diff --git a/api/app/models/file_metadata_model.py b/api/app/models/file_metadata_model.py
index baf9bd97..28e87367 100644
--- a/api/app/models/file_metadata_model.py
+++ b/api/app/models/file_metadata_model.py
@@ -35,7 +35,7 @@ class FileMetadata(Base):
 
     id = Column(UUID(as_uuid=True), primary_key=True, default=uuid.uuid4, index=True)
     tenant_id = Column(UUID(as_uuid=True), nullable=False, index=True, comment="Tenant ID")
-    workspace_id = Column(UUID(as_uuid=True), nullable=False, index=True, comment="Workspace ID")
+    workspace_id = Column(UUID(as_uuid=True), nullable=True, index=True, comment="Workspace ID")
     file_key = Column(String(512), nullable=False, unique=True, index=True, comment="Storage file key")
     file_name = Column(String(255), nullable=False, comment="Original file name")
     file_ext = Column(String(32), nullable=False, comment="File extension")
diff --git a/api/app/services/file_storage_service.py b/api/app/services/file_storage_service.py
index 672e1cff..bb9f1894 100644
--- a/api/app/services/file_storage_service.py
+++ b/api/app/services/file_storage_service.py
@@ -26,7 +26,7 @@ logger = get_business_logger()
 
 def generate_file_key(
     tenant_id: uuid.UUID,
-    workspace_id: uuid.UUID,
+    workspace_id: uuid.UUID | None,
     file_id: uuid.UUID,
     file_ext: str,
 ) -> str:
@@ -56,8 +56,9 @@ def generate_file_key(
     # Ensure file_ext starts with a dot
     if file_ext and not file_ext.startswith('.'):
         file_ext = f'.{file_ext}'
-    
-    return f"{tenant_id}/{workspace_id}/{file_id}{file_ext}"
+    if workspace_id:
+        return f"{tenant_id}/{workspace_id}/{file_id}{file_ext}"
+    return f"{tenant_id}/{file_id}{file_ext}"
 
 
 class FileStorageService:
@@ -96,7 +97,7 @@ class FileStorageService:
     async def upload_file(
         self,
         tenant_id: uuid.UUID,
-        workspace_id: uuid.UUID,
+        workspace_id: uuid.UUID | None,
         file_id: uuid.UUID,
         file_ext: str,
         content: bytes,
diff --git a/api/app/version_info.json b/api/app/version_info.json
index 991369d7..aea03dcd 100644
--- a/api/app/version_info.json
+++ b/api/app/version_info.json
@@ -1,4 +1,38 @@
 {
+  "v0.2.4": {
+    "introduction": {
+      "codeName": "智远",
+      "releaseDate": "2026-2-11",
+      "upgradePosition": "🐻 生产级稳健性升级版本，智慧致远，从容应对复杂场景",
+      "coreUpgrades": [
+        "1. Skills 技能框架 🛠️<br>* Skills 支持：引入全新的Skills技能系统，支持可扩展的能力模块，可在Agent和工作流中动态加载与编排",
+        "2. 多模态与交互 💬<br>* 文件多模态支持：全面支持消息输入、LLM处理和输出渲染中的多模态文件处理，实现更丰富的媒体感知对话<br>* 语音交互：语音交互功能正在积极开发中，为免提对话体验奠定基础（开发中）",
+        "3. 知识库集成 📚<br>* 飞书知识库：无缝对接飞书文档库，支持企业知识检索<br>* 语雀知识库：原生连接语雀文档平台，扩展对国内企业工具生态的覆盖<br>* Web站点知识库：通用Web站点抓取与索引，支持从公开网页内容构建知识库<br>* 视觉模型选择优化：知识库视觉模型配置现已支持LLM和Chat两种模型类型，移除了此前仅限Chat类型的限制",
+        "4. 记忆智能 🧠<br>* 本体工程（二期）：基于本体工程的高级记忆场景分类与萃取，实现结构化、领域感知的记忆组织，提升分类准确性<br>* 默认模型配置：情绪分析、反思和记忆萃取模块现默认使用空间级模型，确保开箱即用的一致性行为<br>* 智能模型回退：当已配置的情绪或反思模型为空或不可用时，系统自动回退至空间默认模型，避免静默失败<br>* 记忆模型回退兜底：当记忆中配置的模型为空或不可用时，系统优雅降级至空间默认模型",
+        "5. 性能与扩展 ⚡<br>* 模型并发（model_api_keys）：支持并发模型API Key管理，实现并行模型调用，提升高负载场景下的吞吐能力",
+        "6. 稳健性与缺陷修复 🔧<br>* 记忆配置版本固定：修复用户记忆配置未跟随应用版本发布固定的问题，消除跨部署的行为不一致<br>* 空间默认记忆保护：空间级默认记忆配置现不可删除；用户级配置仍可删除<br>* Agent与工作流配置兜底：解决Agent和工作流节点中记忆配置可能为空、或已选择但未配置的边界情况——全面的回退处理现可防止运行时错误<br>* 隐形记忆字段重命名：将隐形记忆接口JSON响应中的user_id修正为end_user_id，与规范数据模型对齐<br>* 记忆配置ID迁移：将Agent和工作流记忆配置中的memory_content重命名为memory_config_id，保持API一致性<br>* Worker-Memory告警解决：解决worker-memory服务中的告警级别问题，提升运维监控清晰度<br>* 双语接口修复：修复记忆相关API接口的中英文不一致问题<br>* 新用户记忆配置自动回填：新创建的EndUser若memory_config_id为None，系统自动从最新Release获取memory_config_id并回填<br>* 存量用户记忆配置自动回填：已有EndUser若memory_config_id为None，系统同样从最新Release获取并回填，确保向后兼容，无需手动迁移",
+        "<br>",
+        "Memory Bear v0.2.4 向生产级稳健性迈进，Skills框架与多模态支持开启认知平台新篇章。",
+        "记忆熊，智慧致远，从容应对真实世界的多样性。🐻✨"
+      ]
+    },
+    "introduction_en": {
+      "codeName": "ZhiYuan",
+      "releaseDate": "2026-2-11",
+      "upgradePosition": "🐻 Production-grade resilience release — Wisdom Reaching Far, gracefully handling complex scenarios",
+      "coreUpgrades": [
+        "1. Skills Framework 🛠️<br>* Skills Support: Introduced a new Skills system, enabling extensible capability modules that can be dynamically loaded and orchestrated within agents and workflows",
+        "2. Multimodal & Interaction 💬<br>* File Multimodal Support: Full multimodal file handling across message input, LLM processing, and output rendering — supporting richer, media-aware conversations<br>* Voice Interaction: Voice-based interaction capabilities are under active development, laying the groundwork for hands-free conversational experiences (In Progress)",
+        "3. Knowledge Base Integration 📚<br>* Feishu Knowledge Base: Seamless integration with Feishu (Lark) document repositories for enterprise knowledge retrieval<br>* Yuque Knowledge Base: Native connector for Yuque documentation platforms, expanding coverage of Chinese enterprise tooling<br>* Web Site Knowledge Base: General-purpose web site crawling and indexing for knowledge base construction from public web content<br>* Visual Model Selection: Knowledge base visual model configuration now supports both LLM and Chat model types, removing the previous restriction to Chat-only selection",
+        "4. Memory Intelligence 🧠<br>* Ontology Engineering (Phase 2): Advanced memory scene classification and extraction powered by ontology engineering — enabling structured, domain-aware memory organization with improved categorization accuracy<br>* Default Model Configuration: Emotion analysis, reflection, and memory extraction modules now default to the space-level model, ensuring consistent behavior out of the box<br>* Intelligent Model Fallback: If configured emotion or reflection models are empty or unavailable, the system automatically falls back to the space default model — preventing silent failures<br>* Memory Config Fallback for Models: When any memory-configured model is empty or unavailable, the system gracefully degrades to the space default model",
+        "5. Performance & Scalability ⚡<br>* Model Concurrency (model_api_keys): Support for concurrent model API key management, enabling parallel model invocations and improved throughput for high-load scenarios",
+        "6. Robustness & Bug Fixes 🔧<br>* Memory Config Version Pinning: Fixed an issue where user memory configurations were not pinned to application release versions, causing inconsistent behavior across deployments<br>* Space Default Memory Protection: Space-level default memory configurations are now protected from deletion; user-level configurations remain deletable<br>* Agent & Workflow Config Fallback: Resolved edge cases in Agent and Workflow nodes where memory config could be empty or selected but unconfigured — comprehensive fallback handling now prevents runtime errors<br>* Implicit Memory Field Rename: Corrected user_id to end_user_id in JSON responses from implicit memory interfaces, aligning with the canonical data model<br>* Memory Config ID Migration: Renamed memory_content to memory_config_id in Agent and Workflow memory configurations for API consistency<br>* Worker-Memory Alerts: Resolved warning-level alerts in the worker-memory service, improving operational monitoring clarity<br>* Bilingual Interface Fixes: Fixed Chinese/English language inconsistencies across memory-related API interfaces<br>* EndUser Memory Config Auto-Backfill (New Users): When a newly created EndUser has memory_config_id as None, the system automatically fetches the latest release's memory_config_id and backfills it<br>* EndUser Memory Config Auto-Backfill (Existing Users): For existing EndUsers with memory_config_id as None, the system similarly retrieves and backfills from the latest release — ensuring backward compatibility without manual migration",
+        "<br>",
+        "Memory Bear v0.2.4 advances toward production-grade resilience, with the Skills framework and multimodal support opening a new chapter for the cognitive platform.",
+        "MemoryBear — Wisdom Reaching Far, gracefully handling real-world variability. 🐻✨"
+      ]
+    }
+  },
   "v0.2.3": {
     "introduction": {
       "codeName": "归墟",