All the primitive features would be difficult to expose through the API and are relatively worthless in real-world modeling. Let's look at the box feature as an example. When you run the command interactively, it asks you to select a construction plane or planar face. You then provide two points to define the base corners and then define the height of the box. It's creating a temporary sketch and then extruding it. For the API to be equivalent, you would need to define a plane and then two points that lie on that plane, and finally an offset and direction. Through the API, it would be much better to be able to define a box anywhere in space without being limited to an existing plane.
Also, with the box feature, you can't accurately define the position of the original two points, so you can't control the position of the box. And once it's created the only way to edit it is to change the parameters that control its length, width, and height. There isn't any way to control its position. For real modeling, you'll want to create a sketch and then the appropriate feature. You can fully control the position of sketch geometry using both geometric and dimensional constraints.
---------------------------------------------------------------
Brian EkinsInventor and Fusion 360 API Expert
Website/Blog:
https://EkinsSolutions.com