What tends to not work well with Kerdi membrane are latex modified mortars. There are many thinset modifiers that do not need to dry to attain their strength, but latex modified ones do, and that's really a VERY long-term thing between the waterproof membrane and a nearly impervious tile.
Being composite, it could be mostly plastic, and that may be an issue with any thinset.
FWIW, Schluter doesn't prohibit a modified thinset on Kerdi, they prohibit one that uses latex. They actually manufacture a modified thinset they know will work with their membrane. A modified mortar can act as a combination of the cement bonding effect and a glue, which can make for a stronger bond. Keep in mind, though, that the actual bond created to a fleece covered membrane is because the thinset flows around the fibers, then cures...it does not stick like a glue. So, the actual bond to the membrane is required to be at least 50-psi, and if installed properly, Kerdi tends to test out with a 75-psi bond in shear. Typically, the actual bond to a ceramic tile (your panel doesn't seem to be ceramic) is usually at least 200-psi, and can be over 400-psi, so if it were to break free, it would be the fleece tearing off of the membrane that would occur before the tile comes off. A quality dryset would work with most any ceramic tile...not sure about that composite one.
I've been in some ruins around the Dead Sea...some mosaics on the floors were still intact and some had some glass elements...modified mortars are only about a 70-year old invention, and today's mortars are MUCH more consistent and strong than what was available nearly 2000-years ago.
I'd probably call Schluter and see what they say. My guess is that they would recommend their Allset product, but there are others that will work. Custom's Versabond works for many tasks over Kerdi. The consistency of the thinset is critical when working with a membrane like Kerdi...it needs to be fluid enough to flow around the fleece fibers, but not so runny that it drops off of the trowel. IT should be able to hold a notch when troweling, but just barely for easiest embedding of the membrane. Excessive moisture makes for a weak bond as it holds the particles of cement apart too long, and when it starts to cure, they may not fully interlock which is how it creates its bond. On an extreme situation, it can allow the components to stratify, so you get mostly the sand at the bottom and cement at the top, so it's not homogeneous, and thus weaker. The hassle with a latex modified mortar is that until the latex dries, it remains fluid and can crush the fragile cement crystals if there's any slight load, and that drying between impervious surfaces can literally take months on a larger tile.
If the back of those things is really smooth, it probably wouldn't hurt to rough it up a bit first. As I said, thinset works by growing crystalline structures that interlock...it does this on a microscopic level, but a really smooth, synthetic surface may not provide much 'tooth'.
The TCNA guidelines call for expansion accomodation between changes of plane and materials, otherwise, you can grout the joints. On a composite tile, you might have to be careful about scratching the surface with some sanded grouts, depending on your technique and the one you choose (some sanded grouts use more rounded bits than sharper sand, like you'd seen in sandpaper). And, if you want to avoid using something like silicone on those changes of plane, you could use an engineered joint - called a profile in the industry there, and never have to deal with caulk.
I think you posted over at
www.johnbridge.com, but if not, you might want to take this discussion over there as they specialize in tiling things.
