fix(subagents): raise general-purpose max_turns to 150 and default timeout to 30min (#3610)

* fix(subagents): raise general-purpose max_turns to 150 and default timeout to 30min

Deep-research subtasks failed out of the box with GraphRecursionError (Recursion limit of 100 reached): the built-in general-purpose subagent caps at max_turns=100. Raise it to 150 and bump the default subagent timeout from 900s (15min) to 1800s (30min) so the extra turns have time to run instead of shifting the failure to a timeout. The lead agent recursion_limit (100) is unchanged; the failures are subagent-only.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

* docs(subagents): clarify lead recursion_limit is independent of subagent max_turns

Add comments at both lead recursion_limit=100 sites (gateway services + channel manager) explaining the lead's LangGraph super-step budget is separate from subagent depth, so the two 100s are not conflated. Comment-only, no behavior change.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

* docs(subagents): clarify built-in vs custom timeout scope; pin bash max_turns in test

Review follow-ups: (1) clarify SubagentConfig docstring + global timeout field/comment that the 1800 default applies to built-in subagents (custom agents keep their own timeout_seconds); (2) pin bash.max_turns==60 in the defaults regression test so the config.example.yaml doc cannot drift; (3) rename test_default_timeout_preserved_when_no_config -> test_explicit_global_timeout_propagates_to_general_purpose since it intentionally exercises an explicit non-default 900. No runtime behavior change.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

---------

Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>

backend/CLAUDE.md

@@ -311,7 +311,7 @@ Proxied through nginx: `/api/langgraph/*` → Gateway LangGraph-compatible runti
 
 **Built-in Agents**: `general-purpose` (all tools except `task`) and `bash` (command specialist)
 **Execution**: Dual thread pool - `_scheduler_pool` (3 workers) + `_execution_pool` (3 workers)
-**Concurrency**: `MAX_CONCURRENT_SUBAGENTS = 3` enforced by `SubagentLimitMiddleware` (truncates excess tool calls in `after_model`), 15-minute timeout
+**Concurrency**: `MAX_CONCURRENT_SUBAGENTS = 3` enforced by `SubagentLimitMiddleware` (truncates excess tool calls in `after_model`); default subagent timeout `subagents.timeout_seconds=1800` (30 min) and built-in `general-purpose` `max_turns=150` (raised from 100/15-min so deep-research subtasks stop hitting `GraphRecursionError` out of the box)
 **Flow**: `task()` tool → `SubagentExecutor` → background thread → poll 5s → SSE events → result
 **Events**: `task_started`, `task_running`, `task_completed`/`task_failed`/`task_timed_out`
 **Deferred MCP tools** (if `tool_search.enabled`): `SubagentExecutor._build_initial_state` assembles deferral after policy filtering via the shared `assemble_deferred_tools` (fail-closed), appends the `tool_search` tool, injects the `<available-deferred-tools>` section into the subagent's `SystemMessage`, and threads the setup to `_create_agent`, which attaches `DeferredToolFilterMiddleware` through `build_subagent_runtime_middlewares(deferred_setup=...)`. Subagents thus withhold full MCP schemas until promotion, same as the lead agent; each task run gets a fresh `ThreadState` so promotion is isolated per run

backend/app/channels/manager.py

@@ -42,6 +42,11 @@ DEFAULT_GATEWAY_URL = "http://localhost:8001"
 DEFAULT_ASSISTANT_ID = "lead_agent"
 CUSTOM_AGENT_NAME_PATTERN = re.compile(r"^[A-Za-z0-9-]+$")
 
+# Lead-agent recursion budget (LangGraph super-steps for the lead graph only).
+# This is independent of subagent depth: a `task()` dispatch runs the whole
+# subagent inside ONE lead tools-node step, and subagents enforce their own
+# limit via `subagents.max_turns` (see SubagentExecutor). Do not conflate this
+# 100 with the general-purpose subagent's max_turns.
 DEFAULT_RUN_CONFIG: dict[str, Any] = {"recursion_limit": 100}
 DEFAULT_RUN_CONTEXT: dict[str, Any] = {
     "thinking_enabled": True,

backend/app/gateway/services.py

@@ -224,6 +224,11 @@ def build_run_config(
     the LangGraph Platform-compatible HTTP API and the IM channel path behave
     identically.
     """
+    # Lead-agent recursion budget (LangGraph super-steps for the lead graph
+    # only). Independent of subagent depth: a `task()` dispatch runs the whole
+    # subagent inside ONE lead tools-node step, and subagents enforce their own
+    # limit via `subagents.max_turns`. Do not conflate this 100 with the
+    # general-purpose subagent's max_turns.
     config: dict[str, Any] = {"recursion_limit": 100}
     if request_config:
         # LangGraph >= 0.6.0 introduced ``context`` as the preferred way to

backend/packages/harness/deerflow/config/subagents_config.py

@@ -72,9 +72,9 @@ class SubagentsAppConfig(BaseModel):
     """Configuration for the subagent system."""
 
     timeout_seconds: int = Field(
-        default=900,
+        default=1800,
         ge=1,
-        description="Default timeout in seconds for all subagents (default: 900 = 15 minutes)",
+        description="Default timeout in seconds for built-in subagents (default: 1800 = 30 minutes); custom agents use their own timeout_seconds unless given a per-agent override",
     )
     max_turns: int | None = Field(
         default=None,

backend/packages/harness/deerflow/subagents/builtins/general_purpose.py

@@ -57,5 +57,5 @@ You have access to the same sandbox environment as the parent agent:
     tools=None,  # Inherit all tools from parent
     disallowed_tools=["task", "ask_clarification", "present_files"],  # Prevent nesting and clarification
     model="inherit",
-    max_turns=100,
+    max_turns=150,
 )

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

@@ -20,8 +20,13 @@ class SubagentConfig:
         skills: Optional list of skill names to load. If None, inherits all enabled skills.
                 If an empty list, no skills are loaded.
         model: Model to use - 'inherit' uses parent's model.
-        max_turns: Maximum number of agent turns before stopping.
-        timeout_seconds: Maximum execution time in seconds (default: 900 = 15 minutes).
+        max_turns: Maximum agent turns before stopping. Built-in agents use the
+            value set here (general-purpose=150, bash=60) unless the global
+            ``subagents.max_turns`` is set.
+        timeout_seconds: Bare fallback execution-time cap. For built-in agents the
+            effective limit is the global ``subagents.timeout_seconds`` (default
+            1800 = 30 min), layered on by the registry; this 900 only applies
+            when no differing global value exists.
     """
 
     name: str

backend/tests/test_app_config_reload.py

@@ -316,7 +316,7 @@ def test_get_app_config_resets_singleton_configs_when_sections_removed(tmp_path,
         assert get_title_config().enabled is True
         assert get_summarization_config().enabled is False
         assert get_memory_config().enabled is True
-        assert get_subagents_app_config().timeout_seconds == 900
+        assert get_subagents_app_config().timeout_seconds == 1800
         assert get_tool_search_config().enabled is False
         assert get_guardrails_config().enabled is False
         assert get_checkpointer_config() is None

backend/tests/test_subagent_timeout_config.py

@@ -97,7 +97,7 @@ class TestSubagentOverrideConfig:
 class TestSubagentsAppConfigDefaults:
     def test_default_timeout(self):
         config = SubagentsAppConfig()
-        assert config.timeout_seconds == 900
+        assert config.timeout_seconds == 1800
 
     def test_default_max_turns_override_is_none(self):
         config = SubagentsAppConfig()
@@ -281,7 +281,7 @@ class TestLoadSubagentsConfig:
     def test_load_empty_dict_uses_defaults(self):
         load_subagents_config_from_dict({})
         cfg = get_subagents_app_config()
-        assert cfg.timeout_seconds == 900
+        assert cfg.timeout_seconds == 1800
         assert cfg.max_turns is None
         assert cfg.agents == {}
 
@@ -319,13 +319,30 @@ class TestRegistryGetSubagentConfig:
         assert get_subagent_config("general-purpose") is not None
         assert get_subagent_config("bash") is not None
 
-    def test_default_timeout_preserved_when_no_config(self):
+    def test_explicit_global_timeout_propagates_to_general_purpose(self):
+        """An explicit global timeout (here the non-default 900) propagates to a
+        built-in agent, while max_turns still comes from the builtin def (150).
+        """
         from deerflow.subagents.registry import get_subagent_config
 
         _reset_subagents_config(timeout_seconds=900)
         config = get_subagent_config("general-purpose")
         assert config.timeout_seconds == 900
-        assert config.max_turns == 100
+        assert config.max_turns == 150
+
+    def test_builtin_defaults_have_research_headroom(self):
+        """Out-of-box defaults (no config.yaml subagents section) must give
+        general-purpose enough turns/time for deep research, which previously
+        failed with GraphRecursionError at the old max_turns=100 limit.
+        """
+        from deerflow.subagents.registry import get_subagent_config
+
+        load_subagents_config_from_dict({})  # no subagents config -> model defaults
+        config = get_subagent_config("general-purpose")
+        assert config.max_turns == 150
+        assert config.timeout_seconds == 1800
+        # Pin bash too so the config.example.yaml "bash=60" doc cannot drift.
+        assert get_subagent_config("bash").max_turns == 60
 
     def test_global_timeout_override_applied(self):
         from deerflow.subagents.registry import get_subagent_config