Hi @mzhukovin. I looked at this again a bit later, then decided that this might be 'fun' to look into further, just to see how accurate of results I could get by code. I quickly found out that this task was even more complicated than what I previously had in mind. At first, I was only inspecting faces that were created by fillet features, and were cylindrical in shape, but I saw that this was only highlighting a fraction of all the faces that were created by fillet features. Some of the faces were not cylindrical, such as at corner intersections between two or more fillet features. So, I had to greatly expand my 'inspection' code to 'deal with' all the other possible face shapes that I would commonly expect (Cylinder, Cone, EllipticalCylinder, EllipticalCone, Torus, Sphere). Once I included code for all of these, I was getting much better results. However, there are still apparently some face shapes that I did not cover there, because some of the more complex faces at complex intersections between multiple fillets did not seem to match any of those descriptions, and looked more like 'lofts'. Not super sure how to proceed to check those right now, and frankly do not really want to put that much more time into this project right now, since I do not really have an immediate use for this.
Attached is a text file containing 'a lot' of code that can be used as a single iLogic rule for this task. Right now it just highlights them by their category (concave, or convex), and shows a message while each group is highlighted, so you can pause to review them (for accuracy & functionality). However, the code within the latter half of the Sub Main area could be modified to do something else with them. I included an extra 'Class' block of code, and an extra 'Enum' block of code, in an attempt to shorten, and simplify the code needed within the Sub Main block of code as much as possible.
Whew.😅
Wesley Crihfield

(Not an Autodesk Employee)