chore(sbot): sync release notes for v0.2.7

Author: linyuan.yang

packages/sbot/package.json

@@ -2,8 +2,8 @@
   "name": "@qingfeng346/sbot",
   "version": "0.2.7",
   "description": "Self-hosted AI agent server. Open-source alternative to OpenAI, Claude, and Gemini interfaces. Run LLM agents locally with MCP tool support, memory, multi-channel integrations (Lark/Feishu), and a built-in web UI.",
-  "releasenoteEn": "This release includes the following main changes:\n\n### Added\n\n1. Token Usage now provides full analysis: filter by time, agent, model, provider, and channel, then review trends, breakdowns, and individual calls.\n2. Model and embedding settings now show their references from agents, memory profiles, agenda profiles, notes, and wikis before you change a configuration.\n3. The unified `move` file-system tool can move, rename, or copy files and directories.\n\n### Improved\n\n1. OpenAI-compatible model providers now preserve the server's detailed error message when a request fails, making model setup issues easier to diagnose.\n\n### Fixed\n\n1. Fixed editing an agent unintentionally clearing extended settings that are not shown in the form.",
-  "releasenoteZh": "本次主要更新如下：\n\n### 新增\n\n1. Token 用量页新增完整分析能力：可按时间、智能体、模型、供应商和渠道筛选，并查看趋势、分布和每次调用明细。\n2. 模型与 Embedding 配置会显示被哪些智能体、记忆、日程、Note 和 Wiki 引用，方便调整配置前确认影响范围。\n3. 文件系统新增统一的 `move` 工具，可用于移动、重命名或复制文件和目录。\n\n### 优化\n\n1. OpenAI 兼容模型接口在请求失败时会保留服务端返回的具体错误信息，排查模型配置更方便。\n\n### 修复\n\n1. 修复编辑智能体配置时，表单未展示的扩展配置可能被意外清空的问题。",
+  "releasenoteEn": "This release includes the following main changes:\n\n### Added\n\n1. Token Usage now provides full analysis: filter by time, agent, model, provider, and channel, then review trends, breakdowns, and individual calls.\n2. Model and embedding settings now show their references from agents, memory profiles, agenda profiles, notes, and wikis before you change a configuration.\n3. The unified `move` file-system tool can move, rename, or copy files and directories.\n4. Subtask thinking can be opened while the subtask is still running, and refreshes live instead of only after it finishes.\n5. Settings now has a Stop Service action: it stops accepting new tasks and exits once active agent, memory, and agenda work has completed.\n6. The VS Code extension adds \"Open Chat in Editor\" so the chat can live in an editor tab; the sidebar entry is unchanged.\n\n### Improved\n\n1. OpenAI-compatible model providers now preserve the server's detailed error message when a request fails, making model setup issues easier to diagnose.\n2. Tool-call one-line summaries now pick the arguments that actually distinguish a call, so repeated calls to the same tool are easier to tell apart.\n3. Expanding tool-call details no longer shifts the block's width, and now animates.\n\n### Fixed\n\n1. Fixed editing an agent unintentionally clearing extended settings that are not shown in the form.",
+  "releasenoteZh": "本次主要更新如下：\n\n### 新增\n\n1. Token 用量页新增完整分析能力：可按时间、智能体、模型、供应商和渠道筛选，并查看趋势、分布和每次调用明细。\n2. 模型与 Embedding 配置会显示被哪些智能体、记忆、日程、Note 和 Wiki 引用，方便调整配置前确认影响范围。\n3. 文件系统新增统一的 `move` 工具，可用于移动、重命名或复制文件和目录。\n4. 子任务执行过程中就可以打开思考过程查看，内容会实时刷新，不必等子任务结束。\n5. 管理后台设置页新增「停止服务」，停止接收新任务，等正在执行的智能体、记忆和事项任务完成后退出。\n6. VS Code 插件新增「在编辑器中打开聊天」，可以把聊天开在编辑器标签页里，侧边栏入口保持不变。\n\n### 优化\n\n1. OpenAI 兼容模型接口在请求失败时会保留服务端返回的具体错误信息，排查模型配置更方便。\n2. 工具调用的一行摘要会挑选有区分度的参数展示，同一屏里多次调用同一个工具也能区分开。\n3. 工具调用详情展开时不再发生宽度跳变，并带有展开动画。\n\n### 修复\n\n1. 修复编辑智能体配置时，表单未展示的扩展配置可能被意外清空的问题。",
   "license": "MIT",
   "repository": {
     "type": "git",