The scores were only readable in the interface, which made the MCP server good at reporting ranks and useless at answering the question that follows: what should I change?
Five new endpoints expose ASO Health and its axes, the score history of an app, and the metadata simulator, and the MCP server exposes the matching tools. An assistant can now pull your current listing, read the axis holding it back, simulate a rewrite, and tell you what the new score would be, without you opening a browser.
The AI writer is deliberately not exposed. If you are already talking to a model, you do not need ours to write the draft, you need ours to score it.