Took a look at the file and I can see the problem and indeed it appears to be:
But that’s not actually the problem, to me at least. What I’m seeing is when you rotate IKCtrl_Foot.L around like that is the toes are doing what they should be doing, it’s the ankle bone that’s not moving/rotating as it should. The toe bone is pointing in the same direction as the foot controller is.
When I grab the primary foot controller and it’s pivot point is at the heel, like yours is, I expect to be able to move that controller around and the whole foot follows along. When I grab your foot controller and rotate it around to the side, I see the heel moving to the outside, away from the root of the foot controller bone. The toes are pointing the same direction as the foot controller is.
I went back to an old file to see if it was different, which it wasn’t. I suspect the problem may have always been there, but I never rotated the foot in this way, thus it went unnoticed.
Do your other rigs using this design suffer from the same problems?
Additionally, when the foot controller is pulled below it’s rest position, the toes of the mesh bend downwards. This is because of the ways the bones are moving. The ik constraint on the toes are targeting a bone that’s being pulled away from the foot, so it’s still pointing as it should, and bending the toes of the mesh while doing it’s job.
Rather than re-invent the wheel, here’s a good blender foot rigging tutorial:
It’s a bit older and a bit more complex than you’re rig, from what I have seen. (I did have to go hunting for hidden bones that were part of you’re set up, so I don’t know if there were even more hidden bones or not). It’s a good rig, but it does use something called an action constraint. I don’t like the action constraint, so I changed somethings around to get away from that constraint, and it’s my goto foot rig for a normal character.
I’m not on my main pc where that file is, but when I can, I will post what I use.