there is some literature explaining systematic algorithms for computing Feynman diagrams for scattering processes, but I cannot see why the calculations for such processes require a choice of convention (for example momentum and fermion flow) when there is no such conservation of fermion number required in all processes in the presence of a Majorana fermion.
For example, suppose that $\chi$ is a spin $\dfrac{1}{2}$ Majorana fermion, $\psi$ a Dirac or Majorana $\dfrac{1}{2}$ fermion and $\phi$ a spin one boson. Then both of the following processes should be calculable (and potentially non zero) in the presence of an interaction Lagrangian of Yukawa type, $-\mathcal{L}_Y = g\chi\psi\phi $
$\chi \longrightarrow \psi \phi, \quad \chi\longrightarrow \overline{\psi}\phi$
Here is the literature
https://cds.cern.ch/record/238903/files/th-6549-92.pdf,
https://www.sciencedirect.com/science/article/pii/037026939291045B,
https://arxiv.org/abs/0906.3388
Could someone shine some light on how say at tree level the way to compute such processes is via Feynman Rules? It is not intuitively obvious why the choice of conventions is at all necessary or sufficient considering they seem to do slightly different things in these texts.