subprocess with too many photons

Dear Madgraph experts,
we tried to generate events for e nu gamma gamma final states at NLO in madgraph.

This happens in the scope of event generation for the ATLAS experiment and therefore is wrapped in the gen_tf command, however, the issue seems to be with one of the madgraph subprocesses itself.
Just for completeness, we always setup asetup AthGeneration,23.6.3 in a fresh shell and start in a new folder (here run_3kNLO) where the attached run.sh, with the Gen_tf command, is situated as well as the JO folder with only the attached JO in it.

The entire folder (after execution with all logs) as a copy can be found here: https://cernbox.cern.ch/s/u9yjszCMA7g8vTZ .

The code runs rather far generates the Feynman diagrams and checks for poles, however, then exits with short ERROR STOP 1 messages and later shows that an index, regarding the jobs and events the code is processing, is out of range(line 1505 and 1548 in log.generate). We could find similar errors online, however, were unable to debug it in our case as they seem to be of different origin.
The point where madgraphs probably stops is one of the subprocesses (in PROCNLO_loop_sm-no_b_mass/SubProcesses/P0_dxu_veepaa) where it raises the ERROR ,too many photons 2.

We are happy for every hint you can give us!
Thank you very much,
Cheers,
Thomas