Handle Duplicates

Currently, our "Open" functionality just finds the first match against the bookmark key (ie the oldest entry that matches). This is _probably_ fine, if we were handling duplicates when adding them to the bookmark CSV.

Let's use this ticket to brainstorm how we want to handle duplicate entries when adding a new bookmark.

Start by tracing the Open functionality and the path that adds entries to the bookmark CSV. Review how duplicate bookmark keys are currently handled and decide, with maintainers, what behavior should be expected when adding a duplicate. Done means the duplicate-handling policy is agreed and its implementation scope is clear.

Written by the indexing model from the issue text.