Species index in libraries

I converted a Chemkin file generaged by RMG to a new library. Because species in a Chemikin file had indexes, their indexes were reflected to labels of reactions in a new library of kinetics and thermo as shown below.

entry(
    index = 1,
    label = "NH2(46) + NH2(46) <=> N2H3(58) + H(8)",
    degeneracy = 1.0,
    kinetics = Chebyshev(
        coeffs = [
            [6.844, -7.326e-06, -1.245e-06, -1.424e-07],
            [3.05, 7.244e-06, 1.231e-06, 1.408e-07],
            [0.123, 5.578e-07, 9.478e-08, 1.084e-08],
            [0.04916, -4.475e-07, -7.603e-08, -8.695e-09],
            [0.01741, -2.781e-07, -4.724e-08, -5.403e-09],
            [0.006903, -6.777e-09, -1.153e-09, -1.321e-10],
        ],
        kunits = 'cm^3/(mol*s)',
        Tmin = (500, 'K'),
        Tmax = (3000, 'K'),
        Pmin = (0.493, 'atm'),
        Pmax = (1.974, 'atm'),
    ),
)

NH2 has an index of (46) in the library. Actually, it is just a part of the name of species. But, when I used the this library for a RMG job, RMG handles another species with the same index as well. In my case, NCNO has the same index of (46). A part of RMG.log are shown here.

3 network information
---------------------
Isomers:
    hydrazine(1)                                          86.2423 kJ/mol
Reactant channels:
    NCNO(46) + NCNO(46)                                   616.252 kJ/mol
    NH2(46) + NH2(46)                                     352.223 kJ/mol
Product channels:
    N2H3(58) + H(8)                                       425.149 kJ/mol
Path reactions:
    NCNO(46) + NCNO(46) <=> hydrazine(1)                   617.94 kJ/mol
    hydrazine(1) <=> NH2(46) + NH2(46)                     364.75 kJ/mol
    hydrazine(1) <=> N2H3(58) + H(8)                      425.149 kJ/mol

RMG mixes up two species with the same index. Probably, NH2(46) derived from the library should have a different index like NH2(46)(99). Otherwise, users should remove the indexes from the Chemkin files before converting to libraries. Could you solve this problem?

The issue names no source file, test, or entry point. Start by tracing Chemkin conversion and library species-label/index handling, then reproduce the NH2(46)/NCNO(46) collision from the supplied example. Done means distinct species remain distinguishable when the generated library is used by an RMG job.