Most of that is accurate, but careful you don't drift into group selection, which (as I understand it) isn't well supported by the available evidence. To be selected for, a trait needs to differentially increase the likelihood that genes for that trait are passed on, so there needs to be a sufficiently close relationship that the person being benefited is likely to share the gene responsible for the help (or greater effectiveness of the help, or...). Of course, there is another leap from "X causes a trait to be selected for" to "we should individually be valuing X", but guard-of-terra seems to be thinking in those terms to begin with (and others have questioned that point).