I read somewhere long time ago that developments which give rise to a problem generally also create the conditions that allow solving the problem. But perhaps this is because desparacy of the victims enhances their inventiveness in finding a solution even under given conditions. In this case we might expect one of the social-regulating innovations of proto-humans to be some sort of taboo on murder, and homicide seems indeed to be traditionally seen as just as serious a faux-pas (so long as the victim belongs to the own clan/tribe/culture-sphere) as e.g. incest (in civilized or at least no-longer-"egalitarian" societies one could get clobbered to death for it, but in "egalitarian" communities one usually survived the punishment and even got more or less rehabilitated, I think). However, the invention of taboos perhaps belongs to a too advanced stage of development to have been effective at the time rampant homicide started becoming a potential survival threat.

An the other hand, when the casualty rates in internal feuds starts endangering survival, one would expect natural selection to lead to either subjective or objective checks to extent or instensity of the feuds. One episode described in an article on deception and manipulation among chimps (I think) suggests that objective checks could already have developed at a pre-human stage: A young fellow gets carried too far, teasing a child. The alpha male gets enraged and bears down on the misdemeanant, looking and sounding like he's going to break every bone in his body. The rest of the gang follows with similar intent. While running for his life, the fellow gets a bright idea: he suddenly freezes in the "watch out, I see an eagle" pose. The posse also stops short and starts looking in all directions, trying to sight the non-existent eagle. After a while they forget what they got all excited about.... So, an instinct to stop a fight before the weaker one got seriously hurt could have developed at a pre-human stage.

As for subjective checks, one example could be the wellknown masculine solidarity. This seems to develop from boyhood pal/chum relationships, which I like to see as a further support of my pet theory of "humanity" having grown out of childhood sibling hierarchies. This pal/chum relation doesn't seem to really prevent internal skirmishes, but modifies them into a kind of fitness-training for physique and spirit of the "braves". It permits the hotheads to fight things out without forgetting to see each other as companions upon whose physical fitness one must continue to be able to count on (like there's nothing like a good brawl to fasten bonds of friendship).

Another subjective check is perhaps that the slender lanky "tenderfoot" look of young men tends to provoke a patronizing response in elder men, and that youngters actually tend to solicit such patronization. This not only warrants that they get that final portion of tuition they need for becoming full-fledged members of the community. It also tends to prevent that the over-eager young upstarts immediately get crippled by the much stronger grownup males as they enter the free-for-all competition for the favour of the females. In "primitive" human society, an additional safeguard is introduced in that, after initiation, there is a special ceremony in which the lads and lasses get paired off in monogamous couples, so they don't get a chance to start a fight over brides in the first place. Apart from that, there seems to have been a tendency to frown upon marriage between persons of different generation in pre-civilized society, and in the non-upperclass part of pre-industrial civilized society, which might also reflect an early taboo to prevent youngsters from getting crippled.