These are both data-related issues. Probability tables are used for unresolved resonance region cross sections to account for energy self-shielding effects. These tables are generated by NJOY and should not have negative probabilities, but this does sometimes occur. OpenMC will zero out those values, but it is telling you that the data produced by NJOY may be suspect.
The second message about the maximum neutron energy is due to an erroneous evaluation of Be7 in ENDF/B-VII.1 that has an upper limit on energy of 8.1 MeV. For most problems, this actually isn’t something you should be that worried about because very few particles will have energies exceeding 8 MeV. If you were trying to simulate, say, a D-T fusion with neutrons being produced at 14 MeV, this would obviously be a bigger problem. If you’re really concerned about this, I would suggest removing Be7 from the depletion chain file, which will prevent it from getting inserted in the simulation.