Guinea fowl, African cattle (IIRC) and possibly antelope species like in leopard9's The Shade of Baobabs timeline could provide domestic animals for this civilization. I'm not sure about donkeys, though-don't they live in northeast Africa?
You're right about the donkeys. It looks like they didn't penetrate West Africa during prehistoric times, but that there's evidence of domestic ponies that were hardy and adapted to semi-desert environments. So the *Malians might get an earlier start with those, and ITTL they'd be draft animals as well as pack animals (they were too small to ride).
Regarding 'barbarian invasion', IMO in order to be successful, barbarian invasions will need to happen either at a time when the empire has fallen into chaos and cannot defend itself, when the barbarian bands become super-organized to take on the empire's armies (unlikely) or when the barbarians get mounts such as camels or horses that serve as force multipliers.
I was imagining that the barbarian invasions would occur during times equivalent to the intermediate periods of ancient Egypt, when the empire becomes feudal or collapses temporarily into warring states. Other than that, the empire would have the numerical advantage. Riding animals would be a game-changer for a while - the first incursion by mounted nomads might be *Mali's equivalent of the Hyksos invasion - but the *Malians would improve their fortifications soon enough and would start hiring nomad mercenaries themselves.
I'd expect that the barbarians will assimilate to *Malian civilization fairly quickly, given that they'd already be influenced by *Mali through ordinary trading and raiding. (Oh, and another thing that would be traded for from the arid regions: indigo.)
The interesting thing, though, is that there are two areas of the Niger where rice can be grown, with a long stretch in between where it can't. There could be a second rice-growing empire in the lower Niger, but it wouldn't be under *Malian rule, because it would be separated by a few hundred miles of territory in which *Malian civilization can't be supported. There would be *Malian influence here, but at a considerable remove, and this second empire could be *Malian culture's funhouse mirror.
The effects on the proto-Bantu depend on how much they come to rely on rice instead of yams, and what different environments these crops need. A Bantu migration that uses rice as its main crop is probably going to be pretty happy in the Congo basin where there's a lot of water, but once it moves into the more arid regions of the East African highlands we might see the migration slow down significantly from OTL's pace until new irrigation technologies are (re)discovered or the Bantu reconfigure the way of life and foods they're used to eating.
That sounds about right. The current conventional wisdom is that the Bantu originated in the Cameroon highlands, which wouldn't be suitable for rice agriculture, but if they get an early push into Gabon and the Congo basin, they'd be in a region where rice would do well. If they get the idea from the *Malians, or more likely from a daughter culture in the lower Niger, then rice cultivation could extend down the coast all the way to *Angola.
I wonder if this would create an early cultural split between rice cultivators along the Atlantic coast and Congo basin and yam-and-banana economies in the eastern and southern regions. There might not be nearly as many common cultural threads across the Bantu-speaking peoples ITTL as IOTL. The rice cultivators might also coalesce into kingdoms early on while the yam and banana peoples stay at a pre-state level.
I don't think it's a coincidence that after we evolved a grain/grass based agriculture, that we acquired grass/grazer based domesticates in the form of sheep, cattle, horses. [...] Microlivestocks seem to be relatively easy to domesticate - guineau pigs, turkey, chicken, hutia, geese, rabbits, etc. etc.
OK, so the *Malians would probably have goats, cattle, poultry, dogs and the above-mentioned ponies by the time they get to the city-state stage. Hmmm... pouched rats as microlivestock? Sand cats playing the role of early domestic cats? Grasshopper or even termite cultivation, although it would take a lot of the latter to make a meal?
With all this water control, I can't help but imagine lungfish farming or something of the sort occurring: in a similar way to cats/pigs, rice paddies will attract pests, and following them will be opportunists, like the lungfish, which is also an edible food source and transportable overland, unlike other fish.
The Chinese farmed ducks and fish in rice paddies, didn't they? The *Malians could end up doing the same.
Alcohol will become a bigger part of life in West Africa. A large(r) population with both lots of standing water (in the form of rice paddies and canals) and lots of destabilization (lots of invasion, not as much time to make/maintain sewers) means the Niger will become polluted with feces, and breed a Cholera like enteric virus. The dubious safety of the water will lead to a general gravitation towards a mijiu/sake-esque alcohol.
Palm wine and yam beer too, and on the lower Niger (albeit not in *Mali), East African-style banana beer. Maybe there will still be yam festivals ITTL even though rice is the staple grain, because yams are the source of a favorite alcoholic drink.
To add to your notes on religion, BTW, I'd expect that the *Malians would have a rice festival, and that rice would have many ritual uses.
As the populations grow, the zoonotics, especially crowd-borne ones will be common (think Plague of Athens, but across Western Africa), and just off the cuff estimates of the emergence of plagues in population centers, a plague could emerge as early as 1000 BCE, though certainly it's more likely to happen later, as crowd communicable diseases usually need a population of ~350,000 to 500,000 to sustain themselves. However, IOTL by 2000 BCE West Africa already saw some urbanization, and trade networks had been set up, so it's likely the population would be even higher with the introduction of rice, so *measles could be traveling around Africa as early as the 2nd millennium.
I guess there would also be more opportunities to catch parasitic diseases such as sleeping sickness and river blindness, although at least these wouldn't be pandemics. The *measles would be devastating at first, less so later, although mutated forms would probably sweep through *Mali every few generations.
Religion would probably be rather benevolent Gods ala Egypt, as there is a relatively reliable flooding of the Niger, but there maybe more violent warlike aspects to them with the nomadic invasions and large death toll associated with vector diseases. Somewhere between Semetic mythology and the mythology of the Nilo-Saharan speaking peoples would make sense. There may be primordial tales of a paradise lost due to folkloric memory of the Green Sahara (if we're assuming 3rd Millennium peoples, its likely they were refugees from the drying Sahara in the millennia prior), with little reference to the events which inspired the Flood myth of Western tradition.
The Sahara as the Garden of Eden? I like that. And I think you're right about the plague deities being different from OTL.
There would also be influence from the periodic barbarian conquerors, who would syncretize their gods with the *Malian pantheon as the Hyksos did with the Egyptian.