So, as someone said, there's no such thing as devolving, just a better fit for the environment. The idea that something is "more highly evolved" is a bugbear of biologists. Just ask one of them and they'll rant at you. Lots. Bacteria, for example, have more generations, so they've probably gone through more evolution than we ever will. We are more complex, but not more highly evolved than bacteria.
Environments can be full of changes that encourage complexity or not, and plenty of animals have gone "backwards" in complexity terms. Yet, sometimes, not all is lost. For example, fish versus aquatic mammals like dolphins and whales, in which mammals keep a lot of their complexity on the inside, while on the outside being pretty identical (functionally) to fish. Sometimes going backwards rom a more complex place can help you solve problems better.
This also leads into the phase space adjacent. So, the path through the space of all possible types of creatures (the phase space, in science), matters a lot to where you end up. Evolution can't really jump, it can only move towards phase spaces adjacent, so as long as you can trace an evolutionary path that makes sense (i.e. your species doesn't suffer from having something, but a while ago, it was vitally important), you'll never see a loss of that complexity. This point was made by things we have that are pointless but never got rid of, like our appendixes.
Also, sometimes useless or niche things turn into suddenly useful things. So, there's the belief that birds came to be because feathers were used for display and maybe insulation, while some dinosaurs learned to glide between trees. These bridged the gap between not flying and flying.
So, for your species, you probably want to have a bunch of things that have happened or are happening that would cause them to develop like you wish. For some traits that are historical (e.g. fish have vertical tails, aquatic mammals have flat tails), these will persist for a long time, since it really doesn't matter and the evolutionary difference is small, but other traits need constant selection. So, it's worth identifying any that are one or the other and, for those, making sure that the environment provides the constant stimulation they need.
I think with yours, most of the traits would be prone to constant change, with maybe an exception of polarisation and shells. Anyway, here's some stuff I thought might work as an example:
Anti-freeze and hibernation - any environment with extremes of temperature or water would give rise to this, so you're probably looking at a planet with some really weird variable stuff going on (high eccentricity, maybe, high inclination), to keep this level of complexity in the gene-pool for everything. It would be unlikely to have your species evolve from one of the more extreme environments, but if all animals are similarly disadvantaged, it's not a problem. (Also, the book Three Body Problem comes to mind, there's a planetary environment that's extreme enough to cause issues that force all life to hibernate)
Polarised light - As long as there's a situation where they developed the polarised light as a defense mechanism or hunting mechanism, this would also stay in the gene pool. So, polarised light is meant to be good at glare, so that might be a helpful mechanism in dealing with a planet with extremes of sunlight. Alternately, there may be something (e.g. dust, water vapour) which polarisation helps to cut out, or it was super-useful to their immediate ancestors.
Shells- if you have an extreme environment, these might still be quite useful. If there are intense radiation flares or sudden extreme rains due to temperature variations (e.g. sudden hail) then most animals would have shells for a very long time.
Size - if the extremes can also favour small animals and put a ceiling on size, you'd likely prevent human-scale.