chore(comments): drop epic, sprint and review-round ids from the code

The code cited its plan instead of saying why: `E28 review J4`, `(R6 S3b)`, `S1 justified…`,
`deep-review round 5`. That plan lives in the private ops repo, so for every reader of this one the
citation was noise, and a sentence whose subject was the plan step ("round 2 showed…") hid its
lesson behind an id. Each one is gone and the "why" kept: a bare citation is deleted, a sentence
carrying a lesson is reworded around it ("a later attempt showed…").

Measured (`scripts/audit_metrics.py`, product code): 652 lines in 270 files carried an epic code by
the definition the next commit holds (507 / 241 by the old, two-digit-only one) -> 0. Across all
code, tests, scripts and configs: ~1100 epic-code lines plus ~320 sprint/round ids.

Behaviour unchanged, checked mechanically: every changed .py file has the same AST once docstrings
are blanked; css/sh/toml/yml/json files are identical once comments are stripped; the only
non-comment changes in TS/JS are test titles, one description string in a contract table, and five
ESLint messages, whose text is all the gates snapshot diff holds. Not touched: Markdown (docs and
the changelog keep their history), Alembic migrations, TV episode notation (`S2 · E5`), DESIGN.md
ids (`C1`, `B4`, `S3h`), tracker ids (`SEC-16`, `DX-17`).

One stale claim fixed on the way: test_hls_reaper's docstring said the player checkpoints "every
10 s whether or not it is playing", which plex/stream.py calls wrong; it now says what the code does.

frontend/src/components/PlexPlaylistReorder.dom.test.tsx

@@ -2,10 +2,10 @@
 //
 // One accepted write, one cache entry that matches it — the Plex half.
 //
-// `Playlists.tsx` (the YouTube list) got this rule in round 5: the detail cache is what a rollback
+// `Playlists.tsx` (the YouTube list) got this rule first: the detail cache is what a rollback
 // reads and what a re-entry inside the 30 s staleTime renders without refetching, so an accepted
 // reorder that is not written there leaves the cache disagreeing with the server. The Plex list got
-// the rollback half of that round's fix and not this half (round 6).
+// the rollback half of that fix and not this half.
 //
 // Its own file: `@dnd-kit/core` is mocked module-wide so the test can fire `onDragEnd` directly —
 // dnd-kit's pointer sensor does not activate under jsdom (no layout, no pointer capture).
@@ -64,10 +64,10 @@ const movie = (id: string) =>
   ({ id, type: "movie", title: `Movie ${id}`, year: 2020, thumb: "" }) as unknown as PlexCard;
 
 it("writes the order the server accepted into the detail cache, without waiting for a refetch", async () => {
-  // The refetch that same handler schedules (round 7) must not be what makes the cache right: the
+  // The refetch that same handler schedules must not be what makes the cache right: the
   // point of the write is that a re-entry inside the 30 s staleTime renders the accepted order
   // immediately, with no request at all. So the refetch is HELD OPEN here — if this assertion
-  // needed it to land, deleting the write would still pass and round 6's fix would be unguarded.
+  // needed it to land, deleting the write would still pass and that fix would be unguarded.
   let calls = 0;
   vi.spyOn(api, "plexPlaylist").mockImplementation(() =>
     ++calls === 1
@@ -101,7 +101,7 @@ it("refetches the detail too, so a membership change that landed meanwhile is no
   // surface, an item added elsewhere), and with only the LIST key invalidated nothing corrected it:
   // staleTime is 30 s and refetchOnWindowFocus is off, so the phantom row stayed on screen for as
   // long as the view was mounted and the next drag persisted an order built from it. The YouTube
-  // sibling invalidates the detail key right after writing it; this half did not (round 7).
+  // sibling invalidates the detail key right after writing it; this half did not.
   let calls = 0;
   vi.spyOn(api, "plexPlaylist").mockImplementation(() =>
     Promise.resolve(