There's some kind of floating point error going on with the transform rotation, but nothing that could cause that large amount of thrust deviation. Furthermore, having a look at the config file, it doesn't have even have a gimbal. Points to some kind of unequal mass-distribution you have going on with that TKS craft there (Could you check the fuel levels in the radial tanks for me?), very confusing to say the least. I'll try and recreate this problem myself. Also, thank you for demonstrating the spoiler function, that will be useful :-) Quad-symmetry RCS tanks (100% full), quad-symmetry engines, dual-symmetry solar panels. The stock monopropellant engines have similarly slanted thrust.