feat(app-instance): 集成Beaver后端并更新配置管理

集成新的Beaver后端服务到应用实例中，替换原有的nanobot实现。

主要变更包括：
- 在Dockerfile和环境配置中添加Beaver相关路径和配置变量
- 更新工作目录结构从.nanobot到.beaver
- 实现Beaver引擎加载器，支持配置文件加载和工具组装
- 添加内置工具如ListDirectoryTool、ReadFileTool、SearchFilesTool
- 更新消息处理流程，支持通道适配器和网关模式
- 重构技能系统，支持显式工具提示和嵌入式检索
- 改进错误处理和生命周期管理

此变更使应用实例能够使用统一的Beaver后端进行AI代理运行时管理。

app-instance/backend/beaver/engine/loop.py

@@ -272,12 +272,24 @@ class AgentLoop:
         memory_service = self._require_loaded("memory_service")
         context_builder = self._require_loaded("context_builder")
         tool_registry = self._require_loaded("tool_registry")
+        tool_assembler = self._require_loaded("tool_assembler")
         tool_executor = self._require_loaded("tool_executor")
+        skills_loader = self._require_loaded("skills_loader")
         skill_assembler = self._require_loaded("skill_assembler")
 
+        config = loaded.config
+        configured_provider = config.resolve_provider_target(model=model, provider_name=provider_name)
+
         resolved_session_id = session_id or uuid4().hex
         resolved_run_id = uuid4().hex
-        resolved_model = model or self.profile.default_model
+        resolved_model = configured_provider.get("model") or self.profile.default_model
+        resolved_provider_name = configured_provider.get("provider_name") or provider_name
+        resolved_api_key = api_key or configured_provider.get("api_key")
+        resolved_api_base = api_base or configured_provider.get("api_base")
+        resolved_extra_headers = extra_headers or configured_provider.get("extra_headers")
+        resolved_request_timeout_seconds = configured_provider.get("request_timeout_seconds")
+        resolved_embedding_model = embedding_model or config.default_embedding_model
+        resolved_embedding_target = embedding_target or config.resolve_embedding_target()
         resolved_max_tokens = max_tokens or self.profile.max_tokens
         resolved_temperature = self.profile.temperature if temperature is None else temperature
         resolved_max_tool_iterations = (
@@ -316,20 +328,21 @@ class AgentLoop:
         user_message_recorded = False
         iterations = 0
         final_usage: dict[str, Any] = {}
-        final_provider_name: str | None = provider_name
+        final_provider_name: str | None = resolved_provider_name
         final_model: str | None = resolved_model
         try:
             bundle = provider_bundle or make_provider_bundle(
                 model=resolved_model,
-                provider_name=provider_name,
-                api_key=api_key,
-                api_base=api_base,
-                extra_headers=extra_headers,
+                provider_name=resolved_provider_name,
+                api_key=resolved_api_key,
+                api_base=resolved_api_base,
+                request_timeout_seconds=resolved_request_timeout_seconds,
+                extra_headers=resolved_extra_headers,
                 routing=routing,
                 fallback_target=fallback_target,
                 auxiliary_target=auxiliary_target,
-                embedding_target=embedding_target,
-                embedding_model=embedding_model or "text-embedding-v4",
+                embedding_target=resolved_embedding_target,
+                embedding_model=resolved_embedding_model,
             )
             skill_selector_provider = bundle.auxiliary_provider or bundle.main_provider
             skill_selector_model = (
@@ -364,6 +377,32 @@ class AgentLoop:
                     user_id=user_id,
                 )
 
+            selected_tool_specs = await tool_assembler.assemble(
+                task_description=task,
+                registry=tool_registry,
+                skills_loader=skills_loader,
+                activated_skills=assembled_skills.activated_skills,
+                embedding_runtime=bundle.embedding_runtime,
+                top_k=10,
+            )
+            tool_schemas = tool_registry.export_selected_provider_schemas(selected_tool_specs)
+            session_manager.append_message(
+                resolved_session_id,
+                run_id=resolved_run_id,
+                role="system",
+                event_type="tool_selection_snapshotted",
+                event_payload={
+                    "tools": [spec.to_mcp_descriptor() for spec in selected_tool_specs],
+                    "tool_names": [spec.name for spec in selected_tool_specs],
+                },
+                content=", ".join(spec.name for spec in selected_tool_specs) or None,
+                context_visible=False,
+                source=source,
+                title=title,
+                model=resolved_model,
+                user_id=user_id,
+            )
+
             build_input = ContextBuildInput(
                 base_system_prompt=self.profile.system_prompt,
                 history=session_manager.get_history(resolved_session_id),
@@ -412,7 +451,6 @@ class AgentLoop:
 
             provider = bundle.main_provider
             messages = list(context_result.messages)
-            tool_schemas = tool_registry.export_provider_schemas()
             tool_context = ToolContext(
                 workspace=str(loaded.workspace),
                 session_id=resolved_session_id,