I want to tag InlineShapes with a unique identifier. One way to do this is to convert the InlineShape to a Shape, set the Name property on the Shape, and then convert back to an InlineShape. Then to get the key associated with an InlineShape, repeat the steps: convert to a Shape, get the Name, and convert back to InlineShape.
My question is how much vulnerability does this introduce? My sense is that there's a risk of an InlineShape with complex formatting, situated in a particular way inside of a table or frame, being reformatted or repositioned during this back and forth conversion.