On Sun, Dec 21, 2003 at 05:18:21PM -0000, Mark Peterson wrote: > If an Online Bank utilizes 3rd-party webservices (javascript/.JS) via > either web-analytic measurements or a banner-ad server - Is there not > indeed a theoretical backdoor to the client-side browser if this > 3rd-party webservice/webserver was compromised with malicious code? Any half-way reasonable web browser will strongly enforce a separation of content, javascript, etc between the bank and its third-party content suppliers. Whether or not any specific user's web browser is "reasonable" is a matter of implementation, and I rather suspect some Very Popular browsers are going to let the user down in pretty drastic ways. (I can confirm that at least mozilla 1.4 on linux appears to enforce this separation: MBNA has introduced a new 'throwaway creditcard number' service, ShopSafe, to allow users to generate new numbers per merchant. Yay! Finally! Of course, the server that serves the javascript references images served on a different server, and the mozilla javascript engine refuses to allow this.) So, while you've found a method to slip malicious content to clients that people rarely think about, at least the core problem is several years old at this point, so web browser authors have had more than ample opportunity to repair these problems -- and some have used that opportunity. Hopefully, the efforts of http-equiv, Liu Die Yu, Jelmer, 3APA3A, Guninksi, and other auditors[1], will help browser authors fix bugs before they are exploited by nefarious sorts. Thanks [1]: Sorry this list is not exhaustive -- I certainly don't mean to slight anyone by their absence from this list! -- The DMCA is anti-consumer. The RIAA and MPAA have no right to rewrite copyright laws to suit themselves.
Attachment:
pgp00450.pgp
Description: PGP signature
