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/lib/layerStack.ts

@@ -1,6 +1,6 @@
 import { useEffect, useEffectEvent, useRef } from "react";
 
-// R8 S3 — one shared registry of open "dismissable layers" (modals, popovers, players), ordered by
+// One shared registry of open "dismissable layers" (modals, popovers, players), ordered by
 // open time. It replaces the three mutually-unaware Escape systems that each guessed at ordering with
 // its own `stopPropagation`: Modal's private `modalStack`, the popovers' document listeners, and the
 // two players' `modalCount()` checks. Now EVERY layer keeps its own key handler but coordinates
@@ -24,7 +24,7 @@ const passive = new Set<number>();
 // inside one (the recorder's step field). It must silence the page's keys — an autofocused field
 // does — yet it is not a dialog of its own, so claiming the trap left the dialog around it with
 // none, and Tab from anything outside the prompt walked into the page behind the scrim. The two
-// questions are separate; a layer answers each for itself (deep-review round 3).
+// questions are separate; a layer answers each for itself.
 const defers = new Set<number>();
 
 export function pushLayer(opts?: { passive?: boolean; trapsTab?: boolean }): number {
@@ -88,7 +88,7 @@ export function escapeHandled(e: KeyboardEvent): boolean {
 /** THE close-on-Escape test for layer `id`: an Escape no layer above has taken yet, while `id` is on
  *  top — and it claims the event for `id`. Both halves are needed: a list inside a dialog closes with a
  *  synchronous flush that pops its layer before the dialog's `window` listener runs, so topmost alone
- *  read "yes" and one Escape closed the list AND the dialog (E28, the share dialog's user picker). */
+ *  read "yes" and one Escape closed the list AND the dialog (the share dialog's user picker). */
 export function takeEscape(e: KeyboardEvent, id: number): boolean {
   if (e.key !== "Escape" || escapeHandled(e) || !isTopmost(id)) return false;
   markEscapeHandled(e);