The lists you already keep elsewhere now feed the report, and anything you have since filled in clears off it.
Your lists become gaps. Switch on any of the personal lists in the settings and each arrives as its own group on Discover: your IMDb watchlist or any public IMDb list, your Trakt watchlist, your JustWatch watchlist, your MDBList watchlist, your Discogs wantlist, your OpenLibrary "Want to Read", your favorite series on TheTVDB, and your TMDB watchlist and favorites once you connect your TMDB account (two clicks, nothing calling back, so your server needs no public address). An IMDb list of people is read as a filmography to fill in instead, which is what lets you follow a director or an actor you own nothing by yet.
Discover is sectioned. The lists you keep first, then the lists you pointed the plugin at, then what your own library suggested. A list that was read and holds nothing you are missing says so rather than not appearing at all, and one that could not be read says that instead.
Clear off what you have, without a rescan. A refresh control on every row and every heading checks your library right now and drops what you hold, "Clear what I have" does the whole filtered tab, and an acquired title leaves every tab it appeared on at once. What survives can be re-checked against its providers in the background, which also picks up what has been added to a collection, discography, or bibliography since the last scan. The TODO list checks the same way, and both Markdown exports check before writing.
Also: undated titles are treated as not yet out and badged Announced (music and books exempt); Diagnose reads an announced same-titled item as an earlier, different release rather than a misidentification; scanning and clearing read far less of your library, so a long scan no longer holds it in memory while it waits on providers; Trakt and MDBList lists read to the end instead of stopping at 100 and 1000; an acquisition URL carrying a username and password is redacted before it reaches the log.