There's the same concentration of OH- and H2S at the beginning, so I can make it all HS- and start from there. There must be a way with at most second degree. I know about Cardano, but they tell us all exercises we get are solvable with second degree equations.
I think I might have found a solution: I start from HS- and use the reactions in both directions (H3O+ and OH-). This gives m two constants that are 1/pka1 and pka2/kap (kap meaning autoprotolysis constant)
Then I use c1/c2 and c1*c2 which gives me two equations. I multiply the first with kap again which makes it second degree and the other equation can easily give one variable in dependence of the other, which I can use to solve the second degree equation.