH's StrictHostKeyCheck), Apple pushed it on developers. Amazing.


Like I said:  Remarkably tricky to do well

Autoupdating third party apps is still an unsolved problem, save for the
web where you redownload the client every time (a *wildly successful*
approach, as it happens).  iOS's third party app updating is a hilariously
broken experience.
_______________________________________________
Fun and Misc security discussion for OT posts.
https://linuxbox.org/cgi-bin/mailman/listinfo/funsec
Note: funsec is a public and open mailing list.

Current thread: