merge(main): merge v2.6.1 hotfixes from hotfix/v2.6.1

backend/apps/agent_app.py

@@ -61,7 +61,6 @@
     check_skill_conflicts_impl,
 )
 from services.prompt_service import generate_guardrail_rules_impl
-from services.human_interaction.models import InteractionError
 from services.knowledge_scope_service import get_agent_knowledge_capabilities
 from services.agent_draft_permission_service import AgentDraftEditError
 from services.nl2agent_service import Nl2AgentDraftSaveError, create_nl2agent_stream
@@ -154,8 +153,6 @@
             authorization=authorization,
             resume=resume,
         )
-    except InteractionError as exc:
-        raise HTTPException(status_code=exc.status_code, detail=str(exc)) from exc
     except ForbiddenError as e:
         raise HTTPException(status_code=HTTPStatus.FORBIDDEN, detail=str(e)) from e
     except ValidationError as e:
@@ -193,8 +190,6 @@
             tenant_id=tenant_id,
             skip_user_save=True,
         )
-    except InteractionError as exc:
-        raise HTTPException(status_code=exc.status_code, detail=str(exc)) from exc
     except UnauthorizedError as exc:
         raise HTTPException(
             status_code=HTTPStatus.UNAUTHORIZED,
@@ -279,14 +274,7 @@
     """
     Stop an agent run by conversation ID or ephemeral debug run ID.
     """
     user_id, tenant_id = get_current_user_id(authorization)
-    from consts.const import HITL_ENABLED
-    if HITL_ENABLED:
-        from services.human_interaction.application import get_service
-        service = get_service()
-        durable_id = service.repository.latest(tenant_id, user_id, int(run_id), active_only=True) if run_id.isdigit() else None
-        if durable_id:
-            service.control(durable_id, tenant_id, user_id, "terminate")
     return stop_agent_tasks(int(run_id) if run_id.isdigit() else run_id, user_id)
 
 
@@ -300,23 +288,8 @@
 ):
     """Stop a northbound agent run inside the runtime service."""
     try:
         user_id, tenant_id = verify_internal_runtime_jwt(authorization)
-        from consts.const import HITL_ENABLED
-        if HITL_ENABLED:
-            from services.human_interaction.application import get_service
-            service = get_service()
-            durable_id = await run_blocking(
-                "hitl-service-repository-latest", service.repository.latest, tenant_id, user_id,
-                conversation_id, active_only=True, lane="control-io", owner=__name__,
-            )
-            if durable_id:
-                await run_blocking(
-                    "hitl-service-control", service.control, durable_id, tenant_id, user_id, "terminate",
-                    lane="control-io", owner=__name__,
-                )
         return stop_agent_tasks(conversation_id, user_id)
-    except InteractionError as exc:
-        raise HTTPException(status_code=exc.status_code, detail=str(exc)) from exc
     except UnauthorizedError as exc:
         raise HTTPException(
             status_code=HTTPStatus.UNAUTHORIZED,
@@ -378,7 +351,7 @@
     try:
         return await get_creating_sub_agent_info_impl(authorization)
     except Exception as e:
         logger.error(f"Agent create error: {str(e)}")
         raise HTTPException(
             status_code=HTTPStatus.INTERNAL_SERVER_ERROR, detail="Agent create error.")
 

backend/apps/northbound_app.py

@@ -10,6 +10,7 @@
 from fastapi.responses import JSONResponse, Response, StreamingResponse
 from pydantic import ValidationError as PydanticValidationError
 
+from consts.model import reject_legacy_agent_fields
 from consts.exceptions import (
     ConversationNotFoundError,
     ForbiddenError,
@@ -392,15 +393,10 @@ async def run_chat(
             }
         }],
     ),
-    enable_hitl: bool = Body(
-        False,
-        embed=True,
-        description="Enable human interaction when supported by the runtime. Cards use type=human_interaction; "
-                    "submit answers through /nb/v1/chat/human-interactions/{run_id}/requests/{request_id}/decisions.",
-    ),
     idempotency_key: Optional[str] = Header(None, alias="Idempotency-Key"),
 ):
     try:
+        reject_legacy_agent_fields(await request.json())
         ctx: NorthboundContext = await _get_northbound_context(request)
         return await start_streaming_chat(
             ctx=ctx,
@@ -413,7 +409,6 @@ async def run_chat(
             tool_params=tool_params,
             model_id=model_id,
             idempotency_key=idempotency_key,
-            enable_hitl=enable_hitl,
         )
     except LimitExceededError as e:
         logging.error(f"Too Many Requests: rate limit exceeded: {str(e)}", exc_info=e)

backend/apps/northbound_base_app.py

@@ -30,7 +30,6 @@
 from services.runtime_state_service import runtime_state_service
 from .northbound_app import router as northbound_router
 from .northbound_knowledge_app import router as northbound_knowledge_router
-from .northbound_human_interaction_app import router as northbound_human_interaction_router
 
 
 class A2AServerSettings(BaseModel):
@@ -111,7 +110,6 @@ async def thread_capacity():
 
 northbound_app.include_router(northbound_router)
 northbound_app.include_router(northbound_knowledge_router)
-northbound_app.include_router(northbound_human_interaction_router)
 
 
 # =============================================================================

backend/apps/runtime_app.py

@@ -16,8 +16,6 @@
     file_management_runtime_router as file_management_router,
 )
 from apps.skill_app import skill_creator_router
-from apps.human_interaction_app import router as human_interaction_router
-from apps.human_interaction_app import internal_router as internal_human_interaction_router
 from consts.const import RUNTIME_THREAD_SHUTDOWN_GRACE_SECONDS
 from middleware.exception_handler import ExceptionHandlerMiddleware
 from nexent.core.concurrency import (
@@ -33,9 +31,7 @@
 
 
 async def start_agent_automation_scheduler():
-    from consts.const import HITL_ENABLED
     from services.agent_automation.scheduler import agent_automation_scheduler
-    from services.human_interaction.application import get_service, human_run_scheduler
     from services.startup_recovery_service import recover_runtime_tasks
     from services.workspace_cleanup_service import cleanup_orphaned_agent_workspaces
 
@@ -49,17 +45,12 @@ async def start_agent_automation_scheduler():
     )
     cleanup_orphaned_agent_workspaces()
     await agent_automation_scheduler.start()
-    if HITL_ENABLED:
-        get_service()
-        await human_run_scheduler.start()
 
 
 async def stop_agent_automation_scheduler():
     from services.agent_automation.scheduler import agent_automation_scheduler
-    from services.human_interaction.application import human_run_scheduler
 
     await agent_automation_scheduler.stop()
-    await human_run_scheduler.stop()
 
 
 @asynccontextmanager
@@ -125,5 +116,3 @@ async def thread_capacity():
 app.include_router(file_management_router)
 app.include_router(voice_router)
 app.include_router(skill_creator_router)
-app.include_router(human_interaction_router)
-app.include_router(internal_human_interaction_router)

backend/consts/const.py

@@ -34,12 +34,6 @@ class VectorDatabaseType(str, Enum):
 # Data Processing Service Configuration
 DATA_PROCESS_SERVICE = os.getenv("DATA_PROCESS_SERVICE")
 RUNTIME_SERVICE_URL = os.getenv("RUNTIME_SERVICE_URL", "http://localhost:5014").rstrip("/")
-HITL_ENABLED = os.getenv("HITL_ENABLED", "false").lower() in ("true", "1", "yes")
-HITL_ACCEPT_NEW_RUNS = os.getenv("HITL_ACCEPT_NEW_RUNS", "true").lower() in ("true", "1", "yes")
-HITL_TOOL_APPROVAL_ENABLED = os.getenv("HITL_TOOL_APPROVAL_ENABLED", "false").lower() in ("true", "1", "yes")
-HITL_ENCRYPTION_KEY = os.getenv("HITL_ENCRYPTION_KEY", "")
-HITL_WAIT_SECONDS = int(os.getenv("HITL_WAIT_SECONDS", "86400"))
-HITL_MAX_CONCURRENCY = int(os.getenv("HITL_MAX_CONCURRENCY", "2"))
 CLIP_MODEL_PATH = os.getenv("CLIP_MODEL_PATH")
 TABLE_TRANSFORMER_MODEL_PATH = os.getenv("TABLE_TRANSFORMER_MODEL_PATH")
 UNSTRUCTURED_DEFAULT_MODEL_INITIALIZE_PARAMS_JSON_PATH = os.getenv(

backend/consts/model.py

@@ -825,11 +825,15 @@ class ConversationKnowledgeScopeUpdateRequest(BaseModel):
     scope: Optional[ConversationKnowledgeScopeRequest] = None
 
 
+def reject_legacy_agent_fields(value):
+    """Reject removed execution controls rather than silently starting a new run."""
+    if isinstance(value, dict) and {"enable_hitl", "hitl_run_id", "hitl_after_event"}.intersection(value):
+        raise ValueError("Legacy human interaction fields are no longer supported; send a normal query")
+    return value
+
+
 class AgentRequest(BaseModel):
     query: str
-    enable_hitl: bool = False
-    hitl_run_id: Optional[str] = Field(default=None, min_length=36, max_length=36)
-    hitl_after_event: int = Field(default=0, ge=0)
     conversation_id: Optional[int] = None
     history: Optional[List[HistoryItem]] = None
     # Complete list of attachment information
@@ -855,6 +859,11 @@ class AgentRequest(BaseModel):
         description="Optional optimistic-lock version for runtime metadata updates",
     )
 
+    @model_validator(mode="before")
+    @classmethod
+    def reject_legacy_execution_controls(cls, value):
+        return reject_legacy_agent_fields(value)
+
     @field_validator("context_policy")
     @classmethod
     def validate_context_policy(cls, value):