I’m searching around, and I can’t find a definitive answer.
(1) Are you able to save OmniOutliner 4 (.oo3) files and Scrivener (.scriv) files in DTPO without corrupting the files?
(2) Afterwards, are you able to open, edit, then save them right from DTPO or is this is bad practice since .oo3 and .scriv files contain Package Contents?
Did you read something that told you that DEVONthink dissects or damages packages?
I’ve stored .scriv, .oo3, .rtfd, and other similar packages in DEVONthink for years with nothing gone wrong. Of course, you cannot use DEVONthink to edit .scriv or .oo3 – or most file types, for that matter. But their previews work as expected, and they “Open With” Scrivener or OmniOutliner, respectively, without a problem.
But, why trust the forum for the answer? Like many things, you can test this for yourself and see if the results are what you expect.
I didn’t find an answer one way or the other. The nice folks at Scrivener said to use File > Index, so I started searching for best practices when dealing with files that contain package contents.
Here is a response from Scrivener Support:
It would be fine to Index your Scrivener files within a DTP database. (File
-> Index in DTP.) I’ve done that myself. I wouldn’t recommend allowing DTP
to put Scrivener projects into its own database hierarchy, though. And
allowing DTP to edit the contents of Scrivener projects would be very very bad.
I’m making sure I’m taking the right steps before moving a bunch of these filetypes to DTPO. I appreciate your response about your experiences dealing with these types of files in DTP. 
It’s easy to test what might happen with a copy of a file.
Whether you Import or Index your Scrivener file into DEVONthink, it might be possible to edit the file as plain or rich text (or possibly something else), with results that you wouldn’t like. For example, you could select a Scrivener file that had been Indexed, Control-click on it and choose the Open With… command, then select an application capable of opening/editing the file other than Scrivener. (You would decide not to do that on your working Scrivener files.)
If you Import or Index a Scrivener file and wish to edit it, you should get in the habit of pressing Shift-Command-O to open and edit it under Scrivener, or click on the “Open Externally” icon in the Toolbar to open the file under its parent application as designated in the Finder (if its a Scrivener file, the Toolbar icon will be labeled “Open With Scrivener”).
What is the difference between Import and Index? Is there an advantage?
Import (File > Import > Files & Folders) copies selected items in the Finder to a DEVONthink database, and the copies are stored within the database.
Index (File > Index) indexes the text content of selected items in the Finder and creates metadata about them in the database, but the files themselves remain external to the database. DEVONthink retains the Paths of the files so that they can be displayed, edited or printed.
You will find numerous discussions of these capture methods on the user forum. Whether one is preferable to the other depends on the user’s workflows and personal preferences.
If you need to share files with another database, such as a citation manager, Index would allow that. If you share files on a server with other users, including Windows users, Index would allow that. Those are critical needs for many users, and so would guide their choice of capture mode.
My own preference is Import capture. I like to be able to dive into a database and reorganize it, without worrying about the limitations that would be imposed in an Indexed database, which should stick to the organization of the captured folders and files in the Finder. I like the fact that a Database Archive backup is a complete backup of my Imported (self-contained) databases.
Your choice of capture modes is up to you. DEVONthink accommodates either mode.