Thanks for the insight!
Some of those points you make are really good and, in fact, i'm continuing to evolve the design to solve those flaws. That said, i wouldn't consider all of them too problematic. I'll analyse your observations in order:
I tried tenths of ascent paths and for some reason that one gave me the best results. It really does work, but you risk overheating and really have to be careful. Proof it works: https://www.youtube.com/watch?v=5-MMMswoYCU
The engines can generate electric charge on the ascent, and I don't think power is too important during descent.
The heat shield can be detached consistently by turning on the engines for an instant. It uses about 3 fuel units but saves the weight of a decoupler.
This craft was meant to meet the requirements for this challenge and weight the least possible. That means stuff like the above and being able to orbit wasn't a priority. That said, I'm working on a much more usable version right now.
True, and this is specially a problem at higher speeds/lower altitudes. It worked for me being extremely gentle and even then it wobbled. Part of the reason that is a thing is that I'm kind of obsessed with elegant designs and didn't want to move the radial decouplers inside the main tank. Again, the new design is more utilitarian.
I don't get the parachute thing . The non-water-landing version I had planned worked with two parachutes (albeit at low altitudes), and this one craft was able to land on those 7500m mountains that Eve has where the terminal velocity is about 15m/s with the parachutes deployed.
Again, thanks for the observations!