I came across this plugin in a support case, where it was causing lots of problems in a large instance of Confluence.
The main problem with the implementation is that the PageSearchHelper isn't using the search index. In some cases, it can retrieve an insane amount of data from the database in order to answer its queries.
This class needs to be rewritten to use the SmartListManager and ListQuerys to work efficiently in large instances. We have a ContentPermissionsFilter which can be used to restrict access to search results efficiently. See the SearchQueryBean in Confluence for a sample implementation, and how we do our queries effectively for labels, pages, blogs, etc.