Currently, when Initialize Mesh Reproduction Sprite module is used for skeletal meshes that have sections affected by cloth simulation, the mesh tri coordinate values generated from this module do not reflect the real position of said simulated mesh sections. Instead, the location of tri coordinates is taken as if there is no cloth simulation. This goes the same no matter if the module is used for spawning and/or updating particle/sprite location. Either the calculation should be updated to take into account the cloth simulated sections, or there could be an option to exclude geometry of skeletal mesh affected by cloth simulation.
As it is right now, skeletal meshes with cloth, can’t use these modules without some poor results, as particles spawn in undesired locations, ignoring the actual location of cloth simulated sections.
Edit: I’ve attempted to get the desired results using sampling regions within skeletal mesh. Using a skeletal mesh that has cloth simulation on, and switching the emitter to use CPU, does make the particles spawn on the actual locations of the sections of a particular mesh which simulates cloth (local space must be enabled and in this case the particles will spawn only on cloth simulated sections, regardless of simulated regions). If the CPU emitter has local space enabled on, then the particles will appear everywhere (or on sampled regions if stated) except on the cloth simulated sections. Cloth simulated sections in this case would render particles at 0,0,0 coordinates, which means some particles would appear on skeletal mesh and the others on said coordinates, The best of both worlds would be if GPU could somehow use these locations for particles, where cloth simulated sections are ignored, and leave us to make a dedicated CPU emitter that would spawn particles on cloth simulated sections.