RE "Feasibility issues aside, would this contraption help delaying airflow separation at high AoA?":
According to a team that tested it, yes:
The major advantages of a Magnus effect device are high-lift
forces or rather high wing-loading and stall resistance [...]
Their design housed the treadmill in such a way that only the upper moving surface is exposed. According to their tables, they tested only up to 5° AOA. They acknowledge the complexity and added weight—and I'll add to that, reduced thrust and fuel capacity to power and house the system, respectively.
Side note: if the goal is not a novel STOL RC plane, one popular and tested method for high AOA is leading-edge root extensions (LERX).