While you browse data, entities that are rendered in the UI are checked for their schema. The schema is then used as a backbone for many features.
While querying entities using a Find Query, you create a list of conditions that results conform to. When choosing a field to filter on, there are two ways to provide this field.
Your first option is to write the field name manually. If a matching field is found in the schema, the initially assumed field type is used based on the schema. You can modify the suggested data type, if needed.
The second option is to use the 'Magnifying Glass' icon to the right of the field. Doing so presents you with the schema as MongoKitten implies it. For more information on the meaning of the icons, please read the documentation on Data Types.
Aggregates use the same filtering system as Find Queries. However, when adding certain stages, the schema changes. For example, project stages can move fields to different keys, or omit the value entirely!
These changes are represented in successive stages. When a projection is followed by a filter, the filter's suggested schema is affected by the changes applied by your projection stage.
When adding an entity, you'll initially be presented with a form containing all assumed values based on your schema. You can remove these fields by right-clicking (macOS) or long-pressing the row (iOS). This summons a menu, including an option to fill the field with a null value and removing the field entirely.