ARCH-Comp Problem instance collection

I was wondering how existing ARCH benchmarks come into play reagarding the competition and the presentation of certain benchmarks in the respective groups. As others are maybe wondering as well, here's a more comprehensive task description (Thanks Goran):

The goal is, to provide (until Dec. 15) a set of *concrete* benchmark instances for each category. The instances can either be part of already existing ARCH benchmarks (a concrete instance) or new instances. In any way, the description (the model and the specification) should be precise and not contain any ambiguity. Ideally, the benchmark instance should be provided in SpaceEx format as a separate topic (along with the corresponding files) which then can be discussed. Each participant is allowed to provide at most 2 instances per category.

If I forgot something, please feel free to extend this post.

Best,
Stefan