A quick question: do sky130 offer some kind of non quasi-stactic effect modelling on the MOSFET spice models?

Likely not, as this is rarely done. Two possibilities: (1) Split the L of the MOSFET into smaller L (i.e. split one MOSFET into multiples). However, this only works when L is not at Lmin. (2) Model the NQS by adding an extra resistor in the gate line with value Rnqs=1/(5*g_m) [see Tsividis, Operation and Modeling of the MOS Transistor, fig. 8.24]

I'm trying to design a cascode stage and the flow I'm currently using is too cumbersome. I'm using the RF cells, but in order to have accurate interconnection parasitics I have to manually erase it from the final layout and assign input and output ports to every node. I'm looking for quicker alternatives to have a rough modelling of the transistor response in mmW