Among the vce option oim vs bootstrap vs jacknife, which one should I go for to make my standard errors robust.

with xtlogit depVar indepVar i.Var1 i.Var2, fe vce () ---- for this regression "robust" option is not available in stata.