Let $\phi:F\to G$ be a morphism of presheaves of rings over a topological space $X$. One can define $Ker(\phi)$ and $Coker(\phi)$.
Since $F,G$ are presheaves of rings, for each open $U\subset X$, $Ker(\phi)(U)$ is not a unital ring but rather a module of $F(U)$. For $Coker(\phi)(U)$, I have it as a quotient module of $G(U)$ which is a module over $F(U)$ at best. They are not rings?
What is the meaning of $Ker(\phi),Coker(\phi)$ in the case of morphism of presheave of rings here? If they are presheaves of abelian category objects, it would make sense for some time.
As stated in the comments, the target category is not the usual one of unital rings, but instead the category of (possibly) nonunital rings; i.e. associative $\mathbb{Z}$-algebras.
This category has a zero object, so the (co)kernel of a map is defined as the (co)equalizer of that map with the parallel zero map.
Limits of presheaves are given pointwise, so $\ker(\phi)(U)$ is the kernel of $F(U) \to G(U)$, and similarly for the cokernel.