This is probably wishful thinking but here goes...
Now obviously I don't expect some magical app to crawl a website, summarise each page and detail all the assets within each page (if only!) however is there a crawler of some sort that will trawl a given website and produce a full set of links and respective hierarchy would be invaluable. That would ensure that no pages get missed out as they invariably would through a purely manual inventory.
From there I would be able to see the existing site hierarchies and taxonomy decisions that have been made when building the site, and could use this as the basis for creating a detailed website content inventory.
Whether this be an online application, script or even a useful manual technique other than just manually sitting at the site and following links would be great.