diff --git a/plugins/EXTRACTING-A-PLUGIN.md b/plugins/EXTRACTING-A-PLUGIN.md
index 03fbde06..ac3fde9f 100644
--- a/plugins/EXTRACTING-A-PLUGIN.md
+++ b/plugins/EXTRACTING-A-PLUGIN.md
@@ -116,8 +116,12 @@ A normal refresh is enough; the shell is `no-store`. When the log's last line ap
 ## Music specifically — READ THIS BEFORE STARTING
 
 **Music is not a bigger offscale. It does not fit the shape, and three of its parts have nowhere to go.**
-Mapped 2026-08-15. Decide what to do about each *before* moving a file, or the extraction strands
-half-done.
+Mapped 2026-08-15.
+
+**Decide these yourself. Do not stop to ask.** Every one of them has two defensible answers, the owner has
+said either is fine, and a corrected decision is cheap where a stalled extraction is not. Record what you
+chose and why — in the manifest, in a comment, in the commit message — and keep going. The failure mode to
+avoid is not a wrong call; it is a tree left half-moved with a question attached.
 
 ### What has no home yet
 
@@ -166,16 +170,27 @@ take `homedir()` as their root independently. So a granted member's favourites a
 `/stream` serves them the **owner's** files and cliamp spawns a process in the **owner's** home. The
 app-store entry promises a configurable `library` folder and **nothing reads it**.
 
-### Suggested order
+### How to decide the three, if you want a default
 
-**Do not** invent the visibility model during the move. Extract with the semantics it already has, or the
-two changes tangle and a bug in either is a bug in both.
+Each of these is "close the platform gap" or "leave the piece behind". **Closing the gap is better when
+you have the runway**, because every later plugin needs it too — but a music that works with cliamp left
+in the platform is worth far more than a perfect design that did not land.
 
-1. Decide the three homeless parts first — close the platform gaps, or leave those pieces behind and say
-   so explicitly in the manifest. Either is defensible; drifting is not.
-2. `assertCapabilityTotality` should be fixed **before** this one: the cliamp routes are the live example
-   of the exact drift it cannot currently see.
-3. Then the move, in the order above.
+| Part | Close the gap | Leave it behind |
+| --- | --- | --- |
+| websockets | wire `server.reload({ routes })` — already proven, never called — so a manifest can declare ws providers | leave `cliamp`/`cliamp-audio` and the relay in the platform, note it in the manifest |
+| global overlay | add one slot the shell renders from installed plugins | leave `MusicPlayerHost` in `DashboardLayout`, gated on the permission as it already is |
+| widget | let a plugin contribute `widgetRegistryMetas` | leave `widgets/MusicPlayer` where it is |
+
+Whatever you pick, the plugin must **install, enable, disable and uninstall cleanly** at the end. A piece
+left in the platform is a documented seam; a piece left dangling is a bug.
+
+`assertCapabilityTotality` is worth fixing on the way past if you touch the websocket path — the cliamp
+routes are the live example of the drift it cannot currently see (routes upgrading into commented-out
+handlers). Not required.
+
+**Do not** invent music's per-user visibility model during the move. Extract with the semantics it has, or
+the two changes tangle and a bug in either is a bug in both.
 
 ---