Currently when reverse engineering the database or updating the model from the database we have 3 nodes - tables, views, functions. Tables show all the tables that could be discovered. This means that when reverse engineer a database created by code first we will always have the __MigrationsHistory table which - in most cases - users don't want/need to add to their models. We could try adding a separate node with system tables to not have people always uncheck system tables they most likely don't need.