feat(route): enable topology convergence by default (#311)

Default new Host shares to native-edge topology convergence while preserving the pre-share opt-out and peer-only default.

docs/adr/0005-automatic-hybrid-media-routing.md

@@ -1,8 +1,8 @@
 # ADR-0005: Automatic Hybrid Media Routing
 
-- Status: accepted and deployed; native-edge convergence is opt-in per share
+- Status: accepted; route model deployed; default-on policy pending release
 - Date: 2026-08-20
-- Last updated: 2026-08-27
+- Last updated: 2026-08-28
 
 ## Context
 
@@ -268,9 +268,10 @@ does not promise a static mathematical global optimum.
 
 Two pre-share gates constrain this same controller. Peer-only policy removes
 every SFU tuple and bootstrap path before reservation, so exhaustion keeps the
-existing explicit failure. The initial topology-convergence rollout is opt-in;
-when disabled, native evidence never creates quality work. Both gates are bound
-to share generation and cannot change while sharing.
+existing explicit failure. Topology convergence is enabled by default and may
+be disabled before sharing; when disabled, native evidence never creates
+quality work. Both gates are bound to share generation and cannot change while
+sharing.
 
 ## Consequences
 

docs/product/routing-transport.md

@@ -29,8 +29,9 @@ share generation:
 - peer-only mode excludes SFU publication, subscription, bootstrap, and quality
   candidates, while retaining the same bounded Peer graph and clear exhausted
   failure; and
-- native-edge topology convergence is initially opt-in. When disabled, quality
-  evidence stays diagnostic and availability routing is unchanged.
+- native-edge topology convergence is enabled by default. The Host may disable
+  it before sharing; when disabled, quality evidence stays diagnostic and
+  availability routing is unchanged.
 
 These are route-policy gates, not new routing algorithms. Changing either
 policy requires stopping the current share and starting another generation.

docs/project-memory.md

@@ -38,11 +38,11 @@ implementation and routine UI detail.
   one serial child operation. First decoded frame commits availability work;
   accepted quality work adds native edge proof before commit in that operation.
 - WebRTC and LiveKit own congestion control, media adaptation, reconnect, and SFU
-  layers. Current source uses persistent native edge state for opt-in local P2P
-  convergence. SFU quality work is limited to multi-root Host fanout relief and
-  each Viewer must prove its own candidate non-regression before commit. No
-  weighted score, all-pairs probe, general parent-wide prediction, or periodic
-  rebalance is accepted.
+  layers. Current source defaults persistent native-edge local P2P convergence
+  on, with a pre-share Host opt-out. SFU quality work is limited to multi-root
+  Host fanout relief and each Viewer must prove its own candidate non-regression
+  before commit. No weighted score, all-pairs probe, general parent-wide
+  prediction, or periodic rebalance is accepted.
 - Browser video uses `motion`; each share chooses H.264 with VP8 fallback through
   an actual sender probe unless the Host explicitly selects VP8 or H264. Screen
   audio uses 64/128/192 kbps ceilings with 128 default.

docs/status.md

@@ -64,7 +64,8 @@ This is the current execution index. Git history owns completed timelines;
   Viewer sees first-frame readiness plus three fresh windows that strictly
   improve delivered resolution or rounded FPS without regressing either. Peer
   candidates stay first; bounded SFU is only the suffix for whole-Host-root
-  degradation and retains native publication proof. The gate defaults off.
+  degradation and retains native publication proof. The gate defaults on in
+  source and remains a locked pre-share opt-out.
 - Screener has no custom SFU layer list, quality score, layer selector,
   all-pairs probe, parent-wide prediction or periodic rebalancing. P2P and SFU
   still recover their current route before actual failure enters reassignment.
@@ -86,8 +87,9 @@ subscribers back to P2P, and retired the Host SFU publication.
   screening did not reproduce a Host-page drop; accepted Viewer lifecycle work
   prevents frozen JavaScript wall time from becoming an immediate route failure
   but is not capture keepalive.
-- Current-path quality can drive opt-in local convergence in source and
-  production. Weighted/global optimization remains parked.
+- Current-path quality can drive local convergence in source and production.
+  Source defaults it on; production remains default-off until a later release.
+  Weighted/global optimization remains parked.
 - Repository simplification does not change or deploy product behavior beyond
   deleting proven dead private surfaces.
 - Native/executable senders, shared encode, distribution packages, broad UI

src/shared/protocol.ts

@@ -223,7 +223,7 @@ export const routePolicySchema = z
 export type RoutePolicy = z.infer<typeof routePolicySchema>;
 export const DEFAULT_ROUTE_POLICY = {
   peerOnly: false,
-  topologyOptimization: false,
+  topologyOptimization: true,
 } as const satisfies RoutePolicy;
 
 export const relayDownstreamEdgesSchema = z

tests/protocol.test.ts

@@ -273,6 +273,10 @@ describe("client signaling protocol", () => {
   });
 
   it("accepts a bounded authentication message", () => {
+    expect(DEFAULT_ROUTE_POLICY).toEqual({
+      peerOnly: false,
+      topologyOptimization: true,
+    });
     expect(
       decodeClientMessage(
         JSON.stringify({
@@ -313,6 +317,10 @@ describe("client signaling protocol", () => {
       role: "host",
       shareGeneration: "share_generation_12345678",
       qualitySettings: qualitySettingsWithAudio,
+      routePolicy: {
+        peerOnly: false,
+        topologyOptimization: true,
+      },
     });
     expect(
       clientMessageSchema.safeParse({