A (unital) $R$-module $M$ can also be viewed as given by a (unital) ring morphism $\rho:R\to \text{End}(M)$. This point of view can be extended to many other examples of "compatible actions".
The definition of an $\mathcal{O}_X$-module $F$ is a sheaf of abelian group $F$ equipped with actions $\mathcal{O}_X(U)\times F(U)\to F(U)$ on fibers that are compatible with restriction.
Mimicking the above point of view, I tried to give it an equivalent definition:
An $\mathcal{O}_X$-module $F$ is a sheaf of abelian groups with a morphism of sheafs of rings $\rho:\mathcal{O}_X\to \mathcal{Hom}(F,F)$ from $\mathcal{O}_X$ to the sheaf Hom.
Is it reasonable?