3) If Chrome is to be a platform, it's going to need a shell with a rich interface to manage multiple applications and allow them to interact with each other. It's also going to need an ecosystem and the people it needs in that ecosystem use OSX and Linux. Making it inflexible or "hiding" it are the exact wrong things to do.
Firefox Greasemonkey scripts are used by power users to make minor tweaks to specific sites. I don't see how that functionality will further Google's lofty plans for Chrome. This was a quick hack that nominatively fulfilled a very popular feature request.
I'm certainly not the one in charge of directing Google's vision but it would seem to logical to me that they are not interested in creating a new 'platform' in the sense that xul is a platform. As far as I can see, the purpose of Chrome is to promote web apps as a reasonable alternative to desktop apps.
2) Since Chrome relies heavily on process isolation, I would think that a lot of the architecture of the browser is tied to the platform. Even the html rendering pipeline (based on WebKit) runs out-of-band and the rendered bitmap is send to the browser window via IPC. (I don't work on the team but from my understanding of someone's comments, this is how it roughly works)
3) Regarding extensibility, I agree that I doubt Google is going to provide much extensibility of the actual browser chrome. Google's interests are in the web sites themselves and as such, the greasemonkey-like extensions are what they're probably going to promote. They want to hide the browser container as much as possible and move the focus to the web page inside it.