The symbol rescuer had a potentially fatal flaw. It only rescued symbols that were different than the symbol in the original library. It did not rescue symbols could no longer be found in any library. This meant that you were always using the cached symbols in this case. This was a serious flaw. If the cache library ever got corrupted or deleted, all of the links to the symbols that no longer existed in any library would be broken. This is not the purpose of the cache library. It only should have ever been used as a temporary backup in case the user accidentally deleted a symbol in the schematic while editing. Unfortunately, it has been used a crutch since the beginning of the project albeit a crutch that was hidden from the user for the most part.
To answer your question, it makes perfect sense to rescue symbols that cannot found in the library list or now in the library table to prevent the situation describe above. This way if the cache ever gets corrupted or deleted, there should not be any broke symbol links (unless of course you delete symbol libraries that have links in the schematic) because they are always rescued to the rescue library. So your assumption about it being more dangerous is incorrect. It is now far less likely to have a broken symbol links.