I am thinking that is the following a problem solved by nominal unification.
$\lambda a.X = \lambda b.\lambda c.c $
where we find $X$. The answer is obvious.
The reason is that it seems in nominal unification $X$ always will be an atom, not a function such as $\lambda c.c$. The theory seems works fine with this input, just did not saw any example similar to this one.