feat(coordinator): 添加团队节点默认最大工具迭代次数配置

添加 DEFAULT_TEAM_NODE_MAX_TOOL_ITERATIONS 配置项以控制团队节点的最大工具迭代次数，
并修改 LocalAgentRunner 中的逻辑来使用此默认值当 envelope 中未指定时。

fix(runtime): 修复团队节点运行成功判断逻辑

更新运行成功判断条件，将 finish_reason 为 "max_tool_iterations_finalized" 的情况
视为运行失败，并添加对原始工具调用输出的检测，避免将其误判为成功完成。

feat(mcp): 添加团队工作流MCP工具类别支持

增加新的本地MCP工具类别 "team_workflow" 及其对应的工具创建功能，
为团队工作流提供本地工具支持。

refactor(engine): 调整AgentLoop最大工具迭代次数设置

将 AgentProfile 中的默认 max_tool_iterations 从 30 增加到 100，
同时移除 TaskExecutionPlanner 构造函数中的重复参数传递。

perf(mcp): 优化MCP连接管理避免重复连接

添加 mcp_connected 标志来跟踪MCP连接状态，确保 connect_all 只执行一次，
提高性能并避免不必要的重复连接。

refactor(skills): 移除技能团队模板相关功能

移除与技能团队模板相关的代码，包括解析、存储和处理逻辑，
简化技能记录结构和加载流程。

feat(process): 增强会话过程投影器功能

添加技能激活快照事件处理，改进团队运行完成消息显示，
并增强技能激活事件的时间戳记录功能。

refactor(tasks): 简化任务尝试编排器团队执行逻辑

移除团队执行相关代码，将所有任务统一按单步执行处理，
简化任务编排器的复杂度并提升执行效率。

fix(evidence): 修复节点证据评估中需求验证逻辑

更新节点证据评估逻辑，跳过自然语言证据需求的确定性验证，
只执行机器可读的需求验证，避免因自然语言需求导致的节点失败。

app-instance/backend/tests/unit/test_agent_loop.py

@@ -1,10 +1,13 @@
 import asyncio
 import json
 from contextlib import suppress
+from types import SimpleNamespace
 from typing import Any
 
 from beaver.engine import AgentLoop, AgentRunResult, EngineLoader
 from beaver.engine import loop as loop_module
+from beaver.engine.providers.base import LLMProvider, LLMResponse
+from beaver.engine.providers.factory import ProviderBundle
 
 
 def _run_result(run_id: str, output_text: str) -> AgentRunResult:
@@ -49,7 +52,7 @@ def test_running_loop_handles_reentrant_submit_direct(tmp_path) -> None:
     asyncio.run(run_case())
 
 
-def test_web_search_loop_guard_stops_after_repeated_low_quality_results() -> None:
+def test_web_search_loop_guard_keeps_successful_low_quality_results_available() -> None:
     guard = loop_module._WebSearchLoopGuard()
     low_quality = json.dumps(
         {
@@ -63,21 +66,106 @@ def test_web_search_loop_guard_stops_after_repeated_low_quality_results() -> Non
     assert guard.observe_result("web_search", low_quality) is None
     assert guard.observe_result("web_search", low_quality) is None
 
-    guidance = guard.observe_result("web_search", low_quality)
-
-    assert guidance is not None
-    assert guidance["finish_reason"] == "web_search_low_quality_budget"
-    assert "weather beijing" in guidance["message"]
+    assert guard.observe_result("web_search", low_quality) is None
 
 
 def test_web_search_loop_guard_resets_after_useful_result() -> None:
     guard = loop_module._WebSearchLoopGuard()
-    low_quality = json.dumps({"success": True, "query": "weather", "quality": "low", "results": []})
+    failed_search = json.dumps({"success": False, "query": "weather", "error": "timeout"})
     useful = json.dumps({"success": True, "query": "weather", "quality": "high", "results": []})
 
-    assert guard.observe_result("web_search", low_quality) is None
+    assert guard.observe_result("web_search", failed_search) is None
     assert guard.observe_result("web_search", useful) is None
-    assert guard.observe_result("web_search", low_quality) is None
-    assert guard.observe_result("web_search", low_quality) is None
+    assert guard.observe_result("web_search", failed_search) is None
+    assert guard.observe_result("web_search", failed_search) is None
 
-    assert guard.observe_result("web_search", low_quality) is not None
+    assert guard.observe_result("web_search", failed_search) is not None
+
+
+class RecordingProvider(LLMProvider):
+    def __init__(self) -> None:
+        super().__init__()
+        self.tool_names_by_call: list[list[str]] = []
+
+    async def chat(
+        self,
+        messages: list[dict],
+        tools: list[dict] | None = None,
+        model: str | None = None,
+        max_tokens: int | None = None,
+        temperature: float = 0.7,
+        thinking_enabled: bool | None = None,
+    ) -> LLMResponse:
+        self.tool_names_by_call.append(
+            [
+                str(tool.get("function", {}).get("name") or tool.get("name"))
+                for tool in tools or []
+            ]
+        )
+        return LLMResponse("done", provider_name="stub", model="stub-model")
+
+    def get_default_model(self) -> str:
+        return "stub-model"
+
+
+def _bundle(provider: RecordingProvider) -> ProviderBundle:
+    return ProviderBundle(
+        main_runtime=SimpleNamespace(model="stub-model", provider_name="stub"),
+        main_provider=provider,
+    )
+
+
+def test_task_mode_root_does_not_expose_agent_team_tool(tmp_path) -> None:
+    provider = RecordingProvider()
+    loop = AgentLoop(loader=EngineLoader(workspace=tmp_path))
+
+    asyncio.run(
+        loop.process_direct(
+            "compare financial reports",
+            session_id="session",
+            task_id="task-1",
+            task_mode=True,
+            include_skill_assembly=False,
+            provider_bundle=_bundle(provider),
+        )
+    )
+
+    assert "run_agent_team" not in provider.tool_names_by_call[0]
+
+
+def test_default_engine_registry_does_not_register_agent_team_tool(tmp_path) -> None:
+    loaded = AgentLoop(loader=EngineLoader(workspace=tmp_path)).boot()
+
+    assert loaded.tool_registry is not None
+    assert loaded.tool_registry.get("run_agent_team") is None
+    assert "run_agent_team" not in loaded.tools
+
+
+def test_non_task_and_team_node_do_not_expose_agent_team_tool(tmp_path) -> None:
+    non_task_provider = RecordingProvider()
+    loop = AgentLoop(loader=EngineLoader(workspace=tmp_path))
+    asyncio.run(
+        loop.process_direct(
+            "ordinary chat",
+            session_id="session",
+            include_skill_assembly=False,
+            provider_bundle=_bundle(non_task_provider),
+        )
+    )
+
+    team_node_provider = RecordingProvider()
+    asyncio.run(
+        loop.process_direct(
+            "team child",
+            session_id="session:team:child",
+            parent_session_id="session",
+            source="team:child",
+            task_id="task-1",
+            task_mode=True,
+            include_skill_assembly=False,
+            provider_bundle=_bundle(team_node_provider),
+        )
+    )
+
+    assert "run_agent_team" not in non_task_provider.tool_names_by_call[0]
+    assert "run_agent_team" not in team_node_provider.tool_names_by_call[0]