What you are describing is one form of review article. Although review articles more typically describe a more general state of the art on a problem, they can also be used to tie together and summarize a collection of linked papers into one coherent entry point.
Doing this well will require creating quite a lot of new content---it just won't be new technical results. Rather, the content created in such a paper is the distillation of a much larger body of work into a single coherent picture. This can be quite valuable for readers, because trying to reconstruct the picture of a body of work that is evolving over time and scattered across papers can be painful and difficult; additional detail about particular points can then be obtained by following the citations to the source articles. Note also that if you write this article well there will be little risk of self-plagiarism, because you'll need to rewrite everything pretty much from scratch to fit into the new and more compressed arrangement of ideas.
So, in summary: it's not only ethical, it is legitimate new work and can be highly valuable, just so long as you make the nature of the article extremely clear and include all of the relevant citations.