refactor: thread app_config through middleware factories (#2652)

* refactor: thread app_config through middleware factories

Continues the incremental config-refactor sequence (#2611 root, #2612 lead
path) one layer deeper into the middleware factories. Two ambient lookups
inside _build_runtime_middlewares are eliminated and the LLMErrorHandling
band-aid removed:

- _build_runtime_middlewares / build_lead_runtime_middlewares /
  build_subagent_runtime_middlewares now require app_config: AppConfig.
- get_guardrails_config() inside the factory is replaced with
  app_config.guardrails (semantically identical — same default-factory
  GuardrailsConfig — verified by direct equality check).
- LLMErrorHandlingMiddleware.__init__ now requires app_config and reads
  circuit_breaker fields directly. The class-level
  circuit_failure_threshold / circuit_recovery_timeout_sec defaults are
  removed along with the try/except (FileNotFoundError, RuntimeError):
  pass band-aid — the let-it-crash invariant the rest of the refactor
  enforces.

Caller chain (already-resolved app_config sources):
- _build_middlewares in lead_agent/agent.py: reorder so
  resolved_app_config = app_config or get_app_config() is computed BEFORE
  build_lead_runtime_middlewares is called, then passed as kwarg.
- SubagentExecutor: optional app_config parameter (mirrors the lead-agent
  pattern); _create_agent does the same `or get_app_config()` fallback at
  agent-build time, so task_tool callers don't need to plumb app_config
  through yet (typed-context plumbing for tool runtimes is a separate
  refactor).

Tests:
- test_llm_error_handling_middleware: _make_app_config helper using
  AppConfig(sandbox=SandboxConfig(use="test")) — same minimal-config
  pattern conftest already uses. Three direct LLMErrorHandlingMiddleware()
  calls each followed by post-construction circuit_breaker mutation fold
  cleanly into _build_middleware(circuit_failure_threshold=...,
  circuit_recovery_timeout_sec=...).

Verification:
- tests/test_llm_error_handling_middleware.py — 14 passed
- tests/test_subagent_executor.py — 28 passed
- tests/test_tool_error_handling_middleware.py — 6 passed
- tests/test_task_tool_core_logic.py — 18 passed (verifies task_tool
  unchanged behavior)
- Full suite: 2697 passed, 3 skipped. The single intermittent failure in
  tests/test_client_e2e.py::test_tool_call_produces_events is pre-existing
  LLM flakiness (the test asserts the model decided to call a tool;
  reproduces 1/3 on unchanged main as well).

* fix: address middleware app config review comments

* fix: satisfy app config annotation lint

* test: cover explicit app config middleware wiring

---------

Co-authored-by: greatmengqi <chenmengqi.0376@bytedance.com>

backend/packages/harness/deerflow/agents/lead_agent/agent.py

@@ -259,8 +259,8 @@ def _build_middlewares(
     Returns:
         List of middleware instances.
     """
-    middlewares = build_lead_runtime_middlewares(lazy_init=True)
     resolved_app_config = app_config or get_app_config()
+    middlewares = build_lead_runtime_middlewares(app_config=resolved_app_config, lazy_init=True)
 
     # Add summarization middleware if enabled
     summarization_middleware = _create_summarization_middleware(app_config=resolved_app_config)

backend/packages/harness/deerflow/agents/middlewares/llm_error_handling_middleware.py

@@ -20,7 +20,7 @@ from langchain.agents.middleware.types import (
 from langchain_core.messages import AIMessage
 from langgraph.errors import GraphBubbleUp
 
-from deerflow.config import get_app_config
+from deerflow.config.app_config import AppConfig
 
 logger = logging.getLogger(__name__)
 
@@ -70,20 +70,11 @@ class LLMErrorHandlingMiddleware(AgentMiddleware[AgentState]):
     retry_base_delay_ms: int = 1000
     retry_cap_delay_ms: int = 8000
 
-    circuit_failure_threshold: int = 5
-    circuit_recovery_timeout_sec: int = 60
-
-    def __init__(self, **kwargs: Any) -> None:
+    def __init__(self, *, app_config: AppConfig, **kwargs: Any) -> None:
         super().__init__(**kwargs)
 
-        # Load Circuit Breaker configs from app config if available, fall back to defaults
-        try:
-            app_config = get_app_config()
-            self.circuit_failure_threshold = app_config.circuit_breaker.failure_threshold
-            self.circuit_recovery_timeout_sec = app_config.circuit_breaker.recovery_timeout_sec
-        except (FileNotFoundError, RuntimeError):
-            # Gracefully fall back to class defaults in test environments
-            pass
+        self.circuit_failure_threshold = app_config.circuit_breaker.failure_threshold
+        self.circuit_recovery_timeout_sec = app_config.circuit_breaker.recovery_timeout_sec
 
         # Circuit Breaker state
         self._circuit_lock = threading.Lock()

backend/packages/harness/deerflow/agents/middlewares/tool_error_handling_middleware.py

@@ -11,6 +11,8 @@ from langgraph.errors import GraphBubbleUp
 from langgraph.prebuilt.tool_node import ToolCallRequest
 from langgraph.types import Command
 
+from deerflow.config.app_config import AppConfig
+
 logger = logging.getLogger(__name__)
 
 _MISSING_TOOL_CALL_ID = "missing_tool_call_id"
@@ -67,6 +69,7 @@ class ToolErrorHandlingMiddleware(AgentMiddleware[AgentState]):
 
 def _build_runtime_middlewares(
     *,
+    app_config: AppConfig,
     include_uploads: bool,
     include_dangling_tool_call_patch: bool,
     lazy_init: bool = True,
@@ -91,12 +94,10 @@ def _build_runtime_middlewares(
 
         middlewares.append(DanglingToolCallMiddleware())
 
-    middlewares.append(LLMErrorHandlingMiddleware())
+    middlewares.append(LLMErrorHandlingMiddleware(app_config=app_config))
 
     # Guardrail middleware (if configured)
-    from deerflow.config.guardrails_config import get_guardrails_config
-
-    guardrails_config = get_guardrails_config()
+    guardrails_config = app_config.guardrails
     if guardrails_config.enabled and guardrails_config.provider:
         import inspect
 
@@ -125,18 +126,20 @@ def _build_runtime_middlewares(
     return middlewares
 
 
-def build_lead_runtime_middlewares(*, lazy_init: bool = True) -> list[AgentMiddleware]:
+def build_lead_runtime_middlewares(*, app_config: AppConfig, lazy_init: bool = True) -> list[AgentMiddleware]:
     """Middlewares shared by lead agent runtime before lead-only middlewares."""
     return _build_runtime_middlewares(
+        app_config=app_config,
         include_uploads=True,
         include_dangling_tool_call_patch=True,
         lazy_init=lazy_init,
     )
 
 
-def build_subagent_runtime_middlewares(*, lazy_init: bool = True) -> list[AgentMiddleware]:
+def build_subagent_runtime_middlewares(*, app_config: AppConfig, lazy_init: bool = True) -> list[AgentMiddleware]:
     """Middlewares shared by subagent runtime before subagent-only middlewares."""
     return _build_runtime_middlewares(
+        app_config=app_config,
         include_uploads=False,
         include_dangling_tool_call_patch=True,
         lazy_init=lazy_init,

backend/packages/harness/deerflow/subagents/executor.py

@@ -17,6 +17,7 @@ from langchain_core.messages import AIMessage, HumanMessage, SystemMessage
 from langchain_core.runnables import RunnableConfig
 
 from deerflow.agents.thread_state import SandboxState, ThreadDataState, ThreadState
+from deerflow.config.app_config import AppConfig
 from deerflow.models import create_chat_model
 from deerflow.subagents.config import SubagentConfig
 
@@ -132,6 +133,7 @@ class SubagentExecutor:
         self,
         config: SubagentConfig,
         tools: list[BaseTool],
+        app_config: AppConfig | None = None,
         parent_model: str | None = None,
         sandbox_state: SandboxState | None = None,
         thread_data: ThreadDataState | None = None,
@@ -143,6 +145,9 @@ class SubagentExecutor:
         Args:
             config: Subagent configuration.
             tools: List of all available tools (will be filtered).
+            app_config: Resolved AppConfig; threaded into middleware factories
+                at agent-build time. When None, ``_create_agent`` falls back to
+                ``get_app_config()`` (matches the lead-agent factory's pattern).
             parent_model: The parent agent's model name for inheritance.
             sandbox_state: Sandbox state from parent agent.
             thread_data: Thread data from parent agent.
@@ -150,6 +155,7 @@ class SubagentExecutor:
             trace_id: Trace ID from parent for distributed tracing.
         """
         self.config = config
+        self.app_config = app_config
         self.parent_model = parent_model
         self.sandbox_state = sandbox_state
         self.thread_data = thread_data
@@ -168,13 +174,17 @@ class SubagentExecutor:
 
     def _create_agent(self):
         """Create the agent instance."""
+        # Mirror lead-agent factory pattern: prefer explicit app_config,
+        # fall back to ambient lookup at agent-build time.
+        from deerflow.config import get_app_config
+
+        resolved_app_config = self.app_config or get_app_config()
         model_name = _get_model_name(self.config, self.parent_model)
-        model = create_chat_model(name=model_name, thinking_enabled=False)
+        model = create_chat_model(name=model_name, thinking_enabled=False, app_config=resolved_app_config)
 
         from deerflow.agents.middlewares.tool_error_handling_middleware import build_subagent_runtime_middlewares
 
-        # Reuse shared middleware composition with lead agent.
-        middlewares = build_subagent_runtime_middlewares(lazy_init=True)
+        middlewares = build_subagent_runtime_middlewares(app_config=resolved_app_config, lazy_init=True)
 
         return create_agent(
             model=model,