fix(web,gateway): close P3 re-review#4 findings — sanitize all executor catches; lock pre-start turn boundary (wire turnId deferred)

apps/web/src/spa/chat/use-chat-connection.spec.tsx

@@ -1587,7 +1587,14 @@ describe('useChatConnection', () => {
     expect(fake.emitted.filter((e) => e.event === 'command:approve')).toHaveLength(2);
   });
 
-  it("does not settle turn B when a duplicate agent:end from already-settled turn A is redelivered after B's own ack but before B's own start", async () => {
+  it("P3-5c: locks the achievable pre-start boundary — turn A settled, turn B sent and acked but before B's own start, a stale/duplicate agent:end from A does not release B (a third send stays blocked), and only B's own start/end sequence legitimately releases it", async () => {
+    // This is the provable boundary: in-order Socket.IO delivery guarantees
+    // a same-conversation agent:end arriving before this turn's own
+    // agent:start can only be a stale straggler. Once a turn is 'active',
+    // AgentEndPayload/ErrorPayload carry no turn identity to further
+    // distinguish a genuine end from a duplicate — that residual is not,
+    // and cannot be, asserted here; full correlation needs a wire turnId
+    // (deferred to P5). See the reducer comments in use-chat-connection.ts.
     // Turn A completes normally on c1.
     await act(async () => {
       fake.serverEmit('message:ack', { conversationId: 'c1', messageId: 'm1' });