feat: add 灵犀 Feishu bot + fix agent schedule system + default all tools

- Add 灵犀学习助手 Feishu bot (lingxi_app_service + lingxi_ws_handler)
- Fix agent_schedule_service missing AgentSchedule import (Celery Beat)
- Fix scene_templates default enable_tools=False → True
- Fix workflow_engine LLM node: empty tools list now = all tools (consistent with agent node)
- Add 创建agent.md guide document

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

backend/app/services/workflow_engine.py

@@ -1825,23 +1825,20 @@ class WorkflowEngine:
                 
                 # 如果启用了工具，加载工具定义
                 tools = []
-                if enable_tools and tools_config:
+                if enable_tools:
                     from app.services.tool_registry import tool_registry
-                    # 从注册表加载工具定义
-                    tools = tool_registry.get_tools_by_names(tools_config)
-                    logger.info(f"[rjb] LLM节点启用工具调用: {len(tools)} 个工具, 工具列表: {tools_config}")
+                    if tools_config:
+                        tools = tool_registry.get_tools_by_names(tools_config)
+                    else:
+                        # 空列表 = 全部工具（与 Agent 节点行为一致）
+                        tools = tool_registry.get_all_tool_schemas()
+                    logger.info(f"[rjb] LLM节点启用工具调用: {len(tools)} 个工具, 工具列表: {tools_config or '全部'}")
                     if not tools:
                         logger.warning(
                             "[rjb] LLM 已 enable_tools 但当前进程 tool_registry 中 0 个匹配 schema，"
                             "将无法发起 function calling（常见于 Celery Worker 未加载 tools_bootstrap）。配置=%s",
                             tools_config,
                         )
-                    elif len(tools) < len(tools_config):
-                        missing = [n for n in tools_config if not tool_registry.get_tool_schema(n)]
-                        logger.warning(
-                            "[rjb] LLM 工具部分缺失 schema，缺失=%s（可动手能力不完整）",
-                            missing,
-                        )
                 
                 # 调用LLM服务
                 try: