(1) Using MP2 is never a good choice for transition metal complexes, it performs much worse than a properly chosen DFT functional, such as PBE0-D3(BJ). In addition, LOBA analysis is only applicable to single-determinant wavefunction, that means DFT (except for double-hybrid functionals) is absolutely the best choice.

(2) As mentioned in the prompt, you need to manually modify the Molden file. def2-TZVP for Mo is a pseudopotential basis set, you should open the Molden file via a text editor, and modify the atom index to actual effective atomic charge; alternatively, you should add [Nval] section to Molden file. Please carefully check the paragraph beginning with "One severe drawback of Molden input file" in Section 2.5 of manual.

# shrock
! MP2 RHF def2-TZVP FrozenCore TightSCF
! Pmodel AIM
%maxcore 16128
%PAL NPROC            3
end
* xyz            0           1
atom coordinates

Compound formula

Mo(CH2)(OtBu)2(NC8H10)

I get the message "Orca terminated normally" and the total charge was zero in the Orca listing file
When I try to use MultiWFN to do a LOBA calculation I get

!! Warning! Warning! Warning! Warning! Warning! Warning! Warning! Warning !!
The net charge of this system is quite large (   28)! Probably ECP is employed while you forgot to modify the atomic indices
in [atoms] field of the .molden file to actual nuclear charges, in this case some analysis results will be problematic!
Please check "Molden" part of Section 2.5 of Multiwfn manual to understand why and how to modify the file

The command I use to generate the Molden input file is

c:\ORCA\ORCA_2MKL jobname -molden

I had a look at the molden input file but am unable to follow the advics on modifying the atomic indices.