The Search feature currently will include results from *-cache repositories. This is source of confusion for users as they will pin scripts and deployments to a URL that is returned from the search results. If the cache is deleted, then the URL may no longer work, and the scripts will fail.
We would prefer a way to limit searches to Virtual, or possibly Remote repositories. This way when a user finds a URL in the search results, it is less likely that the link will be become invalid due to flushed caches.