Merge feat/vocabulary-hint: names-and-terms decoder prompt (spec 001 workstream 2)

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

# Conflicts:
#	docs/Architecture.md
#	docs/Decisions.md
#	docs/WorkBreakdown.md
#	server-py/sidecar.py
#	src/App.tsx
#	src/server/live.ts
#	src/server/transcribe.ts
#	tests/fixtures/fake_sidecar.py
#	tests/transcribe.test.ts

tests/server-live.test.ts

@@ -121,7 +121,9 @@ describe("createLiveWebSocket", () => {
     handler.message(ws, toBuffer(speech(1)));
     handler.message(ws, toBuffer(silence(1.5))); // trailing silence commits the segment
 
-    await waitFor(() => sidecar.requests.length > 0);
+    // The fake records the request before it finishes reading the PCM file —
+    // wait for the payload itself, or a loaded machine races past it.
+    await waitFor(() => sidecar.audioPayloads.length > 0);
     expect(sidecar.requests[0]).toEqual(
       // Live passes are short: sequential decoding, no batching/VAD (D-WS5a).
       expect.objectContaining({ model: "base", language: "da", task: "translate", mode: "live" }),
@@ -140,6 +142,21 @@ describe("createLiveWebSocket", () => {
     await waitFor(async () => (await readdir(tempDir)).length === 0);
   });
 
+  test("every pass of a session carries the session's vocabulary hint (D-WS2a)", async () => {
+    const tempDir = await makeTempDir();
+    const sidecar = fakeSidecar();
+    const handler = createLiveWebSocket({ sidecar });
+    const ws = fakeSocket(sessionData({ prompt: "Maxwell, Voices of the Void", tempDir }));
+    handler.open(ws);
+    handler.message(ws, toBuffer(speech(1)));
+    handler.message(ws, toBuffer(silence(1.5)));
+    await waitFor(() => sidecar.requests.length > 0);
+    for (const request of sidecar.requests) {
+      expect(request.prompt).toBe("Maxwell, Voices of the Void");
+    }
+    await waitFor(async () => (await readdir(tempDir)).length === 0);
+  });
+
   test("a sidecar error surfaces on the update stream instead of crashing the session", async () => {
     const sidecar = fakeSidecar({ error: "model load failed" });
     const handler = createLiveWebSocket({ sidecar });