Comment (by Roland Haas):

A similar question was asked on the mailing list and responded to in http://lists.einsteintoolkit.org/pipermail/users/2021-April/007931.html with:

QuasiLocalMeasures regular spin measurement assumes that the black hole
spin is roughly aligned with the z direction. A spin near the equatorial
plane does not work well. In this case you'll have to use the coordinate
spins (qlm_coordspinx, qlm_coordspiny, qlm_coordspinz); these are less
accurate for distorted black holes, but will still work very well under
these circumstances.

so the Killing vector method was only ever designed to work with spins more or less along the z-direction and changing the method to compute the Killing vector will not help.

--
Ticket URL: https://bitbucket.org/einsteintoolkit/tickets/issues/2160/quasilocalmeasures-qlm_spin-doesnt-work