fix(memory): run end-of-turn sync off the turn thread (#41945)

A misconfigured/slow external memory provider could hold the agent in
the 'running' state for minutes after the final response was delivered.
MemoryManager.sync_all / queue_prefetch_all looped provider.sync_turn /
queue_prefetch INLINE on the turn-completion path; a provider making a
blocking network/daemon call (a broken Hindsight daemon was observed
blocking ~298s before failing) blocked run_conversation from returning.
Because every interface (CLI, TUI, gateway) marks the agent 'running'
until run_conversation returns, the agent stayed busy for the full block
and any follow-up message triggered an aggressive interrupt that dropped
the message.

Dispatch provider sync/prefetch to a lazily-created single-worker
background executor. sync_all / queue_prefetch_all return immediately;
work completes (or fails, logged) in the background. A single worker
serializes writes so turn N lands before turn N+1. flush_pending()
provides a barrier for session boundaries and deterministic tests.
shutdown_all() drains the executor with a bounded timeout so a wedged
provider can never hang teardown.

Builtin-only / no-provider sessions spawn no executor (zero new threads
in the common case).

tests/agent/test_memory_provider.py

@@ -229,6 +229,7 @@ class TestMemoryManager:
         mgr.add_provider(p2)
 
         mgr.queue_prefetch_all("next turn")
+        mgr.flush_pending(timeout=5)
         assert p1.queued_prefetches == ["next turn"]
         assert p2.queued_prefetches == ["next turn"]
 
@@ -240,6 +241,7 @@ class TestMemoryManager:
         mgr.add_provider(p2)
 
         mgr.sync_all("user msg", "assistant msg")
+        mgr.flush_pending(timeout=5)
         assert p1.synced_turns == [("user msg", "assistant msg")]
         assert p2.synced_turns == [("user msg", "assistant msg")]
 
@@ -253,7 +255,7 @@ class TestMemoryManager:
         ]
 
         mgr.sync_all("user msg", "assistant msg", session_id="sess-1", messages=messages)
-
+        mgr.flush_pending(timeout=5)
         assert p.synced_turns == [("user msg", "assistant msg", "sess-1", messages)]
 
     def test_sync_all_omits_messages_for_legacy_provider(self):
@@ -262,7 +264,7 @@ class TestMemoryManager:
         mgr.add_provider(p)
 
         mgr.sync_all("user msg", "assistant msg", messages=[{"role": "tool"}])
-
+        mgr.flush_pending(timeout=5)
         assert p.synced_turns == [("user msg", "assistant msg")]
 
     def test_sync_failure_doesnt_block_others(self):
@@ -275,6 +277,7 @@ class TestMemoryManager:
         mgr.add_provider(p2)
 
         mgr.sync_all("user", "assistant")
+        mgr.flush_pending(timeout=5)
         # p1 failed but p2 still synced
         assert p2.synced_turns == [("user", "assistant")]