Laws of logic are propositions, and more precisely, propositions that are always true ( true in all interpretations).
For example : [ (A OR B) & ~B ] --> A
Rules of logic are " commands" , " imperatives" ; as such they are not propositions ( declarative sentenses) for they are neither true nor false; for example :
" from (A OR B) and ~B , infer A".
Is this account correct?
For example, is it helpful to distinguish the "law of modus ponens" and the "modus ponens rule"?
Are there logical languages / systems for which this distinction does not hold?
To the best of my knowledge, neither "law of logic" nor "rule of logic" has a specific technical meaning - at least in the context of mathematical logic (I don't know about the philosophy side of things). That said, the difference between (in a given system) axioms, theorems, and tautologies on the one hand and inference rules on the other is extremely important in proof theory; see e.g. this Mathoverflow question.
Logic is broad enough that there isn't a single "most general" situation, but the following is a pretty good summary of a wide range of cases:
We have a set $S$ of things called sentences, and a deduction relation $\vdash$ on $S$. That is, $\vdash$ is a subset of $\mathcal{P}(S)\times S$, with "$(X,y)\in$ $\vdash$" being interpreted as "from $X$, infer $y$," and abbreviated by "$X\vdash y$." (Usually $S$ is more than just a set, and in particular $S$ is often a free algebra over some signature whose elements are our logical connectives.)
The tautologies of the deductive system $(S,\vdash)$ are the sentences $y$ such that $\emptyset\vdash y$. For a given set $\Gamma\subseteq S$, the theorems of $\Gamma$ are the sentences $y$ such that $\Gamma\vdash y$, and when focusing on a single such $\Gamma$ we call elements of $\Gamma$ axioms.
An inference rule is a pair $(\Gamma,y)$ with $\Gamma\subseteq S$ and $y\in S$ (or a set of such pairs closed under appropriate substitutions). Generally $\vdash$ is "generated by" a collection of inference rules in the obvious way, and we can talk about a given inference rule being admissible (with respect to $\vdash$).
The relevant topic is algebraic logic, and I strongly recommend this paper of Blok and Pigozzi as a good introduction.