docs: update docs, add backlog tasks and change notes

Update configuration, immersion tracking, and mining workflow docs.
Add backlog tasks for upcoming work items and change notes for recent
features and fixes.

changes/2026-03-17-background-stats-server.md

@@ -0,0 +1,6 @@
+type: added
+area: stats
+
+- Added `subminer stats -b` to start or reuse a dedicated background stats server without blocking normal SubMiner instances.
+- Added `subminer stats -s` to stop the dedicated background stats server without closing browser tabs.
+- Stats server startup now reuses a running background stats daemon instead of trying to bind a second local server in another SubMiner instance.

changes/2026-03-17-session-known-word-timeline-fix.md

@@ -0,0 +1,5 @@
+type: fixed
+area: stats
+
+- Fixed session stats so known-word counts track real known-word occurrences without collapsing subtitle-line gaps.
+- Fixed session word totals in session-facing stats views to prefer token counts when available, preventing known words from exceeding total words in the session chart.

changes/2026-03-17-stats-library-tab-label.md

@@ -0,0 +1,4 @@
+type: changed
+area: immersion
+
+- Renamed the stats dashboard's Anime tab to Library so the media browser label matches non-anime sources like YouTube and other yt-dlp-backed content.

changes/2026-03-17-vocabulary-tab-hook-order-fix.md

@@ -0,0 +1,4 @@
+type: fixed
+area: stats
+
+- Fixed the stats Vocabulary tab blank-screen regression caused by a hook-order crash after vocabulary data finished loading.

changes/2026-03-17-watch-threshold-centralization.md

@@ -0,0 +1,5 @@
+type: changed
+area: anilist
+
+- Standardized episode completion threshold by introducing `DEFAULT_MIN_WATCH_RATIO` and using it for both local watched state transitions and AniList post-watch progress updates.
+- Episode auto-marking now uses the same threshold as AniList (`85%`), removing divergent completion behavior.

changes/2026-03-18-remove-session-new-words-series.md

@@ -0,0 +1,4 @@
+type: fixed
+area: stats
+
+- Removed the misleading `New words` series from expanded session charts; session detail now shows only the real total-word and known-word lines.