typedef-data-intelligence 0.32.0__tar.gz → 0.34.0__tar.gz

{typedef_data_intelligence-0.32.0 → typedef_data_intelligence-0.34.0}/CHANGELOG.md

@@ -1,5 +1,64 @@
 # Changelog
 
+## [0.34.0](https://github.com/typedef-ai/typedef/compare/typedef-data-intelligence-v0.33.0...typedef-data-intelligence-v0.34.0) (2026-07-01)
+
+
+### Miscellaneous Chores
+
+* **agent:** prune dead PydanticAI orchestrator builders + .yaml.j2 prompts (TD-3776) ([#1065](https://github.com/typedef-ai/typedef/issues/1065)) ([0d05b9d](https://github.com/typedef-ai/typedef/commit/0d05b9d195a921db341252314fed5c286356a5ce))
+
+
+### Features
+
+* **desktop:** Airflow additional catalogs/databases allowlist + fix DatabaseRef allowlist PATCH (TD-3770, TD-3772) ([#1060](https://github.com/typedef-ai/typedef/issues/1060)) ([60d550c](https://github.com/typedef-ai/typedef/commit/60d550c7beacb9d202e18c2bd93b7901b15e4f50))
+* **desktop:** capture frontend webview console logs (TD-3711) ([#1041](https://github.com/typedef-ai/typedef/issues/1041)) ([cc1a345](https://github.com/typedef-ai/typedef/commit/cc1a345a4f11d6757b4e18334a4f5ac297f58dc4))
+* **testing:** add portable test-home overlays ([#1047](https://github.com/typedef-ai/typedef/issues/1047)) ([56bc013](https://github.com/typedef-ai/typedef/commit/56bc01367995a9c53c66531914302c744548c1bc))
+* **testing:** clone referenced repos during test-home provisioning ([#1048](https://github.com/typedef-ai/typedef/issues/1048)) ([1f38adf](https://github.com/typedef-ai/typedef/commit/1f38adfbd85ad321638429451334069799802794))
+
+
+### Bug Fixes
+
+* **core:** remove silent config/workspace fallbacks — fail-fast (TD-3515) ([#1053](https://github.com/typedef-ai/typedef/issues/1053)) ([b599d0a](https://github.com/typedef-ai/typedef/commit/b599d0a111abd9c57ef20ba7e963002c61d468f8))
+* **desktop:** name failing preflight step and pin error banner by Continue (TD-3657) ([#1039](https://github.com/typedef-ai/typedef/issues/1039)) ([8a0c004](https://github.com/typedef-ai/typedef/commit/8a0c004fc4178edba294771caf3160b4cd848ec7))
+* **desktop:** resolve worktree repo from active worktree (TD-3514) ([#1026](https://github.com/typedef-ai/typedef/issues/1026)) ([d033653](https://github.com/typedef-ai/typedef/commit/d0336531c3d723a7194e5491e729c640098b03f0))
+
+
+### Code Refactoring
+
+* **ingest:** move dbt core behind graph source boundary (TD-3432) ([#1074](https://github.com/typedef-ai/typedef/issues/1074)) ([6c471c5](https://github.com/typedef-ai/typedef/commit/6c471c51e78803568cb52db1fee5d66d023385fc))
+
+## [0.33.0](https://github.com/typedef-ai/typedef/compare/typedef-data-intelligence-v0.32.0...typedef-data-intelligence-v0.33.0) (2026-06-26)
+
+
+### Miscellaneous Chores
+
+* **typecheck:** bootstrap shared ty burndown helpers (TD-3643) ([#1027](https://github.com/typedef-ai/typedef/issues/1027)) ([8661932](https://github.com/typedef-ai/typedef/commit/866193281039c974cae3b4b8d41b5adc28452dee))
+
+
+### Features
+
+* **desktop-sync:** add request-scoped matrix sync overrides (TD-3629) ([#1022](https://github.com/typedef-ai/typedef/issues/1022)) ([5db8f29](https://github.com/typedef-ai/typedef/commit/5db8f2945f41b92772bb3786052d2d843822d209))
+* **desktop:** add isolated test home tooling ([#996](https://github.com/typedef-ai/typedef/issues/996)) ([309a5f3](https://github.com/typedef-ai/typedef/commit/309a5f30f944309414e6ee1489fc816e87fe1d60))
+* **desktop:** kind-aware Airflow worktree + project-relative RTIF snapshot (TD-3446, TD-3409) ([#1018](https://github.com/typedef-ai/typedef/issues/1018)) ([edd4ae4](https://github.com/typedef-ai/typedef/commit/edd4ae4d76c792ef48e3fe59566baf82c41309a2))
+* **just:** worktree disk reclaim recipes (tdh-clean + worktree-reclaim) (TD-3660) ([#1036](https://github.com/typedef-ai/typedef/issues/1036)) ([a1d7e7e](https://github.com/typedef-ai/typedef/commit/a1d7e7e9899f5ed0168130a8b6a44d8f67798424))
+
+
+### Bug Fixes
+
+* **agent:** await memory observer org writes (TD-3645) ([#1032](https://github.com/typedef-ai/typedef/issues/1032)) ([70abe66](https://github.com/typedef-ai/typedef/commit/70abe66ee1d93adcd715ade66ae7f09a68094299))
+* **core:** refuse to guess workspace for graph reads (TD-3378) ([#978](https://github.com/typedef-ai/typedef/issues/978)) ([036c6bf](https://github.com/typedef-ai/typedef/commit/036c6bf7c26061d8161b71542cf2bb2c03c42f81))
+* **desktop:** make worktree DB clone usable by the agent (TD-3507) ([#974](https://github.com/typedef-ai/typedef/issues/974)) ([beb8e2e](https://github.com/typedef-ai/typedef/commit/beb8e2e60d2c216147259e32624048c5493519df))
+* **desktop:** strip server-owned `unused` flag before connector validation (TD-3632) ([#1013](https://github.com/typedef-ai/typedef/issues/1013)) ([557f832](https://github.com/typedef-ai/typedef/commit/557f832aea36cdce13651472311fafa405105853))
+* **desktop:** workspace-key clone state at the source (TD-3565) ([#1000](https://github.com/typedef-ai/typedef/issues/1000)) ([b8f526b](https://github.com/typedef-ai/typedef/commit/b8f526be3fec093b6d68b53a413355221326a9d0))
+* **desktop:** workspace-scope all worktree-keyed state (TD-3562) ([#988](https://github.com/typedef-ai/typedef/issues/988)) ([b7df8d5](https://github.com/typedef-ai/typedef/commit/b7df8d54485518dcb1eed8d4bf109d62fb940c1c))
+* **desktop:** workspace-scope worktree resolution + write-target guard (TD-3610, TD-3623) ([#1005](https://github.com/typedef-ai/typedef/issues/1005)) ([66e56ec](https://github.com/typedef-ai/typedef/commit/66e56ecf591492ffb6cceea741eae7ee0047210c))
+* **ingest:** lock semantic scoped-write contract and surface failures (TD-3633) ([#1031](https://github.com/typedef-ai/typedef/issues/1031)) ([b804356](https://github.com/typedef-ai/typedef/commit/b8043568ea4fc7e85fde505374ad1005268a0036))
+
+
+### Code Refactoring
+
+* **desktop:** collapse worktree clone state into one carrier (TD-3507) ([#998](https://github.com/typedef-ai/typedef/issues/998)) ([a9f25f1](https://github.com/typedef-ai/typedef/commit/a9f25f12410b017e0db67629130a73acd0ee72ed))
+
 ## [0.32.0](https://github.com/typedef-ai/typedef/compare/typedef-data-intelligence-v0.31.1...typedef-data-intelligence-v0.32.0) (2026-06-20)
 
 

{typedef_data_intelligence-0.32.0 → typedef_data_intelligence-0.34.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: typedef-data-intelligence
-Version: 0.32.0
+Version: 0.34.0
 Summary: Unified lineage management and AI-powered analytics platform for dbt projects
 Project-URL: Homepage, https://typedef.ai
 Project-URL: Repository, https://github.com/typedef-ai/data-intelligence

{typedef_data_intelligence-0.32.0 → typedef_data_intelligence-0.34.0}/docs/examples/memory_usage_examples.py

@@ -1,15 +1,18 @@
 """Memory System Usage Examples.
 
-This module demonstrates how to use the memory system in different scenarios.
+This module demonstrates the async MemoryStorage protocol shape in different
+scenarios.
 
-Prerequisites:
-    - FalkorDB running on localhost:6379
-    - Config with memory.enabled=true
-    - Agents initialized with memory backend
+Note:
+    The FalkorDB-backed memory adapters were removed in M6 (TD-3358), and no
+    concrete replacement backend is currently registered. This script exits
+    early until create_memory_backend() can return a real MemoryStorage backend.
 """
 
+import asyncio
 
 from agent.memory_observer import MemoryObserver
+from core.backends.config import MemoryConfig
 from core.backends.memory.factory import create_memory_backend
 from core.backends.memory.models import (
     DataPattern,
@@ -17,32 +20,32 @@ from core.backends.memory.models import (
     EpisodeType,
     UserPreference,
 )
+from core.backends.memory.protocol import MemoryStorage
+
+
+def create_example_memory_backend() -> MemoryStorage | None:
+    """Return the example backend, or None until a post-M6 backend exists."""
+    return create_memory_backend(
+        MemoryConfig(
+            backend="none",
+            enabled=False,
+            default_org_id="example_org",
+            default_user_id="local-user",
+        )
+    )
+
 
 # ============================================================================
 # Example 1: Basic Memory Backend Usage
 # ============================================================================
 
 
-def example_basic_usage():
+async def example_basic_usage(memory: MemoryStorage):
     """Basic memory backend initialization and usage."""
     print("=" * 60)
     print("Example 1: Basic Memory Backend Usage")
     print("=" * 60)
 
-    # Create memory backend
-    memory = create_memory_backend(
-        backend="falkordb",
-        host="localhost",
-        port=6379,
-        username="",
-        password="",
-        default_org_id="example_org",
-    )
-
-    if not memory:
-        print("❌ Memory backend not available")
-        return
-
     print("✅ Memory backend initialized")
 
     # Store a user preference
@@ -54,18 +57,15 @@ def example_basic_usage():
         learned_from="Interactive session with Alice",
     )
 
-    memory.store_user_memory(user_id="alice", episode=preference.to_episode())
+    await memory.store_user_memory(user_id="alice", episode=preference.to_episode())
     print("✅ Stored user preference: default_chart_type=bar")
 
     # Search user memory
-    results = memory.search_user_memory(user_id="alice", query="chart", limit=5)
+    results = await memory.search_user_memory(user_id="alice", query="chart", limit=5)
     print(f"✅ Found {len(results)} memories for alice")
 
     for result in results:
-        print(
-            f"   - {result.episode.name} (score: {result.score:.2f}): "
-            f"{result.episode.content[:50]}..."
-        )
+        print(f"   - {result.rank}. {result.fact[:80]}...")
 
     # Store organization pattern
     pattern = DataPattern(
@@ -78,17 +78,17 @@ def example_basic_usage():
         confidence=0.95,
     )
 
-    memory.store_org_memory(org_id="example_org", episode=pattern.to_episode())
+    await memory.store_org_memory(org_id="example_org", episode=pattern.to_episode())
     print("✅ Stored organization pattern: customers_to_orders join")
 
     # Search organization memory
-    results = memory.search_org_memory(
+    results = await memory.search_org_memory(
         org_id="example_org", query="customer joins", limit=5
     )
     print(f"✅ Found {len(results)} org patterns")
 
     # Hybrid search (both user and org)
-    results = memory.search_all(
+    results = await memory.search_all(
         user_id="alice", org_id="example_org", query="customer", limit=10
     )
     print(f"✅ Hybrid search found {len(results)} total memories")
@@ -101,29 +101,19 @@ def example_basic_usage():
 # ============================================================================
 
 
-def example_memory_observer():
+async def example_memory_observer(memory: MemoryStorage):
     """Demonstrate automatic pattern capture with MemoryObserver."""
     print("=" * 60)
     print("Example 2: Memory Observer for Auto-Capture")
     print("=" * 60)
 
-    memory = create_memory_backend(
-        backend="falkordb",
-        host="localhost",
-        port=6379,
-    )
-
-    if not memory:
-        print("❌ Memory backend not available")
-        return
-
     # Create observer
     observer = MemoryObserver(memory_backend=memory)
     print("✅ Memory observer initialized")
 
     # Scenario 1: User queries a semantic view
     print("\n📊 Scenario: User queries ARR by customer segment")
-    observer.observe_semantic_query(
+    await observer.observe_semantic_query(
         org_id="example_org",
         view_name="sv_arr_reporting",
         query_pattern="Group by customer_segment, aggregate total_arr",
@@ -135,7 +125,7 @@ def example_memory_observer():
 
     # Scenario 2: Agent discovers a join
     print("\n🔗 Scenario: Agent joins ARR to customer dimension")
-    observer.observe_join_pattern(
+    await observer.observe_join_pattern(
         org_id="example_org",
         models=["fct_arr_reporting", "dim_customers"],
         join_condition="customer_id",
@@ -145,7 +135,7 @@ def example_memory_observer():
 
     # Scenario 3: Agent defines a metric
     print("\n📈 Scenario: Agent calculates MRR metric")
-    observer.observe_metric_definition(
+    await observer.observe_metric_definition(
         org_id="example_org",
         metric_name="MRR",
         calculation="SUM(subscription_amount) / 12",
@@ -156,7 +146,7 @@ def example_memory_observer():
 
     # Scenario 4: Agent finds data quality issue
     print("\n⚠️  Scenario: Agent finds null email addresses")
-    observer.observe_data_quality_issue(
+    await observer.observe_data_quality_issue(
         org_id="example_org",
         issue_type="null_values",
         description="dim_customers has approximately 5% null email addresses",
@@ -166,7 +156,9 @@ def example_memory_observer():
     print("✅ Auto-captured data quality issue")
 
     # Verify patterns were stored
-    results = memory.search_org_memory(org_id="example_org", query="patterns", limit=10)
+    results = await memory.search_org_memory(
+        org_id="example_org", query="patterns", limit=10
+    )
     print(f"\n✅ Total patterns stored: {len(results)}")
 
     print()
@@ -177,27 +169,17 @@ def example_memory_observer():
 # ============================================================================
 
 
-def example_user_personalization():
+async def example_user_personalization(memory: MemoryStorage):
     """Demonstrate building user personalization over time."""
     print("=" * 60)
     print("Example 3: User Personalization Flow")
     print("=" * 60)
 
-    memory = create_memory_backend(
-        backend="falkordb",
-        host="localhost",
-        port=6379,
-    )
-
-    if not memory:
-        print("❌ Memory backend not available")
-        return
-
     user_id = "bob@example.com"
 
     # Session 1: User asks about ARR
     print("\n📅 Session 1: Bob asks about ARR trends")
-    memory.store_user_memory(
+    await memory.store_user_memory(
         user_id=user_id,
         episode=Episode(
             name="arr_trends_query",
@@ -210,7 +192,7 @@ def example_user_personalization():
 
     # Session 2: User expresses chart preference
     print("\n📅 Session 2: Bob mentions chart preference")
-    memory.store_user_memory(
+    await memory.store_user_memory(
         user_id=user_id,
         episode=UserPreference(
             key="preferred_chart_for_trends",
@@ -224,7 +206,7 @@ def example_user_personalization():
 
     # Session 3: User sets default date range
     print("\n📅 Session 3: Bob sets default date range")
-    memory.store_user_memory(
+    await memory.store_user_memory(
         user_id=user_id,
         episode=UserPreference(
             key="default_date_range",
@@ -238,11 +220,13 @@ def example_user_personalization():
 
     # Session 4: Recall context for personalization
     print("\n📅 Session 4: Bob returns, agent recalls preferences")
-    results = memory.search_user_memory(user_id=user_id, query="preferences", limit=5)
+    results = await memory.search_user_memory(
+        user_id=user_id, query="preferences", limit=5
+    )
 
     print(f"✅ Recalled {len(results)} preferences:")
     for result in results:
-        print(f"   - {result.episode.name}: {result.episode.content[:60]}...")
+        print(f"   - {result.rank}. {result.fact[:80]}...")
 
     # Agent can now personalize:
     # - Use line charts for trend queries (not bar charts)
@@ -257,22 +241,12 @@ def example_user_personalization():
 # ============================================================================
 
 
-def example_org_knowledge_building():
+async def example_org_knowledge_building(memory: MemoryStorage):
     """Demonstrate building organization-wide knowledge base."""
     print("=" * 60)
     print("Example 4: Organization Knowledge Building")
     print("=" * 60)
 
-    memory = create_memory_backend(
-        backend="falkordb",
-        host="localhost",
-        port=6379,
-    )
-
-    if not memory:
-        print("❌ Memory backend not available")
-        return
-
     org_id = "acme_corp"
 
     # Pattern 1: Common fact-to-dimension joins
@@ -294,7 +268,7 @@ def example_org_knowledge_building():
             discovered_by="data_engineer_agent",
             confidence=0.95,
         )
-        memory.store_org_memory(org_id=org_id, episode=pattern.to_episode())
+        await memory.store_org_memory(org_id=org_id, episode=pattern.to_episode())
 
     print(f"✅ Stored {len(common_joins)} common join patterns")
 
@@ -320,7 +294,7 @@ def example_org_knowledge_building():
             discovered_by="analyst_agent",
             confidence=0.99,
         )
-        memory.store_org_memory(org_id=org_id, episode=pattern.to_episode())
+        await memory.store_org_memory(org_id=org_id, episode=pattern.to_episode())
 
     print(f"✅ Stored {len(metrics)} metric definitions")
 
@@ -341,20 +315,17 @@ def example_org_knowledge_building():
             discovered_by="data_engineer_agent",
             confidence=0.90,
         )
-        memory.store_org_memory(org_id=org_id, episode=pattern.to_episode())
+        await memory.store_org_memory(org_id=org_id, episode=pattern.to_episode())
 
     print(f"✅ Stored {len(grains)} grain patterns")
 
     # Query the built knowledge
     print("\n🔍 Querying organization knowledge")
-    results = memory.search_org_memory(org_id=org_id, query="ARR metric", limit=5)
+    results = await memory.search_org_memory(org_id=org_id, query="ARR metric", limit=5)
     print(f"✅ ARR metric search: {len(results)} results")
 
     for result in results:
-        print(
-            f"   - {result.episode.metadata.get('pattern_type', 'unknown')}: "
-            f"{result.episode.name} (confidence: {result.episode.metadata.get('confidence', 0):.2f})"
-        )
+        print(f"   - {result.rank}. {result.fact[:80]}...")
 
     print()
 
@@ -364,39 +335,29 @@ def example_org_knowledge_building():
 # ============================================================================
 
 
-def example_memory_management():
+async def example_memory_management(memory: MemoryStorage):
     """Demonstrate memory statistics and management."""
     print("=" * 60)
     print("Example 5: Memory Statistics and Management")
     print("=" * 60)
 
-    memory = create_memory_backend(
-        backend="falkordb",
-        host="localhost",
-        port=6379,
-    )
-
-    if not memory:
-        print("❌ Memory backend not available")
-        return
-
     # Get user statistics
     print("\n📊 User Memory Statistics")
-    stats = memory.get_stats(user_id="alice")
+    stats = await memory.get_stats(user_id="alice")
     print(f"User memory: {stats}")
 
     # Get organization statistics
     print("\n📊 Organization Memory Statistics")
-    stats = memory.get_stats(org_id="example_org")
+    stats = await memory.get_stats(org_id="example_org")
     print(f"Org memory: {stats}")
 
     # Clear user memory (GDPR compliance example)
     print("\n🗑️  GDPR: User requests memory deletion")
-    memory.clear_user_memory(user_id="alice")
+    await memory.clear_user_memory(user_id="alice")
     print("✅ Alice's memory cleared")
 
     # Verify deletion
-    stats = memory.get_stats(user_id="alice")
+    stats = await memory.get_stats(user_id="alice")
     print(f"Alice's memory after deletion: {stats}")
 
     print()
@@ -407,18 +368,28 @@ def example_memory_management():
 # ============================================================================
 
 
-if __name__ == "__main__":
+async def main():
+    """Run all memory usage examples."""
     print("\n" + "=" * 60)
     print("MEMORY SYSTEM USAGE EXAMPLES")
     print("=" * 60 + "\n")
 
+    memory = create_example_memory_backend()
+    if memory is None:
+        print("❌ Memory backend not available; examples skipped")
+        return
+
     # Run all examples
-    example_basic_usage()
-    example_memory_observer()
-    example_user_personalization()
-    example_org_knowledge_building()
-    example_memory_management()
+    await example_basic_usage(memory)
+    await example_memory_observer(memory)
+    await example_user_personalization(memory)
+    await example_org_knowledge_building(memory)
+    await example_memory_management(memory)
 
     print("=" * 60)
     print("All examples completed!")
     print("=" * 60)
+
+
+if __name__ == "__main__":
+    asyncio.run(main())

{typedef_data_intelligence-0.32.0 → typedef_data_intelligence-0.34.0}/docs/examples/pydantic_tool_customization_example.py

@@ -3,9 +3,9 @@
 This demonstrates how the same tool can have different descriptions
 depending on which specialist is using it.
 """
-from agent.orchestrator import AgentDeps
 from agent.toolsets.data import add_execute_query_tool
 from agent.toolsets.graph import add_query_graph_tool
+from agent.types import AgentDeps
 from pydantic_ai import Agent
 
 

{typedef_data_intelligence-0.32.0 → typedef_data_intelligence-0.34.0}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "typedef-data-intelligence"
-version = "0.32.0"
+version = "0.34.0"
 description = "Unified lineage management and AI-powered analytics platform for dbt projects"
 readme = "README.pypi.md"
 license = "LicenseRef-Proprietary"

{typedef_data_intelligence-0.32.0 → typedef_data_intelligence-0.34.0}/src/lineage/api/app.py

@@ -52,6 +52,15 @@ _HAS_DESKTOP_API = _API_PROFILE.has(ApiCapability.DESKTOP_ROUTES) and _API_PROFI
     ApiCapability.UNIFIED_GRAPH
 )
 _NEEDS_LEGACY_STATE = False
+DYNAMIC_LOCAL_CORS_REGEX = r"^https?://(localhost|127\.0\.0\.1)(:\d+)?$"
+
+
+def _dynamic_local_cors_regex() -> str | None:
+    """Allow arbitrary local dev ports only when explicitly enabled."""
+    enabled = os.environ.get("TYPEDEF_ENABLE_DYNAMIC_LOCAL_CORS", "").lower()
+    if _HAS_DESKTOP_API and enabled in {"1", "true", "yes"}:
+        return DYNAMIC_LOCAL_CORS_REGEX
+    return None
 
 
 def _legacy_state_from_request(request: Request) -> Any | None:
@@ -296,7 +305,10 @@ async def lifespan(app: FastAPI):
             logger.debug("gold-build temp dir cleanup failed", exc_info=True)
 
     try:
-        from core.paths import cleanup_stale_wizard_scratch
+        from core.paths import (
+            cleanup_stale_frontend_logs,
+            cleanup_stale_wizard_scratch,
+        )
 
         removed_wizard_scratch = cleanup_stale_wizard_scratch()
         if removed_wizard_scratch:
@@ -305,6 +317,15 @@ async def lifespan(app: FastAPI):
                 len(removed_wizard_scratch),
             )
 
+        # TD-3711: prune captured frontend (webview) logs by age + total size
+        # so optional capture can never grow unbounded.
+        removed_frontend_logs = cleanup_stale_frontend_logs()
+        if removed_frontend_logs:
+            logger.info(
+                "Removed %d stale frontend log file(s) on startup",
+                len(removed_frontend_logs),
+            )
+
         # TD-3200 Phase 3: convert any already-workspaces-shape config from the
         # pre-envelope ContextEntry/flat-LineageTarget shape to the
         # ContextProject/BuildTarget envelope.  Operates only on
@@ -408,6 +429,9 @@ async def lifespan(app: FastAPI):
                         "Initialized standalone threads backend: %s",
                         _ucfg_for_threads.threads.db_path,
                     )
+                    _migrate_threads_workspace_scope(
+                        app.state.standalone_threads_backend, _ucfg_for_threads
+                    )
                 except Exception as exc:
                     logger.warning(
                         "Could not initialize threads backend for desktop: %s", exc
@@ -1060,6 +1084,35 @@ def initialize_app_state() -> AppState:
     )
 
 
+def _migrate_threads_workspace_scope(
+    threads_backend: ThreadsBackend, cfg: UnifiedConfig
+) -> None:
+    """Run the one-time TD-3562 worktree-scope re-key on the desktop threads DB.
+
+    Re-keys existing chat history from bare worktree ids to workspace-scoped
+    keys so it isn't silently orphaned by the new scoping.  Idempotent (the
+    backend guards on a versioned sentinel).  Never breaks boot — a migration
+    failure degrades to "history may appear under the wrong key" rather than a
+    dead backend.  TEMPORARY: remove with the migration module one release
+    later (TD-3564).
+    """
+    from core.backends.threads.migration import migrate_workspace_scope
+
+    try:
+        workspaces = getattr(cfg, "workspaces", None) or {}
+        workspace_worktrees = {
+            ws_name: list((getattr(ws, "worktrees", None) or {}).keys())
+            for ws_name, ws in workspaces.items()
+        }
+        migrate_workspace_scope(threads_backend, workspace_worktrees)
+    except Exception:  # noqa: BLE001 — never break boot on migration issues
+        logger.warning(
+            "[td-3562] threads worktree-scope migration failed; existing chat "
+            "history may appear under its pre-migration key until resolved",
+            exc_info=True,
+        )
+
+
 def _load_unified_config() -> UnifiedConfig:
     """Load UnifiedConfig for desktop profile mode.
 
@@ -1182,6 +1235,7 @@ app.add_middleware(
         "tauri://localhost",
         "https://tauri.localhost",
     ],
+    allow_origin_regex=_dynamic_local_cors_regex(),
     allow_credentials=True,
     allow_methods=["*"],
     allow_headers=["*"],
@@ -1271,6 +1325,7 @@ if _HAS_DESKTOP_API:
     )
     from agent.api.desktop_pty import router as desktop_pty_router  # noqa: E402
     from lineage.api.desktop_agents import router as desktop_agents_router  # noqa: E402
+    from lineage.api.desktop_logs import router as desktop_logs_router  # noqa: E402
     from lineage.api.desktop_quickstart import (
         router as desktop_quickstart_router,  # noqa: E402
     )
@@ -1305,6 +1360,7 @@ if _HAS_DESKTOP_API:
     app.include_router(desktop_quickstart_router)
     app.include_router(desktop_sync_router)
     app.include_router(desktop_debug_router)
+    app.include_router(desktop_logs_router)
 
     # Mount desktop-graph MCP server at /mcp/ via streamable-http transport.
     # When running inside the desktop app, the sidecar is always available —