This is a great set of assumptions to start off with. One thing that I would add is that it needs to include the term “aggregations” within its areas of focus, not just “items.” Most levels of archival description are not item-based, so that’ll be key in developing the Schema.org extension for architypes.
Here’s a typical example: http://archiveshub.ac.uk/search/summary.html?recid=gb186-bxb%2Fbxb%2F1%2F1%2Fclb
The description doesn’t get any deeper here, but it’s not describing a single item. It’s describing an aggregation of materials related to Brendan Cleary (in this case, it’s a boxful of documents). Additionally, this description is only really useful when it inherits certain information from its ancestors in the EAD source file (e.g. information this aggregation of documents is described as part of the “Published Poetry and Translations by Author” series, which is part of the Editorial subfonds, within the Bloodaxe Books Archive fonds, which is held by Newcastle University.) These sorts of relationships are equally as important as defining other relationships according to date, subject, creator, etc.