Hi, I'm trying to implement a one-way isolation scheme for proxied content in Firefox. The problem we're addressing is when content aggregators put proxied or imported scripts inside a page, but don't really want to give the third-party JavaScript/CSS access to the surrounding page contents. Our idea is to let content aggregators specify isolated content with a <div type="untrusted"> tag, and then the browser mediates their relationship so that the aggregator's content can access the div innards BUT the inner div contents cannot access the outer page.
We're prototyping this functionality using browser-side script rewriting in a Firefox extension. I would like to try to actually implement it in Firefox, however. Firefox already has one-way isolation schemes built in (since XUL JavaScript can access page contents but not vice versa) and I'd like to take advantage of that. However, I have no idea how to get started. How does this privilege checking work? Where is it being done? I'd really appreciate it if anyone could point me in the right direction. A workshop paper describing the idea in more detail is available at http://www.cs.virginia.edu/felt/secure_mashups.pdf , if you're interested. Thank you, Adrienne Felt _______________________________________________ dev-security mailing list [email protected] https://lists.mozilla.org/listinfo/dev-security
