feat(engine): 添加MCP连接管理和工具集成功能

- 集成MCP连接管理器，支持MCP服务器连接
- 添加多种内置工具：ClarifyTool、CronTool、DelegateTool、ExecuteCodeTool、
  PatchFileTool、ProcessTool、SendMessageTool、SpawnTool、TerminalTool、
  TodoTool、WebFetchTool、WebSearchTool、WriteFileTool等
- 实现工具注册和装配功能
- 添加技能选择上下文参数
- 支持思考模式控制参数thinking_enabled

feat(coordinator): 重构任务执行计划器参数命名

- 将learning_candidate_enabled重命名为allow_candidate_generation
- 更新TeamGraphScheduler中的参数传递
- 修改LocalAgentRunner中的相关参数处理
- 更新README文档中的相应描述

refactor(context): 标准化工具调用参数格式

- 添加_json导入用于参数序列化
- 实现_provider_tool_calls方法标准化OpenAI兼容的工具调用载荷
- 修复工具调用中参数非字符串类型的序列化问题

refactor(session): 优化消息历史记录过滤逻辑

- 修改get_messages_as_conversation为基于运行状态过滤消息
- 排除未完成、失败或错误结束的运行记录
- 改进对话历史的可见性控制机制

fix(store): 修复FTS索引重建逻辑

- 添加异常处理防止FTS索引创建失败
- 实现_rebuild_fts_index方法重新构建全文搜索索引
- 优化索引触发器和表的维护流程

app-instance/backend/beaver/tools/mcp/wrapper.py

@@ -0,0 +1,88 @@
+"""MCP tool wrappers for Beaver's tool contract."""
+
+from __future__ import annotations
+
+import asyncio
+from dataclasses import dataclass
+import json
+from typing import Any, Awaitable, Callable
+
+from beaver.tools.base import BaseTool, ToolContext, ToolResult, ToolSpec
+
+
+def _tool_schema(tool_def: Any) -> dict[str, Any]:
+    schema = getattr(tool_def, "inputSchema", None) or getattr(tool_def, "input_schema", None)
+    if isinstance(schema, dict):
+        return schema
+    return {"type": "object", "properties": {}}
+
+
+def _tool_name(tool_def: Any) -> str:
+    return str(getattr(tool_def, "name", "") or "")
+
+
+def _tool_description(tool_def: Any) -> str:
+    return str(getattr(tool_def, "description", "") or _tool_name(tool_def))
+
+
+def _mcp_result_to_text(result: Any) -> str:
+    parts: list[str] = []
+    for block in list(getattr(result, "content", []) or []):
+        text = getattr(block, "text", None)
+        parts.append(str(text if text is not None else block))
+    if not parts and getattr(result, "structuredContent", None) is not None:
+        return json.dumps(getattr(result, "structuredContent"), ensure_ascii=False, indent=2)
+    return "\n".join(parts) or "(no output)"
+
+
+@dataclass(slots=True)
+class MCPToolWrapper(BaseTool):
+    server_id: str
+    tool_def: Any
+    call_tool: Callable[[str, dict[str, Any]], Awaitable[Any]]
+    tool_timeout: int = 30
+    sensitive: bool = False
+    kind: str = "online"
+    category: str = "online"
+    display_name: str = ""
+
+    @property
+    def original_name(self) -> str:
+        return _tool_name(self.tool_def)
+
+    @property
+    def spec(self) -> ToolSpec:
+        return ToolSpec(
+            name=f"mcp_{self.server_id}_{self.original_name}",
+            description=_tool_description(self.tool_def),
+            input_schema=_tool_schema(self.tool_def),
+            toolset=f"mcp-{self.server_id}",
+            metadata={
+                "server_id": self.server_id,
+                "original_tool_name": self.original_name,
+                "kind": self.kind,
+                "category": self.category,
+                "display_name": self.display_name or self.server_id,
+                "transport": "mcp",
+            },
+        )
+
+    async def invoke(self, arguments: dict[str, Any], context: ToolContext) -> ToolResult:
+        try:
+            result = await asyncio.wait_for(
+                self.call_tool(self.original_name, dict(arguments or {})),
+                timeout=max(1, int(self.tool_timeout or 30)),
+            )
+            return ToolResult(
+                success=True,
+                content=_mcp_result_to_text(result),
+                tool_name=self.spec.name,
+                raw_output=result,
+            )
+        except Exception as exc:
+            return ToolResult(
+                success=False,
+                content=f"MCP tool {self.server_id}.{self.original_name} failed: {exc}",
+                tool_name=self.spec.name,
+                error=str(exc),
+            )