refactor(agent): remove memory agent, simplify runtime config system

backend/src/core/agentscope/prompts/__init__.py

@@ -1,9 +1,11 @@
 from core.agentscope.prompts.agent_prompt import build_agent_prompt
+from core.agentscope.prompts.memory_prompt import build_memory_prompt
 from core.agentscope.prompts.system_prompt import build_system_prompt
 from core.agentscope.prompts.tool_prompt import build_tools_prompt
 
 __all__ = [
     "build_agent_prompt",
+    "build_memory_prompt",
     "build_system_prompt",
     "build_tools_prompt",
 ]

backend/src/core/agentscope/prompts/agent_prompt.py

@@ -66,6 +66,7 @@ def _router_rules(llm_config: SystemAgentLLMConfig | None) -> list[str]:
         "- Router only: extract intent and route strategy; never answer user directly.",
         "- Preserve intent in normalized_task_input.user_text; keep wording concise and faithful.",
         "- Fill multimodal_summary only when image/attachment changes execution decisions.",
+        "- Fill normalized_task_input.context_summary with a brief description of what the provided context messages contain; this is critical for worker to understand the conversational background.",
         "- Return key_entities and constraints that are execution-relevant; low confidence -> omit rather than guess.",
         "- Set execution_mode by complexity: onestep / tool_assisted / multistep.",
         "- Set result_typing.primary to the most suitable response shape; use clarification_request only when required info is missing.",
@@ -97,23 +98,6 @@ def _worker_rules(llm_config: SystemAgentLLMConfig | None) -> list[str]:
     ]
 
 
-def _memory_rules(llm_config: SystemAgentLLMConfig | None) -> list[str]:
-    return [
-        "[Memory Agent]",
-        "- Analyze conversation context and output structured memory-safe conclusions.",
-        "- Return exactly one agent output JSON object matching the runtime-injected schema.",
-        "[Responsibilities]",
-        "- Focus on extracting durable user facts and preferences from context.",
-        "- Keep outputs concise, deterministic, and evidence-backed.",
-        "- Do not invent facts or hidden user intent.",
-        "- Use tool calls only when required by explicit workflow and allowed tool groups.",
-        "[Schema Guidance]",
-        "- The output schema is injected at runtime; follow it exactly.",
-        "- Do not add fields that are not present in the injected schema.",
-        *_config_rules(llm_config),
-    ]
-
-
 def build_worker_contract_prompt(*, router_output: RouterAgentOutput) -> str:
     contract_json = json.dumps(
         router_output.model_dump(mode="json", exclude_none=True),
@@ -125,6 +109,7 @@ def build_worker_contract_prompt(*, router_output: RouterAgentOutput) -> str:
             "[Worker Contract]",
             "- Keep routed objective unchanged.",
             "- Use normalized_task_input as objective text.",
+            "- Use context_summary to understand conversational background from chat history.",
             "- Use multimodal_summary/key_entities/constraints as execution evidence.",
             "- Infer deterministic missing required tool args from evidence + tool schema.",
             "- Ask clarification only when safe inference is impossible.",
@@ -137,7 +122,6 @@ def build_worker_contract_prompt(*, router_output: RouterAgentOutput) -> str:
 AGENT_PROMPT_REGISTRY = AgentPromptRegistry()
 AGENT_PROMPT_REGISTRY.register(agent_type=AgentType.ROUTER, builder=_router_rules)
 AGENT_PROMPT_REGISTRY.register(agent_type=AgentType.WORKER, builder=_worker_rules)
-AGENT_PROMPT_REGISTRY.register(agent_type=AgentType.MEMORY, builder=_memory_rules)
 
 
 def build_agent_prompt(

backend/src/core/agentscope/prompts/memory_prompt.py

@@ -0,0 +1,52 @@
+from __future__ import annotations
+
+import json
+from typing import Any
+
+from schemas.memories import MemoryContext, MemoryListResponse
+
+
+def _wrap_section(section: str, content: str) -> str:
+    marker_map = {
+        "memory": ("<!-- MEMORY_START -->", "<!-- MEMORY_END -->"),
+    }
+    start, end = marker_map[section]
+    body = content.strip()
+    return f"{start}\n{body}\n{end}" if body else f"{start}\n{end}"
+
+
+def _format_memory_content(content: dict[str, Any]) -> str:
+    if isinstance(content, dict):
+        return json.dumps(content, ensure_ascii=True, separators=(",", ":"))
+    return str(content)
+
+
+def _format_memory(ctx: MemoryContext) -> str:
+    parts = [
+        f"[{ctx.memory_type.value.upper()}] {ctx.title or 'Untitled'}",
+        f"  source: {ctx.source.value}",
+        f"  content: {_format_memory_content(ctx.content)}",
+    ]
+    if ctx.created_at:
+        parts.append(f"  created_at: {ctx.created_at.isoformat()}")
+    return "\n".join(parts)
+
+
+def build_memory_prompt(
+    *,
+    memories: MemoryListResponse,
+) -> str | None:
+    if not memories.memories:
+        return None
+
+    lines: list[str] = [
+        "[User Memories]",
+        "- Memories are persistent context from previous sessions.",
+        "- Use them to ground responses in known user facts and preferences.",
+        "- Do not invent facts not present in memories.",
+    ]
+
+    for ctx in memories.memories:
+        lines.append(_format_memory(ctx))
+
+    return _wrap_section("memory", "\n".join(lines))

backend/src/core/agentscope/prompts/system_prompt.py

@@ -9,10 +9,12 @@ from ag_ui.core.types import Tool
 from core.agentscope.prompts.agent_prompt import (
     build_agent_prompt,
 )
+from core.agentscope.prompts.memory_prompt import build_memory_prompt
 from core.agentscope.prompts.route_prompt import build_frontend_route_prompt
 from core.agentscope.prompts.tool_prompt import build_tools_prompt
 from schemas.agent.system_agent import AgentType, SystemAgentLLMConfig
 from schemas.agent.forwarded_props import ClientTimeContext
+from schemas.memories import MemoryListResponse
 from schemas.user.context import UserContext
 
 
@@ -202,12 +204,13 @@ def _build_route_section() -> str:
 def build_system_prompt(
     *,
     agent_type: AgentType,
-    llm_config: SystemAgentLLMConfig | None,
+    llm_config: SystemAgentLLMConfig | None = None,
     user_context: UserContext,
     now_utc: datetime,
     runtime_client_time: ClientTimeContext | None = None,
     extra_context: str | None = None,
     tools: Sequence[Tool | dict[str, Any]] | None = None,
+    memories: MemoryListResponse | None = None,
 ) -> str:
     include_route_section = agent_type == AgentType.WORKER
     sections: list[str | None] = [
@@ -225,6 +228,7 @@ def build_system_prompt(
             llm_config=llm_config,
         ),
         build_tools_prompt(tools=tools) if tools else None,
+        build_memory_prompt(memories=memories) if memories else None,
         _build_output_rules(),
     ]
     return "\n\n".join(item for item in sections if item).strip()

backend/src/core/agentscope/runtime/context_loader_registry.py

@@ -1,53 +0,0 @@
-from __future__ import annotations
-
-from collections.abc import Awaitable, Callable
-from typing import Any
-
-from schemas.agent.system_agent import ContextBuildStrategy
-
-ContextLoader = Callable[[Any, str, int, int], Awaitable[dict[str, object] | None]]
-
-
-class ContextLoaderRegistry:
-    def __init__(self) -> None:
-        self._loaders: dict[ContextBuildStrategy, ContextLoader] = {}
-
-    def register(self, *, mode: ContextBuildStrategy, loader: ContextLoader) -> None:
-        self._loaders[mode] = loader
-
-    def resolve(self, *, mode: ContextBuildStrategy) -> ContextLoader:
-        loader = self._loaders.get(mode)
-        if loader is None:
-            raise ValueError(f"unsupported context mode: {mode.value}")
-        return loader
-
-
-async def _load_number(
-    service: Any,
-    thread_id: str,
-    count: int,
-    visibility_mask: int,
-) -> dict[str, object] | None:
-    return await service.load_by_user_message_window(
-        thread_id=thread_id,
-        user_message_limit=max(count, 1),
-        visibility_mask=visibility_mask,
-    )
-
-
-async def _load_day(
-    service: Any,
-    thread_id: str,
-    count: int,
-    visibility_mask: int,
-) -> dict[str, object] | None:
-    return await service.load_by_day_window(
-        thread_id=thread_id,
-        day_count=max(count, 1),
-        visibility_mask=visibility_mask,
-    )
-
-
-CONTEXT_LOADER_REGISTRY = ContextLoaderRegistry()
-CONTEXT_LOADER_REGISTRY.register(mode=ContextBuildStrategy.NUMBER, loader=_load_number)
-CONTEXT_LOADER_REGISTRY.register(mode=ContextBuildStrategy.DAY, loader=_load_day)

backend/src/core/agentscope/runtime/orchestrator.py

@@ -6,7 +6,8 @@ from ag_ui.core.types import RunAgentInput
 from agentscope.message import Msg
 from core.agentscope.runtime.runner import AgentScopeRunner
 from core.logging import get_logger
-from schemas.automation.config import AutomationJobConfig
+from schemas.automation import RuntimeConfig
+from schemas.memories import MemoryListResponse
 from schemas.user import UserContext
 
 logger = get_logger("core.agentscope.runtime.orchestrator")
@@ -24,8 +25,8 @@ class RunnerLike(Protocol):
         context_messages: list[Msg],
         pipeline: PipelineLike,
         run_input: RunAgentInput,
-        system_agent_mode: str,
-        memory_job_config: AutomationJobConfig | None,
+        runtime_config: RuntimeConfig,
+        memories: MemoryListResponse | None,
     ) -> dict[str, Any]: ...
 
 
@@ -48,8 +49,8 @@ class AgentScopeRuntimeOrchestrator:
         run_input: RunAgentInput,
         context_messages: list[Msg],
         user_context: UserContext,
-        system_agent_mode: str,
-        memory_job_config: AutomationJobConfig | None = None,
+        runtime_config: RuntimeConfig,
+        memories: MemoryListResponse | None = None,
     ) -> dict[str, Any]:
         thread_id = run_input.thread_id
         run_id = run_input.run_id
@@ -68,8 +69,8 @@ class AgentScopeRuntimeOrchestrator:
                 context_messages=context_messages,
                 pipeline=self._pipeline,
                 run_input=run_input,
-                system_agent_mode=system_agent_mode,
-                memory_job_config=memory_job_config,
+                runtime_config=runtime_config,
+                memories=memories,
             )
 
             await self._pipeline.emit(

backend/src/core/agentscope/runtime/pipeline_registry.py

@@ -1,38 +0,0 @@
-from __future__ import annotations
-
-from core.agentscope.schemas.pipeline_spec import ExecutorKind, PipelineSpec, StageSpec
-from schemas.agent.system_agent import AgentType
-
-
-def build_default_pipeline_spec(*, mode: str) -> PipelineSpec:
-    normalized = mode.strip().lower()
-    if normalized == "worker":
-        return PipelineSpec(
-            mode="worker",
-            stages=[
-                StageSpec(
-                    stage_name="router",
-                    agent_type=AgentType.ROUTER,
-                    executor_kind=ExecutorKind.SINGLE_SHOT,
-                ),
-                StageSpec(
-                    stage_name="worker",
-                    agent_type=AgentType.WORKER,
-                    executor_kind=ExecutorKind.REACT,
-                ),
-            ],
-        )
-
-    if normalized == "memory":
-        return PipelineSpec(
-            mode="memory",
-            stages=[
-                StageSpec(
-                    stage_name="memory",
-                    agent_type=AgentType.MEMORY,
-                    executor_kind=ExecutorKind.REACT,
-                )
-            ],
-        )
-
-    raise ValueError(f"unsupported pipeline mode: {normalized}")

backend/src/core/agentscope/runtime/registry_builder.py

@@ -1,22 +0,0 @@
-from __future__ import annotations
-
-from core.agentscope.schemas.consumer_registry import (
-    AgentConsumerBinding,
-    ConsumerRegistry,
-)
-
-
-def build_consumer_registry(
-    *,
-    system_agent_configs: dict[str, dict[str, object]],
-) -> ConsumerRegistry:
-    bindings: list[AgentConsumerBinding] = []
-    for agent_type, payload in system_agent_configs.items():
-        config_obj = payload.get("config") if isinstance(payload, dict) else None
-        if not isinstance(config_obj, dict):
-            raise ValueError(f"invalid system agent config: {agent_type}")
-        raw_bit = config_obj.get("visibility_consumer_bit")
-        if not isinstance(raw_bit, int):
-            raise ValueError(f"visibility_consumer_bit missing for agent: {agent_type}")
-        bindings.append(AgentConsumerBinding(agent_type=agent_type, bit=raw_bit))
-    return ConsumerRegistry(bindings=bindings)

backend/src/core/agentscope/runtime/runner.py

@@ -12,12 +12,12 @@ from agentscope.message import Msg
 from agentscope.model import OpenAIChatModel
 from core.agentscope.prompts.agent_prompt import build_worker_contract_prompt
 from core.agentscope.prompts.system_prompt import build_system_prompt
-from core.agentscope.runtime.pipeline_registry import build_default_pipeline_spec
+from core.agentscope.schemas.agui_input import extract_latest_user_payload
 from core.agentscope.runtime.json_react_agent import JsonReActAgent
 from core.agentscope.runtime.model_tracking import TrackingChatModel
 from core.agentscope.runtime.stage_emitter import PipelineStageEmitter
-from core.agentscope.runtime.tool_selection_registry import TOOL_SELECTION_REGISTRY
-from core.agentscope.tools.toolkit import build_stage_toolkit
+from core.agentscope.tools.tool_config import AgentTool
+from core.agentscope.tools.toolkit import build_toolkit
 from core.agentscope.utils import (
     finalize_json_response,
     patch_agentscope_json_repair_compat,
@@ -31,19 +31,17 @@ from schemas.agent.forwarded_props import (
     ClientTimeContext,
     parse_forwarded_props_client_time,
 )
-from schemas.automation.config import AutomationJobConfig
 from schemas.agent.runtime_models import (
-    AgentOutput,
     RouterAgentOutput,
     WorkerAgentOutputLite,
     resolve_worker_output_model,
 )
 from schemas.agent.system_agent import (
     AgentType,
-    ContextMessagesConfig,
-    ContextBuildStrategy,
     SystemAgentLLMConfig,
 )
+from schemas.automation import RuntimeConfig
+from schemas.memories import MemoryListResponse
 from schemas.user import UserContext
 from services.litellm.service import LiteLLMService
 from sqlalchemy import select
@@ -53,16 +51,6 @@ if TYPE_CHECKING:
     from core.agentscope.runtime.orchestrator import PipelineLike
 
 
-@dataclass(frozen=True)
-class SystemAgentRuntimeConfig:
-    agent_type: AgentType
-    model_code: str
-    api_base_url: str
-    api_key: str
-    llm_config: SystemAgentLLMConfig
-    extra_context: str | None = None
-
-
 @dataclass(frozen=True)
 class StageExecutionResult:
     message: Msg
@@ -82,96 +70,63 @@ class AgentScopeRunner:
         context_messages: list[Msg],
         pipeline: PipelineLike,
         run_input: RunAgentInput,
-        system_agent_mode: str,
-        memory_job_config: AutomationJobConfig | None = None,
+        runtime_config: RuntimeConfig,
+        memories: MemoryListResponse | None = None,
     ) -> dict[str, Any]:
         owner_id = UUID(user_context.id)
         runtime_client_time = self._resolve_runtime_client_time(run_input=run_input)
-        pipeline_spec = build_default_pipeline_spec(mode=system_agent_mode)
-        stage_agent_types = [stage.agent_type for stage in pipeline_spec.stages]
 
         async with AsyncSessionLocal() as session:
-            if stage_agent_types == [AgentType.ROUTER, AgentType.WORKER]:
-                router_config = await self._load_stage_config(
-                    session=session,
-                    agent_type=AgentType.ROUTER,
-                )
-                worker_config = await self._load_stage_config(
-                    session=session,
-                    agent_type=AgentType.WORKER,
-                )
-                worker_toolkit = self._build_stage_toolkit(
-                    session=session,
-                    owner_id=owner_id,
-                    stage_config=worker_config,
-                )
-                router_output = await self._execute_router_step(
-                    pipeline=pipeline,
-                    run_input=run_input,
-                    user_context=user_context,
-                    context_messages=context_messages,
-                    stage_config=router_config,
-                    runtime_client_time=runtime_client_time,
-                )
-                worker_output = await self._execute_worker_step(
-                    pipeline=pipeline,
-                    run_input=run_input,
-                    user_context=user_context,
-                    router_output=router_output,
-                    toolkit=worker_toolkit,
-                    stage_config=worker_config,
-                    runtime_client_time=runtime_client_time,
-                )
-                return {
-                    "router": router_output.model_dump(mode="json", exclude_none=True),
-                    "worker": worker_output.model_dump(mode="json", exclude_none=True),
-                }
-
-            if stage_agent_types[0] == AgentType.MEMORY:
-                if memory_job_config is None:
-                    raise RuntimeError("memory job config is required")
-                stage_config = await self._build_memory_stage_config(
-                    session=session,
-                    memory_job_config=memory_job_config,
-                )
-            else:
-                stage_config = await self._load_stage_config(
-                    session=session,
-                    agent_type=stage_agent_types[0],
-                )
-            stage_toolkit = self._build_stage_toolkit(
+            router_config = await self._load_stage_config(
+                session=session,
+                agent_type=AgentType.ROUTER,
+            )
+            worker_config = await self._load_stage_config(
+                session=session,
+                agent_type=AgentType.WORKER,
+            )
+            worker_toolkit = self._build_toolkit(
                 session=session,
                 owner_id=owner_id,
-                stage_config=stage_config,
+                enabled_tools=runtime_config.enabled_tools,
             )
-            stage_output = await self._execute_single_stage_step(
+
+            router_output = await self._execute_router_step(
                 pipeline=pipeline,
                 run_input=run_input,
                 user_context=user_context,
-                input_messages=context_messages,
-                toolkit=stage_toolkit,
-                stage_config=stage_config,
+                context_messages=context_messages,
+                stage_config=router_config,
                 runtime_client_time=runtime_client_time,
+                memories=memories,
+            )
+            worker_output = await self._execute_worker_step(
+                pipeline=pipeline,
+                run_input=run_input,
+                user_context=user_context,
+                router_output=router_output,
+                toolkit=worker_toolkit,
+                stage_config=worker_config,
+                runtime_client_time=runtime_client_time,
+                memories=memories,
             )
             return {
-                stage_config.agent_type.value: stage_output.model_dump(
-                    mode="json", exclude_none=True
-                ),
+                "router": router_output.model_dump(mode="json", exclude_none=True),
+                "worker": worker_output.model_dump(mode="json", exclude_none=True),
             }
 
-    def _build_stage_toolkit(
+    def _build_toolkit(
         self,
         *,
         session: AsyncSession,
         owner_id: UUID,
-        stage_config: SystemAgentRuntimeConfig,
+        enabled_tools: list[AgentTool],
     ) -> Any:
-        enabled_tool_names = TOOL_SELECTION_REGISTRY.resolve(stage_config=stage_config)
-        return build_stage_toolkit(
-            agent_type=stage_config.agent_type,
+        tool_names = [t.value for t in enabled_tools] if enabled_tools else []
+        return build_toolkit(
             session=session,
             owner_id=owner_id,
-            enabled_tool_names=enabled_tool_names,
+            enabled_tool_names=set(tool_names) if tool_names else None,
         )
 
     async def _load_stage_config(
@@ -179,124 +134,6 @@ class AgentScopeRunner:
         *,
         session: AsyncSession,
         agent_type: AgentType,
-    ) -> SystemAgentRuntimeConfig:
-        return await self._load_system_agent_config(
-            session=session,
-            agent_type=agent_type,
-        )
-
-    async def _execute_router_step(
-        self,
-        *,
-        pipeline: PipelineLike,
-        run_input: RunAgentInput,
-        user_context: UserContext,
-        context_messages: list[Msg],
-        stage_config: SystemAgentRuntimeConfig,
-        runtime_client_time: ClientTimeContext | None,
-    ) -> RouterAgentOutput:
-        await self._emit_step_event(
-            pipeline=pipeline,
-            run_input=run_input,
-            step_name=AgentType.ROUTER.value,
-            event_type="STEP_STARTED",
-        )
-        router_result = await self._run_router_stage(
-            user_context=user_context,
-            context_messages=context_messages,
-            stage_config=stage_config,
-            runtime_client_time=runtime_client_time,
-        )
-        router_output = RouterAgentOutput.model_validate(router_result.payload)
-        await self._emit_step_event(
-            pipeline=pipeline,
-            run_input=run_input,
-            step_name=AgentType.ROUTER.value,
-            event_type="STEP_FINISHED",
-        )
-        return router_output
-
-    async def _execute_worker_step(
-        self,
-        *,
-        pipeline: PipelineLike,
-        run_input: RunAgentInput,
-        user_context: UserContext,
-        router_output: RouterAgentOutput,
-        toolkit: Any,
-        stage_config: SystemAgentRuntimeConfig,
-        runtime_client_time: ClientTimeContext | None,
-    ) -> WorkerAgentOutputLite:
-        worker_output_model = resolve_worker_output_model(router_output.ui.ui_mode)
-        await self._emit_step_event(
-            pipeline=pipeline,
-            run_input=run_input,
-            step_name=AgentType.WORKER.value,
-            event_type="STEP_STARTED",
-        )
-        worker_result = await self._run_worker_stage(
-            user_context=user_context,
-            input_messages=self._build_worker_input_messages(
-                router_output=router_output
-            ),
-            toolkit=toolkit,
-            run_input=run_input,
-            stage_config=stage_config,
-            worker_output_model=worker_output_model,
-            pipeline=pipeline,
-            runtime_client_time=runtime_client_time,
-        )
-        worker_output = worker_output_model.model_validate(worker_result.payload)
-        await self._emit_step_event(
-            pipeline=pipeline,
-            run_input=run_input,
-            step_name=AgentType.WORKER.value,
-            event_type="STEP_FINISHED",
-        )
-        return worker_output
-
-    async def _execute_single_stage_step(
-        self,
-        *,
-        pipeline: PipelineLike,
-        run_input: RunAgentInput,
-        user_context: UserContext,
-        input_messages: list[Msg],
-        toolkit: Any,
-        stage_config: SystemAgentRuntimeConfig,
-        runtime_client_time: ClientTimeContext | None,
-    ) -> AgentOutput:
-        step_name = stage_config.agent_type.value
-        await self._emit_step_event(
-            pipeline=pipeline,
-            run_input=run_input,
-            step_name=step_name,
-            event_type="STEP_STARTED",
-        )
-        stage_result = await self._run_worker_stage(
-            user_context=user_context,
-            input_messages=input_messages,
-            toolkit=toolkit,
-            run_input=run_input,
-            stage_config=stage_config,
-            worker_output_model=AgentOutput,
-            pipeline=pipeline,
-            runtime_client_time=runtime_client_time,
-        )
-        stage_output = AgentOutput.model_validate(stage_result.payload)
-        await self._emit_step_event(
-            pipeline=pipeline,
-            run_input=run_input,
-            step_name=step_name,
-            event_type="STEP_FINISHED",
-        )
-        return stage_output
-
-    async def _load_system_agent_config(
-        self,
-        *,
-        session: AsyncSession,
-        agent_type: AgentType,
     ) -> SystemAgentRuntimeConfig:
         stmt = (
             select(SystemAgents, Llm, LlmFactory)
@@ -320,63 +157,80 @@ class AgentScopeRunner:
             extra_context=None,
         )
 
-    async def _build_memory_stage_config(
+    async def _execute_router_step(
         self,
         *,
-        session: AsyncSession,
-        memory_job_config: AutomationJobConfig,
-    ) -> SystemAgentRuntimeConfig:
-        stmt = (
-            select(Llm, LlmFactory)
-            .join(LlmFactory, Llm.factory_id == LlmFactory.id)
-            .where(Llm.model_code == memory_job_config.model_code)
+        pipeline: PipelineLike,
+        run_input: RunAgentInput,
+        user_context: UserContext,
+        context_messages: list[Msg],
+        stage_config: SystemAgentRuntimeConfig,
+        runtime_client_time: ClientTimeContext | None,
+        memories: MemoryListResponse | None,
+    ) -> RouterAgentOutput:
+        await self._emit_step_event(
+            pipeline=pipeline,
+            run_input=run_input,
+            step_name=AgentType.ROUTER.value,
+            event_type="STEP_STARTED",
         )
-        row = (await session.execute(stmt)).one_or_none()
-        if row is None:
-            raise RuntimeError(
-                f"memory model not found: {memory_job_config.model_code}"
-            )
-        llm, factory = row
-        llm_config = SystemAgentLLMConfig(
-            temperature=0.7,
-            max_tokens=None,
-            timeout_seconds=30,
-            context_messages=ContextMessagesConfig(
-                mode=(
-                    ContextBuildStrategy.DAY
-                    if memory_job_config.context.window_mode.value == "day"
-                    else ContextBuildStrategy.NUMBER
-                ),
-                count=memory_job_config.context.window_count,
-            ),
-            enabled_tools=memory_job_config.enabled_tools,
+        router_result = await self._run_router_stage(
+            user_context=user_context,
+            context_messages=context_messages,
+            stage_config=stage_config,
+            runtime_client_time=runtime_client_time,
+            memories=memories,
+            run_input=run_input,
         )
-        return SystemAgentRuntimeConfig(
-            agent_type=AgentType.MEMORY,
-            model_code=llm.model_code,
-            api_base_url=factory.request_url,
-            api_key=self._resolve_provider_api_key(factory_name=factory.name),
-            llm_config=llm_config,
-            extra_context=(
-                f"[Memory Input Template]\n{memory_job_config.input_template.strip()}"
-            ),
+        router_output = RouterAgentOutput.model_validate(router_result.payload)
+        await self._emit_step_event(
+            pipeline=pipeline,
+            run_input=run_input,
+            step_name=AgentType.ROUTER.value,
+            event_type="STEP_FINISHED",
         )
+        return router_output
 
-    @staticmethod
-    def _resolve_provider_api_key(*, factory_name: str) -> str:
-        normalized_factory_name = factory_name.strip().upper()
-        if normalized_factory_name == "VOLCENGINE":
-            normalized_factory_name = "ARK"
-
-        provider_keys = {
-            str(key).strip().upper(): str(value).strip()
-            for key, value in config.llm.provider_keys.items()
-            if str(value).strip()
-        }
-        api_key = provider_keys.get(normalized_factory_name, "")
-        if not api_key:
-            raise RuntimeError(f"provider api key missing for factory: {factory_name}")
-        return api_key
+    async def _execute_worker_step(
+        self,
+        *,
+        pipeline: PipelineLike,
+        run_input: RunAgentInput,
+        user_context: UserContext,
+        router_output: RouterAgentOutput,
+        toolkit: Any,
+        stage_config: SystemAgentRuntimeConfig,
+        runtime_client_time: ClientTimeContext | None,
+        memories: MemoryListResponse | None,
+    ) -> WorkerAgentOutputLite:
+        worker_output_model = resolve_worker_output_model(router_output.ui.ui_mode)
+        await self._emit_step_event(
+            pipeline=pipeline,
+            run_input=run_input,
+            step_name=AgentType.WORKER.value,
+            event_type="STEP_STARTED",
+        )
+        worker_result = await self._run_worker_stage(
+            user_context=user_context,
+            input_messages=self._build_worker_input_messages(
+                router_output=router_output
+            ),
+            toolkit=toolkit,
+            run_input=run_input,
+            stage_config=stage_config,
+            worker_output_model=worker_output_model,
+            pipeline=pipeline,
+            runtime_client_time=runtime_client_time,
+            memories=memories,
+        )
+        worker_output = worker_output_model.model_validate(worker_result.payload)
+        await self._emit_step_event(
+            pipeline=pipeline,
+            run_input=run_input,
+            step_name=AgentType.WORKER.value,
+            event_type="STEP_FINISHED",
+        )
+        return worker_output
 
     async def _run_router_stage(
         self,
@@ -385,7 +239,13 @@ class AgentScopeRunner:
         context_messages: list[Msg],
         stage_config: SystemAgentRuntimeConfig,
         runtime_client_time: ClientTimeContext | None,
+        memories: MemoryListResponse | None,
+        run_input: RunAgentInput,
     ) -> StageExecutionResult:
+        messages_for_router = self._build_router_messages(
+            context_messages=context_messages,
+            run_input=run_input,
+        )
         tracking_model = self._build_model(stage_config=stage_config)
         response, payload = await finalize_json_response(
             model=tracking_model,
@@ -400,10 +260,11 @@ class AgentScopeRunner:
                         now_utc=datetime.now(timezone.utc),
                         runtime_client_time=runtime_client_time,
                         tools=None,
+                        memories=memories,
                     ),
                     "system",
                 ),
-                *context_messages,
+                *messages_for_router,
             ],
             output_model=RouterAgentOutput,
             retries=0,
@@ -423,6 +284,30 @@ class AgentScopeRunner:
             ),
         )
 
+    def _build_router_messages(
+        self,
+        *,
+        context_messages: list[Msg],
+        run_input: RunAgentInput,
+    ) -> list[Msg]:
+        if context_messages:
+            last = context_messages[-1]
+            if last.role == "user":
+                return context_messages
+
+        user_text, user_blocks = extract_latest_user_payload(run_input)
+        if (
+            user_blocks
+            and isinstance(user_blocks[0], dict)
+            and user_blocks[0].get("type") == "text"
+        ):
+            content: Any = user_text
+        else:
+            content = user_blocks
+
+        user_msg = Msg(name="user", role="user", content=content)
+        return [user_msg, *context_messages]
+
     async def _run_worker_stage(
         self,
         *,
@@ -434,6 +319,7 @@ class AgentScopeRunner:
         worker_output_model: type[WorkerAgentOutputLite],
         pipeline: PipelineLike,
         runtime_client_time: ClientTimeContext | None,
+        memories: MemoryListResponse | None,
     ) -> StageExecutionResult:
         tracking_model = self._build_model(stage_config=stage_config)
         emitter = PipelineStageEmitter(
@@ -454,6 +340,7 @@ class AgentScopeRunner:
                 runtime_client_time=runtime_client_time,
                 extra_context=stage_config.extra_context,
                 tools=None,
+                memories=memories,
             ),
             toolkit=toolkit,
             model=tracking_model,
@@ -553,5 +440,31 @@ class AgentScopeRunner:
             getattr(run_input, "forwarded_props", None)
         )
 
+    @staticmethod
+    def _resolve_provider_api_key(*, factory_name: str) -> str:
+        normalized_factory_name = factory_name.strip().upper()
+        if normalized_factory_name == "VOLCENGINE":
+            normalized_factory_name = "ARK"
+
+        provider_keys = {
+            str(key).strip().upper(): str(value).strip()
+            for key, value in config.llm.provider_keys.items()
+            if str(value).strip()
+        }
+        api_key = provider_keys.get(normalized_factory_name, "")
+        if not api_key:
+            raise RuntimeError(f"provider api key missing for factory: {factory_name}")
+        return api_key
+
+
+@dataclass(frozen=True)
+class SystemAgentRuntimeConfig:
+    agent_type: AgentType
+    model_code: str
+    api_base_url: str
+    api_key: str
+    llm_config: SystemAgentLLMConfig
+    extra_context: str | None = None
+
 
 AgentScopeReActRunner = AgentScopeRunner

backend/src/core/agentscope/runtime/tasks.py

@@ -12,31 +12,28 @@ from core.agentscope.events import (
     RedisStreamBus,
     SqlAlchemyEventStore,
 )
-from core.agentscope.services.context_service import AgentContextService
 from core.agentscope.runtime.orchestrator import AgentScopeRuntimeOrchestrator
-from core.agentscope.runtime.pipeline_registry import build_default_pipeline_spec
 from core.agentscope.schemas.agui_input import parse_run_input
+from core.agentscope.services.context_service import AgentContextService
 from core.auth.models import CurrentUser
 from core.config.settings import config
 from core.db.session import AsyncSessionLocal
 from core.logging import get_logger
 from core.taskiq.app import worker_agent_broker, worker_automation_broker
-from schemas.agent.visibility import SystemVisibilityBit, bit_mask
-from schemas.automation.config import AutomationJobConfig
+from schemas.automation import MemoryContextConfig, RuntimeConfig
+from schemas.memories import MemoryListResponse
 from schemas.messages.chat_message import (
     AgentChatMessageMetadata,
     extract_user_message_attachments,
 )
-from schemas.agent.forwarded_props import parse_forwarded_props_agent_type
 from schemas.user import UserContext
 from services.base.redis import get_or_init_redis_client
 from services.base.supabase import supabase_service
 from v1.agent.repository import AgentRepository
-from v1.memory.repository import MemoryRepository
-from v1.memory.service import MemoryService
+from v1.memories.repository import MemoriesRepository
+from v1.memories.service import MemoriesService
 from v1.users.dependencies import get_user_service
 
-
 logger = get_logger("core.agentscope.runtime.tasks")
 _MAX_CONTEXT_ATTACHMENTS = 3
 
@@ -86,29 +83,14 @@ async def _build_recent_context_messages(
     *,
     session: Any,
     thread_id: str,
-    context_mode: str,
-    memory_job_config: AutomationJobConfig | None = None,
+    context_config: "MemoryContextConfig",
 ) -> list[Msg]:
     context_service = AgentContextService(repository=AgentRepository(session))
-    if memory_job_config is not None:
-        visibility_mask = bit_mask(bit=int(SystemVisibilityBit.UI_HISTORY))
-        if memory_job_config.context.window_mode.value == "day":
-            result = await context_service.load_by_day_window(
-                thread_id=thread_id,
-                day_count=memory_job_config.context.window_count,
-                visibility_mask=visibility_mask,
-            )
-        else:
-            result = await context_service.load_by_user_message_window(
-                thread_id=thread_id,
-                user_message_limit=memory_job_config.context.window_count,
-                visibility_mask=visibility_mask,
-            )
-    else:
-        result = await context_service.load_context_messages(
-            thread_id=thread_id,
-            system_agent_mode=context_mode,
-        )
+    result = await context_service.load_context_messages(
+        thread_id=thread_id,
+        context_config=context_config,
+    )
+
     if not result:
         return []
 
@@ -193,6 +175,7 @@ async def run_agentscope_task(command: dict[str, Any]) -> dict[str, object]:
     command_type = str(command.get("command", "run")).strip().lower()
     raw_owner_id = command.get("owner_id")
     run_input_raw = command.get("run_input")
+    runtime_config_raw = command.get("runtime_config")
 
     if not isinstance(raw_owner_id, str) or not raw_owner_id.strip():
         raise ValueError("owner_id is required")
@@ -200,15 +183,7 @@ async def run_agentscope_task(command: dict[str, Any]) -> dict[str, object]:
         raise ValueError("run_input is required")
 
     run_input = parse_run_input(run_input_raw)
-    system_agent_mode = parse_forwarded_props_agent_type(
-        getattr(run_input, "forwarded_props", None)
-    )
-    raw_automation_job_id = command.get("automation_job_id")
-    if system_agent_mode == "memory" and (
-        not isinstance(raw_automation_job_id, str) or not raw_automation_job_id
-    ):
-        raise ValueError("automation_job_id is required for memory mode")
-    pipeline_spec = build_default_pipeline_spec(mode=system_agent_mode)
+    runtime_config = RuntimeConfig.model_validate(runtime_config_raw or {})
     thread_id = run_input.thread_id
     run_id = run_input.run_id
     owner_id = UUID(raw_owner_id)
@@ -220,15 +195,10 @@ async def run_agentscope_task(command: dict[str, Any]) -> dict[str, object]:
 
     async with AsyncSessionLocal() as session:
         user_context = await _build_user_context(owner_id=owner_id, session=session)
-        memory_job_config: AutomationJobConfig | None = None
-        if system_agent_mode == "memory":
-            assert isinstance(raw_automation_job_id, str)
-            job_uuid = UUID(raw_automation_job_id)
-            memory_service = MemoryService(MemoryRepository(session))
-            memory_job_config = await memory_service.get_memory_job_config(
-                job_id=job_uuid,
-                owner_id=owner_id,
-            )
+        memories_service = MemoriesService(MemoriesRepository(session))
+        memories: MemoryListResponse = await memories_service.get_all_memories(
+            owner_id=owner_id
+        )
 
         redis_client = await get_or_init_redis_client()
         bus = RedisStreamBus(
@@ -251,16 +221,15 @@ async def run_agentscope_task(command: dict[str, Any]) -> dict[str, object]:
         context_messages = await _build_recent_context_messages(
             session=session,
             thread_id=thread_id,
-            context_mode=pipeline_spec.stages[0].agent_type.value,
-            memory_job_config=memory_job_config,
+            context_config=runtime_config.context,
         )
 
         await runtime.run(
             run_input=run_input,
             context_messages=context_messages,
             user_context=user_context,
-            system_agent_mode=system_agent_mode,
-            memory_job_config=memory_job_config,
+            runtime_config=runtime_config,
+            memories=memories,
         )
     logger.info(
         "agentscope runtime task completed",

backend/src/core/agentscope/runtime/tool_selection_registry.py

@@ -1,41 +0,0 @@
-from __future__ import annotations
-
-from collections.abc import Callable
-from typing import Any
-
-from core.agentscope.tools.tool_config import resolve_tool_function_names
-from schemas.agent.system_agent import AgentType
-
-ToolNameResolver = Callable[[Any], set[str] | None]
-
-
-class ToolSelectionRegistry:
-    def __init__(self) -> None:
-        self._resolvers: dict[AgentType, ToolNameResolver] = {}
-
-    def register(self, *, agent_type: AgentType, resolver: ToolNameResolver) -> None:
-        self._resolvers[agent_type] = resolver
-
-    def resolve(self, *, stage_config: Any) -> set[str] | None:
-        resolver = self._resolvers.get(stage_config.agent_type)
-        if resolver is None:
-            return None
-        return resolver(stage_config)
-
-
-def _default_tool_resolver(stage_config: Any) -> set[str] | None:
-    enabled_tools = getattr(stage_config.llm_config, "enabled_tools", [])
-    if not enabled_tools:
-        return None
-    return resolve_tool_function_names(set(enabled_tools))
-
-
-TOOL_SELECTION_REGISTRY = ToolSelectionRegistry()
-TOOL_SELECTION_REGISTRY.register(
-    agent_type=AgentType.WORKER,
-    resolver=_default_tool_resolver,
-)
-TOOL_SELECTION_REGISTRY.register(
-    agent_type=AgentType.MEMORY,
-    resolver=_default_tool_resolver,
-)

backend/src/core/agentscope/schemas/consumer_registry.py

@@ -1,44 +0,0 @@
-from __future__ import annotations
-
-from pydantic import BaseModel, ConfigDict, Field, field_validator, model_validator
-
-
-class AgentConsumerBinding(BaseModel):
-    model_config = ConfigDict(extra="forbid")
-
-    agent_type: str = Field(..., min_length=1, max_length=64)
-    bit: int = Field(..., ge=16, le=63)
-
-    @field_validator("agent_type")
-    @classmethod
-    def _normalize_agent_type(cls, value: str) -> str:
-        normalized = value.strip().lower()
-        if not normalized:
-            raise ValueError("agent_type must not be empty")
-        return normalized
-
-
-class ConsumerRegistry(BaseModel):
-    model_config = ConfigDict(extra="forbid")
-
-    bindings: list[AgentConsumerBinding] = Field(default_factory=list)
-
-    @model_validator(mode="after")
-    def _validate_unique_bindings(self) -> "ConsumerRegistry":
-        by_agent: set[str] = set()
-        by_bit: set[int] = set()
-        for item in self.bindings:
-            if item.agent_type in by_agent:
-                raise ValueError(f"duplicate agent_type binding: {item.agent_type}")
-            if item.bit in by_bit:
-                raise ValueError(f"duplicate visibility bit binding: {item.bit}")
-            by_agent.add(item.agent_type)
-            by_bit.add(item.bit)
-        return self
-
-    def resolve_agent_bit(self, *, agent_type: str) -> int:
-        target = agent_type.strip().lower()
-        for item in self.bindings:
-            if item.agent_type == target:
-                return item.bit
-        raise ValueError(f"agent visibility bit not configured: {target}")

backend/src/core/agentscope/schemas/pipeline_spec.py

@@ -1,43 +0,0 @@
-from __future__ import annotations
-
-from enum import Enum
-
-from pydantic import BaseModel, ConfigDict, Field, field_validator
-
-from schemas.agent.system_agent import AgentType
-
-
-class ExecutorKind(str, Enum):
-    SINGLE_SHOT = "single_shot"
-    REACT = "react"
-
-
-class StageSpec(BaseModel):
-    model_config = ConfigDict(extra="forbid")
-
-    stage_name: str = Field(..., min_length=1, max_length=64)
-    agent_type: AgentType
-    executor_kind: ExecutorKind
-
-    @field_validator("stage_name")
-    @classmethod
-    def _normalize_stage_name(cls, value: str) -> str:
-        normalized = value.strip().lower()
-        if not normalized:
-            raise ValueError("stage_name must not be empty")
-        return normalized
-
-
-class PipelineSpec(BaseModel):
-    model_config = ConfigDict(extra="forbid")
-
-    mode: str = Field(..., min_length=1, max_length=64)
-    stages: list[StageSpec] = Field(..., min_length=1)
-
-    @field_validator("mode")
-    @classmethod
-    def _normalize_mode(cls, value: str) -> str:
-        normalized = value.strip().lower()
-        if not normalized:
-            raise ValueError("mode must not be empty")
-        return normalized

backend/src/core/agentscope/services/context_service.py

@@ -1,14 +1,15 @@
 from __future__ import annotations
 
+from collections.abc import Awaitable, Callable
 from datetime import date
-from typing import Protocol
+from typing import Any, Protocol
 
-from core.agentscope.runtime.context_loader_registry import CONTEXT_LOADER_REGISTRY
-from schemas.agent.system_agent import SystemAgentLLMConfig
 from schemas.agent.visibility import SystemVisibilityBit, bit_mask
 
+from schemas.automation import ContextWindowMode, MemoryContextConfig
+
+
 _DEFAULT_CONTEXT_WINDOW_USER_MESSAGES = 20
-_DEFAULT_ROUTER_CONTEXT_DAY_COUNT = 20
 
 
 class ContextRepositoryLike(Protocol):
@@ -28,9 +29,53 @@ class ContextRepositoryLike(Protocol):
         visibility_mask: int | None = None,
     ) -> list[dict[str, object]]: ...
 
-    async def get_system_agent_config(
-        self, *, agent_type: str
-    ) -> dict[str, object] | None: ...
+
+ContextLoader = Callable[[Any, str, int, int], Awaitable[dict[str, object] | None]]
+
+
+class ContextLoaderRegistry:
+    def __init__(self) -> None:
+        self._loaders: dict[ContextWindowMode, ContextLoader] = {}
+
+    def register(self, *, mode: ContextWindowMode, loader: ContextLoader) -> None:
+        self._loaders[mode] = loader
+
+    def resolve(self, *, mode: ContextWindowMode) -> ContextLoader:
+        loader = self._loaders.get(mode)
+        if loader is None:
+            raise ValueError(f"unsupported context mode: {mode.value}")
+        return loader
+
+
+async def _load_number(
+    service: Any,
+    thread_id: str,
+    count: int,
+    visibility_mask: int,
+) -> dict[str, object] | None:
+    return await service.load_by_user_message_window(
+        thread_id=thread_id,
+        user_message_limit=max(count, 1),
+        visibility_mask=visibility_mask,
+    )
+
+
+async def _load_day(
+    service: Any,
+    thread_id: str,
+    count: int,
+    visibility_mask: int,
+) -> dict[str, object] | None:
+    return await service.load_by_day_window(
+        thread_id=thread_id,
+        day_count=max(count, 1),
+        visibility_mask=visibility_mask,
+    )
+
+
+CONTEXT_LOADER_REGISTRY = ContextLoaderRegistry()
+CONTEXT_LOADER_REGISTRY.register(mode=ContextWindowMode.NUMBER, loader=_load_number)
+CONTEXT_LOADER_REGISTRY.register(mode=ContextWindowMode.DAY, loader=_load_day)
 
 
 class AgentContextService:
@@ -41,32 +86,16 @@ class AgentContextService:
         self,
         *,
         thread_id: str,
-        system_agent_mode: str,
+        context_config: MemoryContextConfig,
     ) -> dict[str, object] | None:
-        mode = system_agent_mode.strip().lower() if system_agent_mode else "worker"
-        runtime_config = await self._repository.get_system_agent_config(agent_type=mode)
-        raw_llm_config: dict[str, object] = {}
-        if isinstance(runtime_config, dict):
-            raw_config = runtime_config.get("config")
-            if isinstance(raw_config, dict):
-                raw_llm_config = raw_config
-
-        if mode == "router" and not raw_llm_config:
-            raw_llm_config = {
-                "context_messages": {
-                    "mode": "day",
-                    "count": _DEFAULT_ROUTER_CONTEXT_DAY_COUNT,
-                }
-            }
-
-        normalized_config = self._normalize_system_agent_config(raw_llm_config)
-        context_config = normalized_config.context_messages
         visibility_mask = bit_mask(bit=int(SystemVisibilityBit.CONTEXT_ASSEMBLY))
-        context_loader = CONTEXT_LOADER_REGISTRY.resolve(mode=context_config.mode)
+        context_loader = CONTEXT_LOADER_REGISTRY.resolve(
+            mode=context_config.window_mode
+        )
         return await context_loader(
             self,
             thread_id,
-            context_config.count,
+            context_config.window_count,
             visibility_mask,
         )
 
@@ -114,22 +143,6 @@ class AgentContextService:
             return None
         return {"messages": messages}
 
-    def _normalize_system_agent_config(
-        self,
-        raw_config: dict[str, object],
-    ) -> SystemAgentLLMConfig:
-        default_payload = {
-            "context_messages": {
-                "mode": "number",
-                "count": _DEFAULT_CONTEXT_WINDOW_USER_MESSAGES,
-            },
-            "enabled_tools": [],
-        }
-        if not raw_config:
-            return SystemAgentLLMConfig.model_validate(default_payload)
-        merged = {**default_payload, **raw_config}
-        return SystemAgentLLMConfig.model_validate(merged)
-
     def _parse_history_day(self, value: object) -> date | None:
         if isinstance(value, date):
             return value

backend/src/core/agentscope/tools/toolkit.py

@@ -33,7 +33,6 @@ AGENT_TYPE_TO_DEFAULT_TOOLS: dict[AgentType, set[str]] = {
         "calendar_share",
         "user_lookup",
     },
-    AgentType.MEMORY: {"calendar_read", "user_lookup"},
 }
 
 

backend/src/core/automation/scheduler.py

@@ -1,263 +1,15 @@
 from __future__ import annotations
 
-from dataclasses import dataclass
-from datetime import datetime, timedelta, timezone
-from typing import Protocol
-from uuid import UUID, uuid4
-
-from sqlalchemy import select
-from sqlalchemy.ext.asyncio import AsyncSession
+from datetime import datetime, timezone
+from uuid import UUID
 
 from core.config.settings import config
 from core.logging import get_logger
-from models.agent_chat_session import AgentChatSession, SessionType
-from models.automation_jobs import AutomationJob, ScheduleType
-from schemas.automation.config import AutomationJobConfig
-from schemas.automation.scheduler import DueAutomationJob, SchedulerDispatchCommand
+from schemas.automation import RuntimeConfig
 
 logger = get_logger("core.automation.scheduler")
 
 
-class _BulkQueueAdapter:
-    async def enqueue(
-        self,
-        *,
-        command: dict[str, object],
-        dedup_key: str | None,
-    ) -> str:
-        del dedup_key
-        from core.agentscope.runtime.tasks import run_command_task_bulk
-
-        result = await run_command_task_bulk.kiq(command)
-        return str(result.task_id)
-
-
-class QueueLike(Protocol):
-    async def enqueue(
-        self,
-        *,
-        command: dict[str, object],
-        dedup_key: str | None,
-    ) -> str: ...
-
-
-class AutomationSchedulerRepositoryLike(Protocol):
-    async def list_due_jobs(
-        self,
-        *,
-        now_utc: datetime,
-        limit: int,
-    ) -> list[DueAutomationJob]: ...
-
-    async def get_job_config(self, *, job_id: UUID) -> AutomationJobConfig: ...
-
-    async def ensure_latest_chat_session(self, *, owner_id: UUID) -> UUID: ...
-
-    async def mark_job_dispatched(
-        self,
-        *,
-        job_id: UUID,
-        next_run_at: datetime,
-        last_run_at: datetime,
-    ) -> None: ...
-
-    async def commit(self) -> None: ...
-
-    async def rollback(self) -> None: ...
-
-
-@dataclass(slots=True)
-class DispatchResult:
-    scanned: int
-    dispatched: int
-
-
-class AutomationSchedulerService:
-    def __init__(
-        self,
-        *,
-        repository: AutomationSchedulerRepositoryLike,
-        queue: QueueLike,
-    ) -> None:
-        self._repository = repository
-        self._queue = queue
-
-    async def scan_and_dispatch(
-        self,
-        *,
-        now_utc: datetime,
-        limit: int,
-    ) -> DispatchResult:
-        safe_limit = max(int(limit), 1)
-        due_jobs = await self._repository.list_due_jobs(
-            now_utc=now_utc, limit=safe_limit
-        )
-        dispatched = 0
-        for job in due_jobs:
-            try:
-                config = await self._repository.get_job_config(job_id=job.id)
-                thread_id = await self._repository.ensure_latest_chat_session(
-                    owner_id=job.owner_id
-                )
-                command = self._build_dispatch_command(
-                    job=job,
-                    thread_id=thread_id,
-                    input_text=config.input_template,
-                    now_utc=now_utc,
-                )
-                await self._queue.enqueue(command=command, dedup_key=None)
-                await self._repository.mark_job_dispatched(
-                    job_id=job.id,
-                    next_run_at=_compute_next_run_at(
-                        current_next_run_at=job.next_run_at,
-                        now_utc=now_utc,
-                        schedule_type=job.schedule_type,
-                    ),
-                    last_run_at=now_utc,
-                )
-                await self._repository.commit()
-                dispatched += 1
-            except Exception as exc:
-                await self._repository.rollback()
-                logger.exception(
-                    "automation job dispatch failed",
-                    job_id=str(job.id),
-                    owner_id=str(job.owner_id),
-                    error=str(exc),
-                )
-        return DispatchResult(scanned=len(due_jobs), dispatched=dispatched)
-
-    def _build_dispatch_command(
-        self,
-        *,
-        job: DueAutomationJob,
-        thread_id: UUID,
-        input_text: str,
-        now_utc: datetime,
-    ) -> dict[str, object]:
-        run_id = f"auto-{job.id}-{int(now_utc.timestamp())}"
-        payload = SchedulerDispatchCommand(
-            owner_id=job.owner_id,
-            automation_job_id=job.id,
-            thread_id=thread_id,
-            run_id=run_id,
-            input_text=input_text.strip(),
-        )
-        return {
-            "command": "run",
-            "owner_id": str(payload.owner_id),
-            "automation_job_id": str(payload.automation_job_id),
-            "queue": "bulk",
-            "run_input": {
-                "threadId": str(payload.thread_id),
-                "runId": payload.run_id,
-                "state": {},
-                "messages": [
-                    {
-                        "id": str(uuid4()),
-                        "role": "user",
-                        "content": payload.input_text,
-                    }
-                ],
-                "tools": [],
-                "context": [],
-                "forwardedProps": {
-                    "agent_type": "memory",
-                },
-            },
-        }
-
-
-class SqlAlchemyAutomationSchedulerRepository:
-    def __init__(self, *, session: AsyncSession) -> None:
-        self._session = session
-
-    async def list_due_jobs(
-        self,
-        *,
-        now_utc: datetime,
-        limit: int,
-    ) -> list[DueAutomationJob]:
-        stmt = (
-            select(AutomationJob)
-            .where(AutomationJob.deleted_at.is_(None))
-            .where(AutomationJob.status == "active")
-            .where(AutomationJob.next_run_at <= now_utc)
-            .order_by(AutomationJob.next_run_at.asc())
-            .limit(max(limit, 1))
-        )
-        rows = (await self._session.execute(stmt)).scalars().all()
-        return [
-            DueAutomationJob(
-                id=row.id,
-                owner_id=row.owner_id,
-                schedule_type=row.schedule_type,
-                timezone=row.timezone,
-                next_run_at=row.next_run_at,
-            )
-            for row in rows
-        ]
-
-    async def get_job_config(self, *, job_id: UUID) -> AutomationJobConfig:
-        stmt = select(AutomationJob.config).where(AutomationJob.id == job_id)
-        config_payload = (await self._session.execute(stmt)).scalar_one()
-        return AutomationJobConfig.model_validate(config_payload or {})
-
-    async def ensure_latest_chat_session(self, *, owner_id: UUID) -> UUID:
-        stmt = (
-            select(AgentChatSession.id)
-            .where(AgentChatSession.user_id == owner_id)
-            .where(AgentChatSession.deleted_at.is_(None))
-            .where(AgentChatSession.session_type == SessionType.CHAT)
-            .order_by(AgentChatSession.last_activity_at.desc())
-            .limit(1)
-        )
-        existing = (await self._session.execute(stmt)).scalar_one_or_none()
-        if existing is not None:
-            return existing
-
-        session = AgentChatSession(
-            id=uuid4(),
-            user_id=owner_id,
-            session_type=SessionType.CHAT,
-        )
-        self._session.add(session)
-        await self._session.flush()
-        return session.id
-
-    async def mark_job_dispatched(
-        self,
-        *,
-        job_id: UUID,
-        next_run_at: datetime,
-        last_run_at: datetime,
-    ) -> None:
-        stmt = select(AutomationJob).where(AutomationJob.id == job_id)
-        row = (await self._session.execute(stmt)).scalar_one()
-        row.next_run_at = next_run_at
-        row.last_run_at = last_run_at
-        await self._session.flush()
-
-    async def commit(self) -> None:
-        await self._session.commit()
-
-    async def rollback(self) -> None:
-        await self._session.rollback()
-
-
-def _compute_next_run_at(
-    *,
-    current_next_run_at: datetime,
-    now_utc: datetime,
-    schedule_type: ScheduleType,
-) -> datetime:
-    delta = timedelta(days=1 if schedule_type == ScheduleType.DAILY else 7)
-    next_run_at = current_next_run_at
-    while next_run_at <= now_utc:
-        next_run_at = next_run_at + delta
-    return next_run_at
-
-
 def utc_now() -> datetime:
     return datetime.now(timezone.utc)
 
@@ -272,22 +24,85 @@ async def run_automation_scheduler_scan(
         if isinstance(limit, int)
         else int(config.automation_scheduler.batch_limit)
     )
+
     from core.db.session import AsyncSessionLocal
+    from v1.automation_jobs.repository import AutomationJobsRepository
+    from v1.automation_jobs.service import AutomationJobsService
 
     async with AsyncSessionLocal() as session:
-        repository = SqlAlchemyAutomationSchedulerRepository(session=session)
-        service = AutomationSchedulerService(
-            repository=repository,
-            queue=_BulkQueueAdapter(),
+        repository = AutomationJobsRepository(session=session)
+        service = AutomationJobsService(repository=repository, session=session)
+
+        result = await service.scan_and_dispatch(
+            now_utc=now,
+            limit=safe_limit,
+            dispatch_fn=_dispatch_automation_run,
         )
-        result = await service.scan_and_dispatch(now_utc=now, limit=safe_limit)
+
     logger.info(
         "automation scheduler scan completed",
         scanned=result.scanned,
         dispatched=result.dispatched,
-        now_utc=now.astimezone(timezone.utc).isoformat(),
+        now_utc=now.isoformat(),
     )
     return {
-        "scanned": int(result.scanned),
-        "dispatched": int(result.dispatched),
+        "scanned": result.scanned,
+        "dispatched": result.dispatched,
     }
+
+
+async def _dispatch_automation_run(
+    *,
+    owner_id: UUID,
+    thread_id: UUID,
+    run_id: str,
+    input_text: str,
+    runtime_config: RuntimeConfig,
+) -> None:
+    from uuid import uuid4
+
+    from ag_ui.core import RunAgentInput
+    from core.auth.models import CurrentUser
+    from core.agentscope.tools.tool_result_storage import create_tool_result_storage
+    from schemas.agent.forwarded_props import RuntimeMode
+    from v1.agent.dependencies import TaskiqQueueClient, RedisEventStream
+    from v1.agent.repository import AgentRepository
+    from v1.agent.service import AgentService
+
+    current_user = CurrentUser(id=owner_id)
+    tool_result_storage = create_tool_result_storage()
+
+    run_input = {
+        "threadId": str(thread_id),
+        "runId": run_id,
+        "state": {},
+        "messages": [
+            {
+                "id": str(uuid4()),
+                "role": "user",
+                "content": input_text,
+            }
+        ],
+        "forwardedProps": {
+            "runtimeMode": RuntimeMode.AUTOMATION.value,
+        },
+    }
+
+    parsed_run_input = RunAgentInput.model_validate(run_input)
+
+    from core.db.session import AsyncSessionLocal
+
+    async with AsyncSessionLocal() as session:
+        repository = AgentRepository(
+            session=session, tool_result_storage=tool_result_storage
+        )
+        service = AgentService(
+            repository=repository,
+            queue=TaskiqQueueClient(),
+            stream=RedisEventStream(),
+        )
+        await service.enqueue_run(
+            run_input=parsed_run_input,
+            current_user=current_user,
+            runtime_config=runtime_config,
+        )

backend/src/core/automation/tasks.py

@@ -1,14 +0,0 @@
-from __future__ import annotations
-
-
-from core.logging import get_logger
-from core.taskiq.app import worker_automation_broker
-
-logger = get_logger("core.automation.tasks")
-
-
-@worker_automation_broker.task(task_name="tasks.automation.scan_due_jobs")
-async def scan_due_automation_jobs_task(limit: int | None = None) -> dict[str, int]:
-    from core.automation.scheduler import run_automation_scheduler_scan
-
-    return await run_automation_scheduler_scan(limit=limit)

backend/src/core/config/static/database/system_agents.yaml

@@ -18,9 +18,7 @@ agents:
       temperature: 0.7
       max_tokens: null
       timeout_seconds: 30
-      context_messages:
-        mode: number
-        count: 20
+      context_messages: null
       enabled_tools:
         - calendar.read
         - calendar.write

backend/src/schemas/__init__.py

@@ -12,7 +12,7 @@ from schemas.inbox.messages import (
     parse_calendar_content,
 )
 from schemas.invite_codes import InviteCodeRewardConfig
-from schemas.memories import MemoryContent
+from schemas.memories import MemoryContext
 from schemas.messages import AgentChatMessageMetadata
 from schemas.schedule.items import (
     AttachmentType,
@@ -36,7 +36,7 @@ __all__ = [
     "InboxMessageStatus",
     "InboxMessageType",
     "InviteCodeRewardConfig",
-    "MemoryContent",
+    "MemoryContext",
     "ScheduleItemMetadata",
     "ScheduleItemMetadataAttachment",
     "ScheduleItemSourceType",

backend/src/schemas/agent/runtime_models.py

@@ -100,6 +100,7 @@ class NormalizedTaskInput(BaseModel):
 
     user_text: str
     multimodal_summary: list[str] = Field(default_factory=list)
+    context_summary: str = Field(default="", max_length=2000)
 
 
 class RouterUiDecision(BaseModel):

backend/src/schemas/automation/__init__.py

@@ -1,20 +1,77 @@
-from schemas.automation.config import (
-    AutomationAgentType,
-    AutomationContextSource,
-    AutomationContextWindowMode,
-    AutomationJobConfig,
-    AutomationMemoryContextConfig,
-    default_memory_job_config,
-)
-from schemas.automation.scheduler import DueAutomationJob, SchedulerDispatchCommand
+from __future__ import annotations
 
-__all__ = [
-    "AutomationAgentType",
-    "AutomationContextSource",
-    "AutomationContextWindowMode",
-    "AutomationJobConfig",
-    "AutomationMemoryContextConfig",
-    "default_memory_job_config",
-    "DueAutomationJob",
-    "SchedulerDispatchCommand",
-]
+from datetime import datetime
+from enum import Enum
+from uuid import UUID
+
+from core.agentscope.tools.tool_config import AgentTool
+from pydantic import BaseModel, ConfigDict, Field
+
+from models.automation_jobs import AutomationJob as OrmAutomationJob
+from models.automation_jobs import AutomationJobStatus, ScheduleType
+
+
+class ContextSource(str, Enum):
+    LATEST_CHAT = "latest_chat"
+
+
+class ContextWindowMode(str, Enum):
+    DAY = "day"
+    NUMBER = "number"
+
+
+class MemoryContextConfig(BaseModel):
+    model_config = ConfigDict(extra="forbid")
+
+    source: ContextSource = ContextSource.LATEST_CHAT
+    window_mode: ContextWindowMode = ContextWindowMode.DAY
+    window_count: int = Field(default=2, ge=1, le=200)
+
+
+class RuntimeConfig(BaseModel):
+    model_config = ConfigDict(extra="forbid")
+
+    enabled_tools: list[AgentTool] = Field(default_factory=list, max_length=32)
+    context: MemoryContextConfig = Field(default_factory=MemoryContextConfig)
+
+
+class AutomationJobConfig(RuntimeConfig):
+    model_config = ConfigDict(extra="forbid")
+
+    input_template: str = Field(..., min_length=1, max_length=4000)
+
+
+class AutomationJob(BaseModel):
+    model_config = ConfigDict(extra="forbid")
+
+    id: UUID
+    owner_id: UUID
+    title: str = Field(..., min_length=1, max_length=255)
+    config: AutomationJobConfig
+    schedule_type: ScheduleType
+    run_at: datetime
+    next_run_at: datetime
+    timezone: str = Field(default="UTC", min_length=1, max_length=50)
+    last_run_at: datetime | None = None
+    status: AutomationJobStatus
+    created_by: UUID | None = None
+    created_at: datetime
+    updated_at: datetime
+
+    @classmethod
+    def from_orm(cls, obj: OrmAutomationJob) -> "AutomationJob":
+        return cls(
+            id=obj.id,
+            owner_id=obj.owner_id,
+            title=obj.title,
+            config=AutomationJobConfig.model_validate(obj.config or {}),
+            schedule_type=obj.schedule_type,
+            run_at=obj.run_at,
+            next_run_at=obj.next_run_at,
+            timezone=obj.timezone,
+            last_run_at=obj.last_run_at,
+            status=obj.status,
+            created_by=obj.created_by,
+            created_at=obj.created_at,
+            updated_at=obj.updated_at,
+        )

backend/src/schemas/automation/config.py

@@ -1,62 +0,0 @@
-from __future__ import annotations
-
-from enum import Enum
-
-from pydantic import BaseModel, ConfigDict, Field, field_validator
-
-from core.agentscope.tools.tool_config import AgentTool
-
-
-class AutomationAgentType(str, Enum):
-    MEMORY = "memory"
-
-
-class AutomationContextSource(str, Enum):
-    LATEST_CHAT = "latest_chat"
-
-
-class AutomationContextWindowMode(str, Enum):
-    DAY = "day"
-    NUMBER = "number"
-
-
-class AutomationMemoryContextConfig(BaseModel):
-    model_config = ConfigDict(extra="forbid")
-
-    source: AutomationContextSource = AutomationContextSource.LATEST_CHAT
-    window_mode: AutomationContextWindowMode = AutomationContextWindowMode.DAY
-    window_count: int = Field(default=2, ge=1, le=200)
-
-
-class AutomationJobConfig(BaseModel):
-    model_config = ConfigDict(extra="forbid")
-
-    agent_type: AutomationAgentType = AutomationAgentType.MEMORY
-    model_code: str = Field(default="qwen3.5-flash", min_length=1, max_length=64)
-    enabled_tools: list[AgentTool] = Field(default_factory=list, max_length=32)
-    input_template: str = Field(..., min_length=1, max_length=4000)
-    context: AutomationMemoryContextConfig = Field(
-        default_factory=AutomationMemoryContextConfig
-    )
-
-    @field_validator("model_code")
-    @classmethod
-    def _validate_model_code(cls, value: str) -> str:
-        normalized = value.strip()
-        if normalized != "qwen3.5-flash":
-            raise ValueError("model_code must be qwen3.5-flash")
-        return normalized
-
-
-def default_memory_job_config() -> AutomationJobConfig:
-    return AutomationJobConfig(
-        agent_type=AutomationAgentType.MEMORY,
-        model_code="qwen3.5-flash",
-        enabled_tools=[AgentTool.CALENDAR_READ, AgentTool.USER_LOOKUP],
-        input_template="请基于最近聊天上下文生成一段可执行的记忆总结与建议。",
-        context=AutomationMemoryContextConfig(
-            source=AutomationContextSource.LATEST_CHAT,
-            window_mode=AutomationContextWindowMode.DAY,
-            window_count=2,
-        ),
-    )

backend/src/schemas/automation/scheduler.py

@@ -1,28 +0,0 @@
-from __future__ import annotations
-
-from datetime import datetime
-from uuid import UUID
-
-from pydantic import BaseModel, ConfigDict, Field
-
-from models.automation_jobs import ScheduleType
-
-
-class DueAutomationJob(BaseModel):
-    model_config = ConfigDict(extra="forbid")
-
-    id: UUID
-    owner_id: UUID
-    schedule_type: ScheduleType
-    timezone: str = Field(..., min_length=1, max_length=50)
-    next_run_at: datetime
-
-
-class SchedulerDispatchCommand(BaseModel):
-    model_config = ConfigDict(extra="forbid")
-
-    owner_id: UUID
-    automation_job_id: UUID
-    thread_id: UUID
-    run_id: str = Field(..., min_length=1, max_length=128)
-    input_text: str = Field(..., min_length=1, max_length=4000)

backend/src/schemas/memories/__init__.py

@@ -1,11 +1,70 @@
 from __future__ import annotations
 
-from typing import ClassVar
+from datetime import datetime
+from enum import Enum
+from typing import Any, ClassVar, Literal
+from uuid import UUID
 
-from pydantic import BaseModel, ConfigDict
+from pydantic import BaseModel, ConfigDict, Field
 
 
-class MemoryContent(BaseModel):
-    model_config: ClassVar[ConfigDict] = ConfigDict(extra="allow")
+class MemoryType(str, Enum):
+    USER = "user"
+    WORK = "work"
 
-    pass
+
+class MemorySource(str, Enum):
+    MANUAL = "manual"
+    AGENT = "agent"
+    IMPORTED = "imported"
+
+
+class MemoryStatus(str, Enum):
+    ACTIVE = "active"
+    DISABLED = "disabled"
+
+
+class MemoryModel(BaseModel):
+    model_config: ClassVar[ConfigDict] = ConfigDict(
+        extra="forbid", from_attributes=True
+    )
+
+    id: UUID
+    owner_id: UUID
+    agent_id: UUID | None = None
+    memory_type: Literal["user", "work"]
+    title: str | None = None
+    content: dict[str, Any]
+    source: MemorySource
+    status: MemoryStatus
+    created_at: datetime
+    updated_at: datetime
+
+
+class MemoryContext(BaseModel):
+    model_config: ClassVar[ConfigDict] = ConfigDict(extra="forbid")
+
+    memory_type: MemoryType
+    source: MemorySource
+    title: str | None = None
+    content: dict[str, Any]
+    created_at: datetime
+    updated_at: datetime
+
+
+class MemoryListResponse(BaseModel):
+    model_config: ClassVar[ConfigDict] = ConfigDict(extra="forbid")
+
+    owner_id: UUID
+    memories: list[MemoryContext] = Field(default_factory=list)
+    total: int
+
+
+__all__ = [
+    "MemoryContext",
+    "MemoryListResponse",
+    "MemoryModel",
+    "MemorySource",
+    "MemoryStatus",
+    "MemoryType",
+]

backend/src/v1/agent/service.py

@@ -18,6 +18,7 @@ from schemas.agent.forwarded_props import (
     RuntimeMode,
 )
 from schemas.agent.visibility import SystemVisibilityBit, bit_mask
+from schemas.automation import RuntimeConfig
 from schemas.messages.chat_message import (
     AgentChatMessageMetadata,
     UserMessageAttachment,
@@ -72,6 +73,7 @@ class AgentService:
         *,
         run_input: RunAgentInput,
         current_user: CurrentUser,
+        runtime_config: RuntimeConfig | None = None,
     ) -> TaskAccepted:
         created = False
         thread_id = run_input.thread_id
@@ -82,6 +84,13 @@ class AgentService:
         except ValueError as exc:
             raise HTTPException(status_code=422, detail=str(exc)) from exc
 
+        if runtime_config is None:
+            from v1.agent.system_agents_config import (
+                build_runtime_config_from_system_agents,
+            )
+
+            runtime_config = build_runtime_config_from_system_agents()
+
         try:
             owner = await self._repository.get_session_owner(session_id=thread_id)
         except HTTPException as exc:
@@ -124,6 +133,9 @@ class AgentService:
                 "run_input": run_input.model_dump(
                     mode="json", by_alias=True, exclude_none=True
                 ),
+                "runtime_config": runtime_config.model_dump(
+                    mode="json", by_alias=True, exclude_none=True
+                ),
                 "queue": queue,
             },
             dedup_key=None,

backend/src/v1/agent/system_agents_config.py

@@ -0,0 +1,103 @@
+"""
+System agents 配置加载工具
+
+从 system_agents.yaml 加载配置并构建 RuntimeConfig
+"""
+
+from pathlib import Path
+
+import yaml
+from pydantic import ValidationError
+
+from schemas.agent.system_agent import SystemAgentLLMConfig
+from schemas.automation import (
+    ContextSource,
+    ContextWindowMode,
+    MemoryContextConfig,
+    RuntimeConfig,
+)
+
+
+def _default_system_agents_path() -> Path:
+    return (
+        Path(__file__).resolve().parents[2]
+        / "core"
+        / "config"
+        / "static"
+        / "database"
+        / "system_agents.yaml"
+    )
+
+
+def _load_system_agents_yaml(path: Path | None = None) -> dict:
+    target_path = path or _default_system_agents_path()
+    with target_path.open("r", encoding="utf-8") as f:
+        loaded = yaml.safe_load(f) or {}
+    if not isinstance(loaded, dict):
+        raise ValueError(f"Invalid system agents format: {target_path}")
+    return loaded
+
+
+def _parse_context_messages_config(yaml_config: dict | None) -> MemoryContextConfig:
+    if not yaml_config:
+        return MemoryContextConfig()
+    mode_str = yaml_config.get("mode", "day")
+    count = yaml_config.get("count", 2)
+    try:
+        source = ContextSource.LATEST_CHAT
+    except ValueError:
+        source = ContextSource.LATEST_CHAT
+    try:
+        window_mode = ContextWindowMode(mode_str)
+    except ValueError:
+        window_mode = ContextWindowMode.DAY
+    return MemoryContextConfig(
+        source=source,
+        window_mode=window_mode,
+        window_count=count,
+    )
+
+
+def build_runtime_config_from_system_agents(
+    yaml_path: Path | None = None,
+) -> RuntimeConfig:
+    """
+    从 system_agents.yaml 构建 RuntimeConfig
+
+    chat 模式使用:
+    - router.context_messages 配置 context
+    - worker.enabled_tools 配置 tools
+    """
+    raw = _load_system_agents_yaml(yaml_path)
+    agents_list = raw.get("agents", [])
+
+    router_config: SystemAgentLLMConfig | None = None
+    worker_config: SystemAgentLLMConfig | None = None
+
+    for agent in agents_list:
+        agent_type = str(agent.get("agent_type", "")).strip().lower()
+        if agent_type == "router":
+            config_dict = agent.get("config") or {}
+            try:
+                router_config = SystemAgentLLMConfig.model_validate(config_dict)
+            except ValidationError:
+                router_config = SystemAgentLLMConfig()
+        elif agent_type == "worker":
+            config_dict = agent.get("config") or {}
+            try:
+                worker_config = SystemAgentLLMConfig.model_validate(config_dict)
+            except ValidationError:
+                worker_config = SystemAgentLLMConfig()
+
+    context_cfg = _parse_context_messages_config(
+        router_config.context_messages.model_dump() if router_config else None
+    )
+
+    enabled_tools: list[str] = []
+    if worker_config and worker_config.enabled_tools:
+        enabled_tools = [str(t) for t in worker_config.enabled_tools]
+
+    return RuntimeConfig(
+        enabled_tools=enabled_tools,
+        context=context_cfg,
+    )

backend/src/v1/automation_jobs/__init__.py

@@ -0,0 +1,3 @@
+from v1.automation_jobs.service import AutomationJobsService
+
+__all__ = ["AutomationJobsService"]

backend/src/v1/automation_jobs/repository.py

@@ -0,0 +1,77 @@
+from __future__ import annotations
+
+from datetime import datetime
+from typing import TYPE_CHECKING
+from uuid import UUID
+
+from sqlalchemy import select, update
+from sqlalchemy.ext.asyncio import AsyncSession
+
+from core.db.base_repository import BaseRepository
+from models.agent_chat_session import AgentChatSession, SessionType
+from models.automation_jobs import AutomationJob
+
+if TYPE_CHECKING:
+    pass
+
+
+class AutomationJobsRepository(BaseRepository[AutomationJob]):
+    def __init__(self, session: AsyncSession) -> None:
+        super().__init__(session=session, model=AutomationJob)
+
+    async def list_due_jobs(
+        self,
+        *,
+        now_utc: datetime,
+        limit: int,
+    ) -> list[AutomationJob]:
+        stmt = (
+            select(AutomationJob)
+            .where(AutomationJob.deleted_at.is_(None))
+            .where(AutomationJob.status == "active")
+            .where(AutomationJob.next_run_at <= now_utc)
+            .order_by(AutomationJob.next_run_at.asc())
+            .limit(max(limit, 1))
+        )
+        rows = (await self._session.execute(stmt)).scalars().all()
+        return list(rows)
+
+    async def update_job_schedule(
+        self,
+        *,
+        job_id: UUID,
+        next_run_at: datetime,
+        last_run_at: datetime,
+    ) -> None:
+        stmt = (
+            update(AutomationJob)
+            .where(AutomationJob.id == job_id)
+            .where(AutomationJob.deleted_at.is_(None))
+            .values(next_run_at=next_run_at, last_run_at=last_run_at)
+        )
+        await self._session.execute(stmt)
+        await self._session.flush()
+
+    async def get_or_create_chat_session(self, *, owner_id: UUID) -> UUID:
+        stmt = (
+            select(AgentChatSession.id)
+            .where(AgentChatSession.user_id == owner_id)
+            .where(AgentChatSession.deleted_at.is_(None))
+            .where(AgentChatSession.session_type == SessionType.CHAT)
+            .order_by(AgentChatSession.last_activity_at.desc())
+            .limit(1)
+        )
+        existing = (await self._session.execute(stmt)).scalar_one_or_none()
+        if existing is not None:
+            return existing
+
+        from uuid import uuid4
+
+        new_session = AgentChatSession(
+            id=uuid4(),
+            user_id=owner_id,
+            session_type=SessionType.CHAT,
+        )
+        self._session.add(new_session)
+        await self._session.flush()
+        return new_session.id

backend/src/v1/automation_jobs/service.py

@@ -0,0 +1,104 @@
+from __future__ import annotations
+
+from dataclasses import dataclass
+from datetime import datetime, timedelta
+from typing import TYPE_CHECKING, Protocol
+from uuid import UUID
+
+from models.automation_jobs import ScheduleType
+from schemas.automation import AutomationJob as AutomationJobSchema, RuntimeConfig
+
+if TYPE_CHECKING:
+    from sqlalchemy.ext.asyncio import AsyncSession
+
+    from v1.automation_jobs.repository import AutomationJobsRepository
+
+
+class DispatchFn(Protocol):
+    async def __call__(
+        self,
+        *,
+        owner_id: UUID,
+        thread_id: UUID,
+        run_id: str,
+        input_text: str,
+        runtime_config: RuntimeConfig,
+    ) -> None: ...
+
+
+def _compute_next_run_at(
+    *,
+    current_next_run_at: datetime,
+    now_utc: datetime,
+    schedule_type: ScheduleType,
+) -> datetime:
+    delta = timedelta(days=1 if schedule_type == ScheduleType.DAILY else 7)
+    next_run_at = current_next_run_at
+    while next_run_at <= now_utc:
+        next_run_at = next_run_at + delta
+    return next_run_at
+
+
+@dataclass(slots=True)
+class ScanResult:
+    scanned: int
+    dispatched: int
+
+
+class AutomationJobsService:
+    def __init__(
+        self,
+        repository: "AutomationJobsRepository",
+        session: "AsyncSession",
+    ) -> None:
+        self._repository = repository
+        self._session = session
+
+    async def scan_and_dispatch(
+        self,
+        *,
+        now_utc: datetime,
+        limit: int,
+        dispatch_fn: DispatchFn,
+    ) -> ScanResult:
+        rows = await self._repository.list_due_jobs(now_utc=now_utc, limit=limit)
+        scanned = len(rows)
+        dispatched = 0
+
+        for row in rows:
+            try:
+                job = AutomationJobSchema.from_orm(row)
+                thread_id = await self.get_or_create_chat_session(owner_id=job.owner_id)
+                run_id = f"auto-{job.id}-{int(now_utc.timestamp())}"
+
+                await dispatch_fn(
+                    owner_id=job.owner_id,
+                    thread_id=thread_id,
+                    run_id=run_id,
+                    input_text=job.config.input_template.strip(),
+                    runtime_config=RuntimeConfig(
+                        enabled_tools=job.config.enabled_tools,
+                        context=job.config.context,
+                    ),
+                )
+
+                await self._repository.update_job_schedule(
+                    job_id=job.id,
+                    next_run_at=_compute_next_run_at(
+                        current_next_run_at=job.next_run_at,
+                        now_utc=now_utc,
+                        schedule_type=job.schedule_type,
+                    ),
+                    last_run_at=now_utc,
+                )
+                await self._session.commit()
+                dispatched += 1
+
+            except Exception:
+                await self._session.rollback()
+                raise
+
+        return ScanResult(scanned=scanned, dispatched=dispatched)
+
+    async def get_or_create_chat_session(self, *, owner_id: UUID) -> UUID:
+        return await self._repository.get_or_create_chat_session(owner_id=owner_id)

backend/src/v1/memories/__init__.py

@@ -0,0 +1,3 @@
+from v1.memories.service import MemoriesService
+
+__all__ = ["MemoriesService"]

backend/src/v1/memories/repository.py

@@ -0,0 +1,31 @@
+from __future__ import annotations
+
+from typing import TYPE_CHECKING, Protocol
+from uuid import UUID
+
+from sqlalchemy import select
+
+from core.db.base_repository import BaseRepository
+from models.memories import Memory
+
+if TYPE_CHECKING:
+    from sqlalchemy.ext.asyncio import AsyncSession
+
+
+class MemoriesRepositoryLike(Protocol):
+    async def get_active_memories(self, *, owner_id: UUID) -> list[Memory]: ...
+
+
+class MemoriesRepository(BaseRepository[Memory]):
+    def __init__(self, session: AsyncSession) -> None:
+        super().__init__(session=session, model=Memory)
+
+    async def get_active_memories(self, *, owner_id: UUID) -> list[Memory]:
+        stmt = (
+            select(Memory)
+            .where(Memory.owner_id == owner_id)
+            .where(Memory.status == "active")
+            .order_by(Memory.created_at.desc())
+        )
+        result = await self._session.execute(stmt)
+        return list(result.scalars().all())

backend/src/v1/memories/service.py

@@ -0,0 +1,53 @@
+from __future__ import annotations
+
+from uuid import UUID
+
+from models.memories import Memory
+from schemas.memories import MemoryContext, MemoryListResponse, MemorySource, MemoryType
+from v1.memories.repository import MemoriesRepositoryLike
+
+
+class MemoriesService:
+    _repository: MemoriesRepositoryLike
+
+    def __init__(self, repository: MemoriesRepositoryLike) -> None:
+        self._repository = repository
+
+    def _to_context(self, memory: Memory) -> MemoryContext:
+        return MemoryContext(
+            memory_type=MemoryType(memory.memory_type.value),
+            source=MemorySource(memory.source.value),
+            title=memory.title,
+            content=memory.content,
+            created_at=memory.created_at,
+            updated_at=memory.updated_at,
+        )
+
+    async def get_user_memories(self, *, owner_id: UUID) -> MemoryListResponse:
+        memories = await self._repository.get_active_memories(owner_id=owner_id)
+        user_memories = [
+            self._to_context(memory)
+            for memory in memories
+            if memory.memory_type.value == "user"
+        ]
+        return MemoryListResponse(
+            owner_id=owner_id, memories=user_memories, total=len(user_memories)
+        )
+
+    async def get_agent_memories(self, *, owner_id: UUID) -> MemoryListResponse:
+        memories = await self._repository.get_active_memories(owner_id=owner_id)
+        agent_memories = [
+            self._to_context(memory)
+            for memory in memories
+            if memory.memory_type.value == "work"
+        ]
+        return MemoryListResponse(
+            owner_id=owner_id, memories=agent_memories, total=len(agent_memories)
+        )
+
+    async def get_all_memories(self, *, owner_id: UUID) -> MemoryListResponse:
+        memories = await self._repository.get_active_memories(owner_id=owner_id)
+        memory_contexts = [self._to_context(memory) for memory in memories]
+        return MemoryListResponse(
+            owner_id=owner_id, memories=memory_contexts, total=len(memory_contexts)
+        )

backend/src/v1/memory/__init__.py

@@ -1,3 +0,0 @@
-from v1.memory.service import MemoryService
-
-__all__ = ["MemoryService"]

backend/src/v1/memory/repository.py

@@ -1,35 +0,0 @@
-from __future__ import annotations
-
-from typing import TYPE_CHECKING, Protocol
-from uuid import UUID
-
-from sqlalchemy import select
-
-from core.db.base_repository import BaseRepository
-from models.automation_jobs import AutomationJob
-
-if TYPE_CHECKING:
-    from sqlalchemy.ext.asyncio import AsyncSession
-
-
-class MemoryRepositoryLike(Protocol):
-    async def get_job_by_id_and_owner(
-        self, *, job_id: UUID, owner_id: UUID
-    ) -> AutomationJob | None: ...
-
-
-class MemoryRepository(BaseRepository[AutomationJob]):
-    def __init__(self, session: AsyncSession) -> None:
-        super().__init__(session=session, model=AutomationJob)
-
-    async def get_job_by_id_and_owner(
-        self, *, job_id: UUID, owner_id: UUID
-    ) -> AutomationJob | None:
-        stmt = (
-            select(AutomationJob)
-            .where(AutomationJob.id == job_id)
-            .where(AutomationJob.owner_id == owner_id)
-            .where(AutomationJob.deleted_at.is_(None))
-        )
-        result = await self._session.execute(stmt)
-        return result.scalar_one_or_none()

backend/src/v1/memory/service.py

@@ -1,25 +0,0 @@
-from __future__ import annotations
-
-from uuid import UUID
-
-from fastapi import HTTPException
-
-from schemas.automation.config import AutomationJobConfig
-from v1.memory.repository import MemoryRepositoryLike
-
-
-class MemoryService:
-    _repository: MemoryRepositoryLike
-
-    def __init__(self, repository: MemoryRepositoryLike) -> None:
-        self._repository = repository
-
-    async def get_memory_job_config(
-        self, *, job_id: UUID, owner_id: UUID
-    ) -> AutomationJobConfig:
-        job = await self._repository.get_job_by_id_and_owner(
-            job_id=job_id, owner_id=owner_id
-        )
-        if job is None:
-            raise HTTPException(status_code=404, detail="Automation job not found")
-        return AutomationJobConfig.model_validate(job.config or {})

backend/tests/unit/core/agentscope/runtime/test_orchestrator.py

@@ -6,6 +6,7 @@ import pytest
 from ag_ui.core import RunAgentInput
 
 from core.agentscope.runtime.orchestrator import AgentScopeRuntimeOrchestrator
+from schemas.automation import MemoryContextConfig, RuntimeConfig
 from schemas.user import UserContext, parse_profile_settings
 
 
@@ -42,11 +43,18 @@ def _run_input() -> RunAgentInput:
             "messages": [{"id": "u1", "role": "user", "content": "hello"}],
             "tools": [],
             "context": [],
-            "forwardedProps": {"agent_type": "worker"},
+            "forwardedProps": {"runtime_mode": "automation"},
         }
     )
 
 
+def _runtime_config() -> RuntimeConfig:
+    return RuntimeConfig(
+        enabled_tools=[],
+        context=MemoryContextConfig(),
+    )
+
+
 @pytest.mark.asyncio
 async def test_orchestrator_emits_run_lifecycle_events() -> None:
     pipeline = _FakePipeline()
@@ -58,7 +66,7 @@ async def test_orchestrator_emits_run_lifecycle_events() -> None:
         run_input=_run_input(),
         context_messages=[],
         user_context=_user_context(),
-        system_agent_mode="worker",
+        runtime_config=_runtime_config(),
     )
 
     assert result["worker"]["answer"] == "done"

backend/tests/unit/core/agentscope/runtime/test_runner.py

@@ -5,7 +5,6 @@ from ag_ui.core import RunAgentInput
 
 import core.agentscope.runtime.runner as runner_module
 from core.agentscope.runtime.runner import AgentScopeRunner
-from schemas.automation.config import default_memory_job_config
 from schemas.agent.runtime_models import (
     ExecutionMode,
     NormalizedTaskInput,
@@ -19,6 +18,7 @@ from schemas.agent.runtime_models import (
     WorkerAgentOutputLite,
 )
 from schemas.agent.system_agent import AgentType
+from schemas.automation import MemoryContextConfig, RuntimeConfig
 from schemas.user import UserContext, parse_profile_settings
 
 
@@ -31,7 +31,7 @@ def _run_input() -> RunAgentInput:
             "messages": [{"id": "u1", "role": "user", "content": "hello"}],
             "tools": [],
             "context": [],
-            "forwardedProps": {"agent_type": "worker"},
+            "forwardedProps": {"runtime_mode": "automation"},
         }
     )
 
@@ -45,10 +45,20 @@ def _user_context() -> UserContext:
     )
 
 
+def _runtime_config() -> RuntimeConfig:
+    return RuntimeConfig(
+        enabled_tools=[],
+        context=MemoryContextConfig(),
+    )
+
+
 def test_build_worker_input_messages_only_contains_router_contract() -> None:
     runner = AgentScopeRunner()
     router_output = RouterAgentOutput(
-        normalized_task_input=NormalizedTaskInput(user_text="安排明天会议"),
+        normalized_task_input=NormalizedTaskInput(
+            user_text="安排明天会议",
+            context_summary="用户询问天气",
+        ),
         key_entities=[],
         constraints=[],
         task_typing=TaskTyping(primary=TaskType.SCHEDULING),
@@ -67,6 +77,43 @@ def test_build_worker_input_messages_only_contains_router_contract() -> None:
     assert "[RouterAgentOutput]" in str(input_messages[0].content)
 
 
+def test_build_router_messages_injects_user_input_when_context_last_not_user() -> None:
+    runner = AgentScopeRunner()
+    run_input = _run_input()
+
+    messages = runner._build_router_messages(
+        context_messages=[],
+        run_input=run_input,
+    )
+
+    assert len(messages) == 1
+    assert messages[0].role == "user"
+    assert messages[0].content == "hello"
+
+
+def test_build_router_messages_skips_injection_when_context_last_is_user() -> None:
+    runner = AgentScopeRunner()
+    run_input = _run_input()
+
+    from agentscope.message import Msg
+
+    existing_context = [
+        Msg(name="user", role="user", content="之前的问题"),
+        Msg(name="assistant", role="assistant", content="回答"),
+        Msg(name="user", role="user", content="最新用户消息"),
+    ]
+
+    messages = runner._build_router_messages(
+        context_messages=existing_context,
+        run_input=run_input,
+    )
+
+    assert len(messages) == len(existing_context)
+    for i, msg in enumerate(messages):
+        assert msg.role == existing_context[i].role
+        assert msg.content == existing_context[i].content
+
+
 @pytest.mark.asyncio
 async def test_resolve_runtime_client_time_from_forwarded_props() -> None:
     runner = AgentScopeRunner()
@@ -79,7 +126,7 @@ async def test_resolve_runtime_client_time_from_forwarded_props() -> None:
             "tools": [],
             "context": [],
             "forwardedProps": {
-                "agent_type": "worker",
+                "runtime_mode": "automation",
                 "client_time": {
                     "device_timezone": "America/Los_Angeles",
                     "client_now_iso": "2026-03-16T09:12:33-07:00",
@@ -95,7 +142,7 @@ async def test_resolve_runtime_client_time_from_forwarded_props() -> None:
 
 
 @pytest.mark.asyncio
-async def test_execute_worker_mode_runs_router_then_worker(
+async def test_execute_runs_router_then_worker(
     monkeypatch: pytest.MonkeyPatch,
 ) -> None:
     class _FakePipeline:
@@ -127,7 +174,10 @@ async def test_execute_worker_mode_runs_router_then_worker(
     async def _fake_execute_router_step(**kwargs: object) -> RouterAgentOutput:
         del kwargs
         return RouterAgentOutput(
-            normalized_task_input=NormalizedTaskInput(user_text="安排会议"),
+            normalized_task_input=NormalizedTaskInput(
+                user_text="安排会议",
+                context_summary="用户询问天气",
+            ),
             key_entities=[],
             constraints=[],
             task_typing=TaskTyping(primary=TaskType.SCHEDULING),
@@ -145,7 +195,7 @@ async def test_execute_worker_mode_runs_router_then_worker(
 
     monkeypatch.setattr(runner_module, "AsyncSessionLocal", lambda: _FakeSessionCtx())
     monkeypatch.setattr(runner, "_load_stage_config", _fake_load_stage_config)
-    monkeypatch.setattr(runner, "_build_stage_toolkit", lambda **kwargs: object())
+    monkeypatch.setattr(runner, "_build_toolkit", lambda **kwargs: object())
     monkeypatch.setattr(runner, "_execute_router_step", _fake_execute_router_step)
     monkeypatch.setattr(runner, "_execute_worker_step", _fake_execute_worker_step)
 
@@ -154,84 +204,9 @@ async def test_execute_worker_mode_runs_router_then_worker(
         context_messages=[],
         pipeline=_FakePipeline(),
         run_input=_run_input(),
-        system_agent_mode="worker",
+        runtime_config=_runtime_config(),
     )
 
     assert load_calls == [AgentType.ROUTER, AgentType.WORKER]
     assert result["router"]["normalized_task_input"]["user_text"] == "安排会议"
     assert result["worker"]["answer"] == "ok"
-
-
-@pytest.mark.asyncio
-async def test_execute_memory_mode_requires_memory_job_config() -> None:
-    runner = AgentScopeRunner()
-
-    class _FakePipeline:
-        async def emit(self, *, session_id: str, event: dict[str, object]) -> str:
-            del session_id, event
-            return "1-0"
-
-    with pytest.raises(RuntimeError, match="memory job config is required"):
-        await runner.execute(
-            user_context=_user_context(),
-            context_messages=[],
-            pipeline=_FakePipeline(),
-            run_input=_run_input(),
-            system_agent_mode="memory",
-            memory_job_config=None,
-        )
-
-
-@pytest.mark.asyncio
-async def test_execute_memory_mode_uses_memory_job_config(
-    monkeypatch: pytest.MonkeyPatch,
-) -> None:
-    class _FakePipeline:
-        async def emit(self, *, session_id: str, event: dict[str, object]) -> str:
-            del session_id, event
-            return "1-0"
-
-    class _FakeSessionCtx:
-        async def __aenter__(self) -> object:
-            return object()
-
-        async def __aexit__(self, exc_type: object, exc: object, tb: object) -> None:
-            del exc_type, exc, tb
-
-    runner = AgentScopeRunner()
-
-    async def _fake_build_memory_stage_config(**kwargs: object):
-        del kwargs
-        return runner_module.SystemAgentRuntimeConfig(
-            agent_type=AgentType.MEMORY,
-            model_code="qwen3.5-flash",
-            api_base_url="https://example.com",
-            api_key="test",
-            llm_config=runner_module.SystemAgentLLMConfig(),
-        )
-
-    async def _fake_execute_single_stage_step(**kwargs: object):
-        del kwargs
-        return runner_module.AgentOutput(answer="memory")
-
-    monkeypatch.setattr(runner_module, "AsyncSessionLocal", lambda: _FakeSessionCtx())
-    monkeypatch.setattr(
-        runner, "_build_memory_stage_config", _fake_build_memory_stage_config
-    )
-    monkeypatch.setattr(runner, "_build_stage_toolkit", lambda **kwargs: object())
-    monkeypatch.setattr(
-        runner,
-        "_execute_single_stage_step",
-        _fake_execute_single_stage_step,
-    )
-
-    result = await runner.execute(
-        user_context=_user_context(),
-        context_messages=[],
-        pipeline=_FakePipeline(),
-        run_input=_run_input(),
-        system_agent_mode="memory",
-        memory_job_config=default_memory_job_config(),
-    )
-
-    assert result["memory"]["answer"] == "memory"

backend/tests/unit/core/agentscope/runtime/test_tasks.py

@@ -7,6 +7,7 @@ import pytest
 
 import core.agentscope.runtime.tasks as tasks_module
 from schemas.agent import ToolStatus
+from schemas.automation import ContextWindowMode, MemoryContextConfig
 from schemas.user import UserContext, parse_profile_settings
 
 
@@ -15,21 +16,36 @@ def _run_input_payload() -> dict[str, Any]:
         "threadId": str(uuid4()),
         "runId": "run-1",
         "state": {},
-        "messages": [],
+        "messages": [{"id": "u1", "role": "user", "content": "现在几点"}],
         "tools": [],
         "context": [],
-        "forwardedProps": {"agent_type": "worker"},
+        "forwardedProps": {"runtime_mode": "automation"},
     }
 
 
 class _FakeSessionCtx:
-    async def __aenter__(self) -> object:
-        return object()
+    async def __aenter__(self) -> "_FakeSession":
+        return _FakeSession()
 
     async def __aexit__(self, exc_type: Any, exc: Any, tb: Any) -> None:
         del exc_type, exc, tb
 
 
+class _FakeSession:
+    async def execute(self, stmt: object) -> object:
+        del stmt
+
+        class FakeResult:
+            def scalars(self) -> object:
+                class FakeScalars:
+                    def all(self) -> list[object]:
+                        return []
+
+                return FakeScalars()
+
+        return FakeResult()
+
+
 async def _fake_user_context(**kwargs: object) -> UserContext:
     del kwargs
     return UserContext(
@@ -81,6 +97,10 @@ async def test_run_agentscope_task_calls_runtime_run(
             "command": "run",
             "owner_id": str(uuid4()),
             "run_input": _run_input_payload(),
+            "runtime_config": {
+                "enabled_tools": [],
+                "context": {"window_mode": "day", "window_count": 2},
+            },
         }
     )
 
@@ -89,34 +109,28 @@ async def test_run_agentscope_task_calls_runtime_run(
 
 
 @pytest.mark.asyncio
-async def test_run_agentscope_task_includes_recent_context_messages(
+async def test_run_agentscope_task_injects_runtime_config(
     monkeypatch: pytest.MonkeyPatch,
 ) -> None:
-    captured_messages: list[dict[str, Any]] = []
+    captured_config: dict[str, Any] = {}
 
     class _FakeRuntime:
         def __init__(self, **kwargs: object) -> None:
             del kwargs
 
         async def run(self, **kwargs: object) -> object:
-            raw_context_messages = kwargs.get("context_messages")
-            raw_run_input = kwargs.get("run_input")
-            if isinstance(raw_context_messages, list):
-                captured_messages.extend(raw_context_messages)
-            if raw_run_input is not None:
-                raw_messages = getattr(raw_run_input, "messages", [])
-                if isinstance(raw_messages, list):
-                    captured_messages.extend(raw_messages)
+            captured_config.update(
+                {
+                    "runtime_config": kwargs.get("runtime_config"),
+                    "context_messages": kwargs.get("context_messages"),
+                }
+            )
             return object()
 
     async def _fake_get_redis_client() -> object:
         return object()
 
     async def _empty_context(**kwargs: object) -> list[dict[str, Any]]:
-        del kwargs
-        return []
-
-    async def _fake_context(**kwargs: object) -> list[dict[str, Any]]:
         del kwargs
         return [{"id": "ctx-1", "role": "assistant", "content": "历史上下文"}]
 
@@ -133,25 +147,23 @@ async def test_run_agentscope_task_includes_recent_context_messages(
         "_build_recent_context_messages",
         _empty_context,
     )
-    monkeypatch.setattr(
-        tasks_module,
-        "_build_recent_context_messages",
-        _fake_context,
-    )
 
-    run_input = _run_input_payload()
-    run_input["messages"] = [{"id": "u1", "role": "user", "content": "现在几点"}]
     await tasks_module.run_agentscope_task(
         {
             "command": "run",
             "owner_id": str(uuid4()),
-            "run_input": run_input,
+            "run_input": _run_input_payload(),
+            "runtime_config": {
+                "enabled_tools": [],
+                "context": {"window_mode": "day", "window_count": 2},
+            },
         }
     )
 
-    assert len(captured_messages) == 2
-    assert captured_messages[0]["id"] == "ctx-1"
-    assert getattr(captured_messages[1], "id", None) == "u1"
+    assert captured_config["context_messages"] == [
+        {"id": "ctx-1", "role": "assistant", "content": "历史上下文"}
+    ]
+    assert captured_config["runtime_config"] is not None
 
 
 @pytest.mark.asyncio
@@ -177,38 +189,6 @@ async def test_run_agentscope_task_rejects_invalid_command_type() -> None:
         )
 
 
-@pytest.mark.asyncio
-async def test_run_agentscope_task_requires_forwarded_props_agent_type() -> None:
-    payload = _run_input_payload()
-    payload["forwardedProps"] = {}
-
-    with pytest.raises(ValueError, match="invalid RunAgentInput.forwardedProps"):
-        await tasks_module.run_agentscope_task(
-            {
-                "command": "run",
-                "owner_id": str(uuid4()),
-                "run_input": payload,
-            }
-        )
-
-
-@pytest.mark.asyncio
-async def test_run_agentscope_task_memory_mode_requires_automation_job_id() -> None:
-    payload = _run_input_payload()
-    payload["forwardedProps"] = {"agent_type": "memory"}
-
-    with pytest.raises(
-        ValueError, match="automation_job_id is required for memory mode"
-    ):
-        await tasks_module.run_agentscope_task(
-            {
-                "command": "run",
-                "owner_id": str(uuid4()),
-                "run_input": payload,
-            }
-        )
-
-
 @pytest.mark.asyncio
 async def test_build_recent_context_messages_includes_all_user_attachments(
     monkeypatch: pytest.MonkeyPatch,
@@ -221,9 +201,9 @@ async def test_build_recent_context_messages_includes_all_user_attachments(
             self,
             *,
             thread_id: str,
-            system_agent_mode: str,
+            context_config: MemoryContextConfig,
         ) -> dict[str, object] | None:
-            del thread_id, system_agent_mode
+            del thread_id, context_config
             return {
                 "messages": [
                     {
@@ -257,7 +237,10 @@ async def test_build_recent_context_messages_includes_all_user_attachments(
     messages = await tasks_module._build_recent_context_messages(
         session=object(),
         thread_id=str(uuid4()),
-        context_mode="worker",
+        context_config=MemoryContextConfig(
+            window_mode=ContextWindowMode.DAY,
+            window_count=2,
+        ),
     )
 
     assert len(messages) == 1
@@ -281,9 +264,9 @@ async def test_build_recent_context_messages_uses_tool_metadata_output(
             self,
             *,
             thread_id: str,
-            system_agent_mode: str,
+            context_config: MemoryContextConfig,
         ) -> dict[str, object] | None:
-            del thread_id, system_agent_mode
+            del thread_id, context_config
             return {
                 "messages": [
                     {
@@ -312,7 +295,7 @@ async def test_build_recent_context_messages_uses_tool_metadata_output(
     messages = await tasks_module._build_recent_context_messages(
         session=object(),
         thread_id=str(uuid4()),
-        context_mode="worker",
+        context_config=MemoryContextConfig(),
     )
 
     assert len(messages) == 1
@@ -336,9 +319,9 @@ async def test_build_recent_context_messages_skips_tool_without_metadata_output(
             self,
             *,
             thread_id: str,
-            system_agent_mode: str,
+            context_config: MemoryContextConfig,
         ) -> dict[str, object] | None:
-            del thread_id, system_agent_mode
+            del thread_id, context_config
             return {
                 "messages": [
                     {
@@ -354,17 +337,17 @@ async def test_build_recent_context_messages_skips_tool_without_metadata_output(
     messages = await tasks_module._build_recent_context_messages(
         session=object(),
         thread_id=str(uuid4()),
-        context_mode="worker",
+        context_config=MemoryContextConfig(),
     )
 
     assert messages == []
 
 
 @pytest.mark.asyncio
-async def test_build_recent_context_messages_passes_context_mode_through(
+async def test_build_recent_context_messages_passes_context_config(
     monkeypatch: pytest.MonkeyPatch,
 ) -> None:
-    captured_mode: dict[str, str | None] = {"mode": None}
+    captured_config: dict[str, Any] = {"config": None}
 
     class _FakeContextService:
         def __init__(self, *, repository: object) -> None:
@@ -374,19 +357,21 @@ async def test_build_recent_context_messages_passes_context_mode_through(
             self,
             *,
             thread_id: str,
-            system_agent_mode: str,
+            context_config: MemoryContextConfig,
         ) -> dict[str, object] | None:
             del thread_id
-            captured_mode["mode"] = system_agent_mode
+            captured_config["config"] = context_config
             return None
 
     monkeypatch.setattr(tasks_module, "AgentContextService", _FakeContextService)
 
+    cfg = MemoryContextConfig(window_mode=ContextWindowMode.NUMBER, window_count=10)
     messages = await tasks_module._build_recent_context_messages(
         session=object(),
         thread_id=str(uuid4()),
-        context_mode="worker",
+        context_config=cfg,
     )
 
     assert messages == []
-    assert captured_mode["mode"] == "worker"
+    assert captured_config["config"].window_mode == ContextWindowMode.NUMBER
+    assert captured_config["config"].window_count == 10

backend/tests/unit/core/agentscope/test_agent_prompt.py

@@ -23,19 +23,17 @@ def test_build_agent_prompt_for_worker_contains_runtime_config() -> None:
     assert "enabled_tools=calendar.read,calendar.write" in prompt
 
 
-def test_build_agent_prompt_for_memory_uses_memory_rules() -> None:
+def test_build_agent_prompt_for_router_contains_task_typing_rules() -> None:
     prompt = build_agent_prompt(
-        agent_type=AgentType.MEMORY,
+        agent_type=AgentType.ROUTER,
         llm_config=SystemAgentLLMConfig.model_validate(
             {
                 "context_messages": {"mode": "day", "count": 2},
-                "enabled_tools": ["user.lookup"],
             }
         ),
     )
 
-    assert "- type: memory" in prompt
-    assert "[Memory Agent]" in prompt
+    assert "- type: router" in prompt
+    assert "[Router Agent]" in prompt
     assert "context_messages.mode=day" in prompt
     assert "context_messages.count=2" in prompt
-    assert "enabled_tools=user.lookup" in prompt

backend/tests/unit/core/agentscope/test_system_prompt.py

@@ -156,3 +156,48 @@ def test_build_system_prompt_keeps_sections_focused_without_language_duplication
     assert "[Answer Style]" in prompt
     assert "Default reply language:" not in prompt
     assert "Follow agent contracts strictly" not in prompt
+
+
+def test_build_system_prompt_includes_memory_section_when_memories_provided() -> None:
+    from schemas.memories import (
+        MemoryContext,
+        MemoryListResponse,
+        MemorySource,
+        MemoryType,
+    )
+
+    memories = MemoryListResponse(
+        owner_id=uuid4(),
+        memories=[
+            MemoryContext(
+                memory_type=MemoryType.USER,
+                source=MemorySource.MANUAL,
+                title="User prefers morning meetings",
+                content={"text": "User likes meetings before 10am"},
+                created_at=datetime(2026, 3, 1, tzinfo=timezone.utc),
+                updated_at=datetime(2026, 3, 1, tzinfo=timezone.utc),
+            ),
+        ],
+        total=1,
+    )
+
+    prompt = build_system_prompt(
+        agent_type=AgentType.WORKER,
+        user_context=_build_user_context(),
+        now_utc=datetime(2026, 3, 11, 0, 0, tzinfo=timezone.utc),
+        memories=memories,
+    )
+
+    assert "<!-- MEMORY_START -->" in prompt
+    assert "[User Memories]" in prompt
+    assert "User prefers morning meetings" in prompt
+
+
+def test_build_system_prompt_omits_memory_section_when_no_memories() -> None:
+    prompt = build_system_prompt(
+        agent_type=AgentType.WORKER,
+        user_context=_build_user_context(),
+        now_utc=datetime(2026, 3, 11, 0, 0, tzinfo=timezone.utc),
+    )
+
+    assert "<!-- MEMORY_START -->" not in prompt

backend/tests/unit/core/automation/test_scheduler.py

@@ -5,17 +5,23 @@ from uuid import UUID, uuid4
 
 import pytest
 
-from core.automation.scheduler import (
-    AutomationSchedulerService,
-    _compute_next_run_at,
+from models.automation_jobs import AutomationJob as OrmAutomationJob, ScheduleType
+from schemas.automation import (
+    RuntimeConfig,
 )
-from models.automation_jobs import ScheduleType
-from schemas.automation.config import AutomationJobConfig
-from schemas.automation.scheduler import DueAutomationJob
+from v1.automation_jobs.service import AutomationJobsService, _compute_next_run_at
+
+
+class _FakeSession:
+    async def commit(self) -> None:
+        pass
+
+    async def rollback(self) -> None:
+        pass
 
 
 class _FakeRepository:
-    def __init__(self, jobs: list[DueAutomationJob]) -> None:
+    def __init__(self, jobs: list[OrmAutomationJob]) -> None:
         self.jobs = jobs
         self.marked: list[tuple[UUID, datetime, datetime]] = []
         self.commits = 0
@@ -23,30 +29,14 @@ class _FakeRepository:
 
     async def list_due_jobs(
         self, *, now_utc: datetime, limit: int
-    ) -> list[DueAutomationJob]:
+    ) -> list[OrmAutomationJob]:
         del now_utc
         return self.jobs[:limit]
 
-    async def get_job_config(self, *, job_id: UUID) -> AutomationJobConfig:
-        del job_id
-        return AutomationJobConfig.model_validate(
-            {
-                "agent_type": "memory",
-                "model_code": "qwen3.5-flash",
-                "enabled_tools": ["calendar.read", "user.lookup"],
-                "input_template": "auto input",
-                "context": {
-                    "source": "latest_chat",
-                    "window_mode": "day",
-                    "window_count": 2,
-                },
-            }
-        )
-
-    async def ensure_latest_chat_session(self, *, owner_id: UUID) -> UUID:
+    async def get_or_create_chat_session(self, *, owner_id: UUID) -> UUID:
         return owner_id
 
-    async def mark_job_dispatched(
+    async def update_job_schedule(
         self,
         *,
         job_id: UUID,
@@ -55,57 +45,65 @@ class _FakeRepository:
     ) -> None:
         self.marked.append((job_id, next_run_at, last_run_at))
 
-    async def commit(self) -> None:
-        self.commits += 1
 
-    async def rollback(self) -> None:
-        self.rollbacks += 1
-
-
-class _FakeQueue:
-    def __init__(self) -> None:
-        self.commands: list[dict[str, object]] = []
-
-    async def enqueue(
-        self,
-        *,
-        command: dict[str, object],
-        dedup_key: str | None,
-    ) -> str:
-        del dedup_key
-        self.commands.append(command)
-        return "task-1"
+def _make_orm_job(
+    *,
+    job_id: UUID | None = None,
+    owner_id: UUID | None = None,
+    schedule_type: ScheduleType = ScheduleType.DAILY,
+    next_run_at: datetime | None = None,
+) -> OrmAutomationJob:
+    now = datetime(2026, 3, 19, 12, 0, tzinfo=timezone.utc)
+    return OrmAutomationJob(
+        id=job_id or uuid4(),
+        owner_id=owner_id or uuid4(),
+        title="Test Job",
+        config={
+            "enabled_tools": ["calendar.read", "user.lookup"],
+            "context": {
+                "source": "latest_chat",
+                "window_mode": "day",
+                "window_count": 2,
+            },
+            "input_template": "auto input: {date}",
+        },
+        schedule_type=schedule_type,
+        run_at=now - timedelta(hours=1),
+        next_run_at=next_run_at or now - timedelta(minutes=1),
+        timezone="UTC",
+        last_run_at=None,
+        status="active",
+        created_by=None,
+        created_at=now - timedelta(days=1),
+        updated_at=now - timedelta(hours=1),
+        deleted_at=None,
+    )
 
 
 @pytest.mark.asyncio
-async def test_scan_and_dispatch_enqueues_memory_run_command() -> None:
+async def test_scan_and_dispatch_calls_dispatch_fn_with_runtime_config() -> None:
     now = datetime(2026, 3, 19, 12, 0, tzinfo=timezone.utc)
     owner_id = uuid4()
     job_id = uuid4()
-    repo = _FakeRepository(
-        jobs=[
-            DueAutomationJob(
-                id=job_id,
-                owner_id=owner_id,
-                schedule_type=ScheduleType.DAILY,
-                timezone="UTC",
-                next_run_at=now - timedelta(minutes=1),
-            )
-        ]
-    )
-    queue = _FakeQueue()
-    service = AutomationSchedulerService(repository=repo, queue=queue)
+    repo = _FakeRepository(jobs=[_make_orm_job(job_id=job_id, owner_id=owner_id)])
+    dispatched_calls: list[dict] = []
 
-    result = await service.scan_and_dispatch(now_utc=now, limit=10)
+    async def dispatch_fn(**kwargs: object) -> None:
+        dispatched_calls.append(kwargs)
+
+    service = AutomationJobsService(repository=repo, session=_FakeSession())
+
+    result = await service.scan_and_dispatch(
+        now_utc=now, limit=10, dispatch_fn=dispatch_fn
+    )
 
     assert result.scanned == 1
     assert result.dispatched == 1
-    assert len(queue.commands) == 1
-    run_input = queue.commands[0]["run_input"]
-    assert isinstance(run_input, dict)
-    assert run_input["forwardedProps"] == {"agent_type": "memory"}
-    assert queue.commands[0]["automation_job_id"] == str(job_id)
-    assert repo.commits == 1
+    assert len(dispatched_calls) == 1
+    assert dispatched_calls[0]["owner_id"] == owner_id
+    assert dispatched_calls[0]["runtime_config"] is not None
+    cfg: RuntimeConfig = dispatched_calls[0]["runtime_config"]
+    assert len(cfg.enabled_tools) == 2
 
 
 def test_compute_next_run_at_daily() -> None:

docs/protocols/agent/run-agent-input.md

@@ -202,10 +202,12 @@ interface ForwardedProps {
 
 ### 运行模式说明
 
-| runtime_mode | 说明 | 后端 Pipeline |
-|--------------|------|---------------|
-| `chat` | 标准对话模式 | `router` -> `worker` |
-| `automation` | 自动化任务模式 | 由后端业务逻辑决定具体 Agent 类型 |
+| runtime_mode | 说明 | Pipeline | 差异 |
+|--------------|------|----------|------|
+| `chat` | 标准对话模式 | `router` -> `worker` | `enabled_tools` 和 `context` 来自 `system_agents.yaml` |
+| `automation` | 自动化任务模式 | `router` -> `worker` | `enabled_tools` 和 `context` 来自 `AutomationJob.config`（通过 `runtime_config` 注入）|
+
+> `runtime_mode` 仅影响 `RuntimeConfig`（工具列表与上下文配置），不改变执行阶段。两模式均使用固定两阶段 pipeline。
 
 ### 时间来源优先级（固定）
 

docs/protocols/agent/sse-events.md

@@ -326,6 +326,48 @@ cost = uncached_prompt_tokens * input_cost_per_token
 
 ## 8) 可见性与上下文装载说明
 
-- 持久化消息使用单字段 `visibility_mask`（位掩码）控制 consumer 可见性。
-- `/history` 仅投影 `ui.history` 可见消息。
-- 运行时上下文按当前 stage 对应 consumer 位过滤装载，不依赖前端展示可见性。
+### visibility_mask 位掩码系统
+
+持久化消息使用单字段 `visibility_mask`（位掩码）控制不同 consumer 的可见性：
+
+| Bit | 常量名 | 说明 |
+|-----|--------|------|
+| 0 | `UI_HISTORY` | `/history` API 投影可见的消息 |
+| 1 | `CONTEXT_ASSEMBLY` | 运行时上下文装配（context assembly）可见 |
+
+> 新消息入库时，`chat` 模式设置 `mask = UI_HISTORY | CONTEXT_ASSEMBLY`（值为 3），`automation` 模式设置 `mask = 0`。
+
+### /history API
+
+`GET /api/v1/agent/history` 仅投影包含 `UI_HISTORY` 位的消息：
+
+```sql
+WHERE (visibility_mask & 1) != 0
+```
+
+### 运行时上下文装配
+
+`load_context_messages` 查询上下文时使用 `CONTEXT_ASSEMBLY` 位过滤：
+
+```sql
+WHERE (visibility_mask & 2) != 0
+```
+
+**影响**：
+- `chat` 模式用户输入：mask=3 → 进入 `/history` ✅，进入 context assembly ✅
+- `automation` 模式用户输入：mask=0 → 进入 `/history` ❌，进入 context assembly ❌
+
+### Automation 模式上下文注入
+
+由于 automation 用户输入 `mask=0` 不进入 context assembly，router 调用前会从 `RunAgentInput.messages` 注入最新用户消息到 context 头部（条件：context 为空 或 最后一条非 user）。
+
+### runtime_mode 差异总结
+
+| 维度 | `chat` | `automation` |
+|------|--------|--------------|
+| Pipeline | `router` -> `worker` | `router` -> `worker` |
+| 用户输入 visibility_mask | `UI_HISTORY \| CONTEXT_ASSEMBLY` | `0` |
+| 进入 /history | ✅ | ❌ |
+| 进入 context assembly | ✅（自动） | ❌（通过 run_input 注入） |
+| enabled_tools 来源 | `system_agents.yaml` worker 配置 | `AutomationJob.config.enabled_tools` |
+| context 配置来源 | `system_agents.yaml` router context_messages | `AutomationJob.config.context` |