docs(comments): no WebSocket and no CI in the present tense

The chat cut removed the only WebSocket and CI went before it, but comments still leaned on both:
`?account=` was justified by the socket, job polling by an in-process WS registry, the e2e flows as
"proxy-safe (none use the WebSocket)", and pyrightconfig named a CI container the engine contract
test never checked. Each now says what holds today (the wallet gate, no push channel, the two
contract test halves, the image build's `npm ci`, the gate).

backend/app/models.py

@@ -889,8 +889,8 @@ class DownloadJob(Base, TimestampMixin, UpdatedAtMixin):
     """A user's request to download one source — the per-user layer over `MediaAsset`.
 
     The worker claims `queued` jobs with ``FOR UPDATE SKIP LOCKED`` and writes live
-    `progress`/`speed_bps`/`eta_s`/`phase` here (the frontend polls via useLiveQuery — the
-    worker is a separate process and can't reach the in-process WS registry). `profile_snapshot`
+    `progress`/`speed_bps`/`eta_s`/`phase` here (the frontend polls via useLiveQuery: there is
+    no server push channel, and the worker is a separate process anyway). `profile_snapshot`
     freezes the spec at enqueue so later edits/deletion of the profile don't change this job.
     `display_name` is the user's own name for the file (display + the Content-Disposition on
     local download); the on-disk asset is never renamed. `asset_id` links to the shared file

backend/pyrightconfig.json

@@ -13,8 +13,8 @@
 // image installs — so the lane's version IS the pin, with no host-vs-pin check to drift.
 //
 // `pythonVersion` is not free-standing: `/.python-version` is the single source of truth and
-// `tests/test_python_engine_contract.py` fails this gate if this file, the two Dockerfiles, the CI
-// container or ruff's target-version disagree with it.
+// `tests/test_python_engine_contract.py` (with its host half, pythonEngine.contract.test.ts) fails
+// this gate if this file, the two Dockerfiles or ruff's target-version disagree with it.
 //
 // `include` covers `alembic` for the same reason the ruff lane was widened to it: migrations run on
 // every container start, so one that raises there crash-loops prod, and "it is generated code" is

backend/tests/test_plex_stream.py

@@ -331,7 +331,7 @@ class TestWantHw:
 
     @staticmethod
     def _opencl(monkeypatch, present):
-        # Pin the OpenCL-runtime probe so the routing doesn't depend on the CI box's /etc/OpenCL.
+        # Pin the OpenCL-runtime probe so the routing doesn't depend on the test machine's /etc/OpenCL.
         monkeypatch.setattr(stream, "_opencl_present", lambda: present)
 
     def test_no_hw_when_not_transcoding(self, monkeypatch):

frontend/e2e/README.md

@@ -32,7 +32,7 @@ login-free `/watch/<token>` share link).
 - **`:8080`** (baked SPA) is authoritative and same-origin. After a **frontend change you must
   `siftlode dev` (rebuild)** for its baked SPA to carry the latest test hooks.
 - **`:5173`** (Vite dev) serves source live — the fast refactor loop; it proxies `/api` to `:8080`,
-  so the stack must still be up. All five flows are proxy-safe (none use the WebSocket):
+  so the stack must still be up. All five flows are proxy-safe:
 
   ```sh
   PW_BASE_URL=http://localhost:5173 siftlode e2e

frontend/playwright.config.ts

@@ -4,8 +4,8 @@ import { defineConfig, devices } from "@playwright/test";
  * E2E characterization (golden-master) net for the Platform refactor.
  *
  * Target (PW_BASE_URL): defaults to the baked localdev stack on :8080 (production-faithful,
- * same-origin, no HMR/WS gaps). Point it at http://localhost:5173 for the fast Vite loop — all
- * five characterization flows are proxy-safe (none use the WebSocket).
+ * same-origin, no HMR gaps). Point it at http://localhost:5173 for the fast Vite loop — all
+ * five characterization flows are proxy-safe.
  *
  * The stack must already be UP (owned by `siftlode dev`); there is deliberately no `webServer`.
  * `e2e/global.setup.ts` seeds the dedicated test account and logs in once, sharing the resulting

frontend/src/components/ui/ItemCount.dom.test.tsx

@@ -29,7 +29,7 @@ describe("ItemCount", () => {
   it("pluralises and thousands-separates a normal count", () => {
     render(<ItemCount count={192045} />);
     // Separators come from the shared formatNumber — compare against it, not a fixed en-US string,
-    // so the test asserts ItemCount's formatting path rather than the CI locale.
+    // so the test asserts ItemCount's formatting path rather than the runner's locale.
     expect(text()).toBe(`${formatNumber(192045, "en")} items`);
   });
 

frontend/src/lib/api/downloads.ts

@@ -200,7 +200,7 @@ export const downloadsApi = {
   // The body IS the file, and the two names ride as headers. Not a style choice: a multipart part is
   // a DECLARED body, which FastAPI reads and spools BEFORE the route's `current_user` dependency
   // runs — so with the old shape an unauthenticated caller could make the server write the whole
-  // per-file allowance to temp (P1). Headers are latin-1 on the wire, hence `encodeURIComponent`;
+  // per-file allowance to temp. Headers are latin-1 on the wire, hence `encodeURIComponent`;
   // the server percent-decodes both. Content-Type is left to the browser (the File's own type).
   uploadDownload: (
     file: File,
@@ -274,8 +274,9 @@ export const downloadsApi = {
   revokeDownloadLink: (linkId: number): Promise<{ deleted: number }> =>
     req(`/api/downloads/links/${linkId}`, { method: "DELETE" }),
   // A plain <a> navigation can't send the X-Siftlode-Account header, so pass the active
-  // account via ?account= (same wallet-gated selection the WebSocket uses) — otherwise the
-  // download resolves to the session-default account and 404s for a per-tab account's file.
+  // account via ?account= (the server honours it only for an account in this browser's wallet) —
+  // otherwise the download resolves to the session-default account and 404s for a per-tab
+  // account's file.
   downloadFileUrl: (id: number): string => {
     const a = getActiveAccount();
     return a != null ? `/api/downloads/${id}/file?account=${a}` : `/api/downloads/${id}/file`;

frontend/src/lib/api/facade.test.ts

@@ -3,7 +3,8 @@ import { describe, it, expect, beforeAll } from "vitest";
 // The façade guard: lib/api.ts is a spread-only façade over per-domain slice modules. These tests
 // mechanise the two invariants that make that structure safe — so a future edit that reintroduces
 // the interleave (a method dropped into two slices → a silent last-wins spread override) or bloats
-// the barrel (a method defined inline instead of in a domain module) fails CI instead of shipping.
+// the barrel (a method defined inline instead of in a domain module) fails the gate instead of
+// shipping.
 //
 // Scope, on purpose: this covers RUNTIME method-key collisions — the only *silent* failure mode,
 // since tsc does not flag overlapping spread keys. It deliberately does NOT cover exported-TYPE-name

frontend/src/lib/format.test.ts

@@ -133,7 +133,7 @@ describe("relativeFromMs", () => {
 // printed "1 234 567" (and a Hungarian UI in an English browser "1,234,567"). The case below asserts
 // BOTH languages and that the two differ, rather than comparing against the runtime's own default:
 // the default is whatever locale the machine running the suite has (Hungarian on the dev box, English
-// on a CI image), so a one-language assertion passes under the old bug on half the machines.
+// in the test image), so a one-language assertion passes under the old bug on half the machines.
 describe("formatNumber", () => {
   afterEach(() => {
     i18nStub.language = "en";

frontend/src/lib/store.ts

@@ -1,8 +1,8 @@
 import { useSyncExternalStore } from "react";
 
 // A minimal observable value shared across React components AND non-React code (the api
-// layer, websocket handlers, etc.). Replaces the hand-rolled `listeners`/`subscribe`/`emit`
-// triad that several modules each re-implemented slightly differently.
+// layer, the debug recorder, the error dialog, etc.). Replaces the hand-rolled
+// `listeners`/`subscribe`/`emit` triad that several modules each re-implemented slightly differently.
 export interface Store<T> {
   get(): T;
   set(next: T | ((prev: T) => T)): void;

scripts/check-node-modules.mjs

@@ -4,8 +4,8 @@
 // The local gate's frontend lanes run against the HOST's node_modules, and nothing tied that tree to
 // the lock: a `git pull` that bumped a dependency, or an `npm install` that was never committed, left
 // tsc/eslint/vitest judging the code against packages the build image (`npm ci`) does not use — a
-// green that proves nothing about what ships, in either direction. CI is immune (it runs `npm ci`);
-// this makes the local lanes say so instead of quietly differing.
+// green that proves nothing about what ships, in either direction. The image build is immune (it
+// runs `npm ci`); this makes the local lanes say so instead of quietly differing.
 //
 // Checks every package the lock places under node_modules/: present, at the pinned version. A package
 // marked `optional` may be absent (a platform binary for another OS). Exit 1 with the list and the