I would do that as an assembly with a skeleton part in there with curves in the skeleton part driving the shape of the teeth (with clearance) and the helix. It LOOKS like the section is normal to the long axis, not the helical trajectory, though I may be wrong. You can sweep features with the section normal to the spine with the option: "Normal To Trajectory", or remain constant to a reference using the option: "Constant Normal Direction", in your case the flat end. You can do a straight spine for your trajectory and use trajpar in the section relations for the twist (simplest), or you can wrap a curve around a cylinder to get the helix and use that to control the twist as a second trajectory and again using the straight spine, I'd go with the first option since you don't need the variable twist rate or reversal the wrap could give you.
Here's something I did for fun, I got a true involute geartooth form without using a CS and equation, and then made a sort of "herringbone" gear.
Best of luck!
