<!-- qwen-release-notes:v1 -->
## 亮点
- 多工作区支持现已覆盖 ACP 传输、守护进程工作进程、分屏会话以及工作区感知型操作。([#6621](https://github.com/QwenLM/qwen-code/pull/6621), [#6635](https://github.com/QwenLM/qwen-code/pull/6635), [#6746](https://github.com/QwenLM/qwen-code/pull/6746), [#6724](https://github.com/QwenLM/qwen-code/pull/6724))
- Web Shell 新增了工件面板、现代化的侧边栏与设置体验，以及可编辑的用户级模型配置。([#6591](https://github.com/QwenLM/qwen-code/pull/6591), [#6804](https://github.com/QwenLM/qwen-code/pull/6804), [#6817](https://github.com/QwenLM/qwen-code/pull/6817), [#6768](https://github.com/QwenLM/qwen-code/pull/6768))
- 会话与模型恢复机制对畸形流、协议轮次泄漏、OAuth 过期及守护进程重启的容错能力更强。([#6794](https://github.com/QwenLM/qwen-code/pull/6794), [#6683](https://github.com/QwenLM/qwen-code/pull/6683), [#6732](https://github.com/QwenLM/qwen-code/pull/6732), [#6680](https://github.com/QwenLM/qwen-code/pull/6680))
- 子代理的追踪更加便捷，提供更丰富的实时命令上下文、转录路径、时间线以及继承的 Explore 模型。([#6580](https://github.com/QwenLM/qwen-code/pull/6580), [#6772](https://github.com/QwenLM/qwen-code/pull/6772), [#6807](https://github.com/QwenLM/qwen-code/pull/6807))
- CLI 工作流新增项目级提示词暂存、可见的 Git 分支、可配置的 Shell 超时设置以及更安全的内存压力处理机制。([#6709](https://github.com/QwenLM/qwen-code/pull/6709), [#6725](https://github.com/QwenLM/qwen-code/pull/6725), [#6628](https://github.com/QwenLM/qwen-code/pull/6628), [#6682](https://github.com/QwenLM/qwen-code/pull/6682))
- SDK 与守护进程集成现可控制推理力度、模型、用量、上下文、传输选项及运行时通道。([#6492](https://github.com/QwenLM/qwen-code/pull/6492), [#6491](https://github.com/QwenLM/qwen-code/pull/6491), [#6741](https://github.com/QwenLM/qwen-code/pull/6741))
## 破坏性变更
暂无已知的破坏性变更。
## 完整变更列表
### 新功能
- 托管于 SDK 的代理现可将 `ask_user_question` 工具审批的用户回答转发回模型，从而在 TypeScript 和 Python SDK 集成中启用交互式问答流程。([#6655](https://github.com/QwenLM/qwen-code/pull/6655)) by @TianYuan1024
- 钉钉频道现可在机器人回复中可选地使用可见的 @ 符号提及群消息发送者，且对未提及的回复保留 Markdown 渲染。([#6679](https://github.com/QwenLM/qwen-code/pull/6679)) by @qqqys
- Web Shell 现包含右侧审查面板，用于检查编辑过的文件、工件和定时任务，支持差异对比、文件导航及会话级标签页。([#6591](https://github.com/QwenLM/qwen-code/pull/6591)) by @ytahdn
- 守护进程现支持带工作区限定的 ACP 端点，允许 SDK 和 Web Shell 客户端在多工作区部署中打开限定于特定工作区的会话。([#6621](https://github.com/QwenLM/qwen-code/pull/6621)) by @doudouOUC
- 新增提示词编辑器头部渲染插槽及可选的移动端欢迎页底部布局，以便嵌入方在小屏幕上于编辑器上方显示自定义内容。([#6584](https://github.com/QwenLM/qwen-code/pull/6584)) by @yuanyuanAli
- 新增 SDK 方法以控制推理力度、列出可用模型，并在运行时查询账户与上下文用量统计。([#6492](https://github.com/QwenLM/qwen-code/pull/6492)) by @juhuan
- 流式传输期间使用 Alt+T 展开思考块时，现会实时显示完整的推理内容，而非截断的预览。([#6678](https://github.com/QwenLM/qwen-code/pull/6678)) by @huww98
- 定时任务提示词现支持通过浮动选择器内联扩展、技能及 MCP 引用标签，这些标签可干净地序列化到保存的提示词文本中。([#6589](https://github.com/QwenLM/qwen-code/pull/6589)) by @BZ-D
- 在非空提示词下按 Ctrl+S 现会将其暂存至当前项目，并在下次启动时恢复，提交后清除暂存。([#6709](https://github.com/QwenLM/qwen-code/pull/6709)) by @destire-mio
- 会话恢复现在所有入口点采用统一计划，一致地处理中断的提示词、悬空的工具调用及受损的历史记录。([#6731](https://github.com/QwenLM/qwen-code/pull/6731)) by @water-in-stone
- 子代理视图现显示未截断的实时命令、更深层的进度历史、执行转录路径以及内联审批时的近期工具上下文。([#6580](https://github.com/QwenLM/qwen-code/pull/6580)) by @TianYuan1024
- `qwen serve` 守护进程现可为多个可信工作区托管频道工作进程，并将每个工作进程绑定至其对应的工作区目录。([#6635](https://github.com/QwenLM/qwen-code/pull/6635)) by @doudouOUC
- 新增 `MessageDisplay` 钩子，在助手流式传输期间增量触发，允许观察者在回合完成前跟踪回复文本。([#6489](https://github.com/QwenLM/qwen-code/pull/6489)) by @yanchenko
- 双击渲染后的 Markdown 表格中的单元格，现会打开一个包含完整单元格值的对话框，便于阅读和复制。([#6530](https://github.com/QwenLM/qwen-code/pull/6530)) by @jifeng
- 守护进程现为非可信工作区提供只读会话目录，允许客户端在无权写入的情况下浏览会话。([#6717](https://github.com/QwenLM/qwen-code/pull/6717)) by @doudouOUC
- 运行时添加的工作区注册现已持久化，使其在守护进程重启后无需手动重新配置即可保留。([#6716](https://github.com/QwenLM/qwen-code/pull/6716)) by @doudouOUC
- Web Shell 中的用户消息现以内联方式渲染扩展、技能及 MCP 引用标签，与提示词编辑器预览保持一致。([#6537](https://github.com/QwenLM/qwen-code/pull/6537)) by @ever-o
- 守护进程重启时，通道工作进程现在会自动重新连接，无需人工干预即可恢复钉钉和飞书会话。 ([#6680](https://github.com/QwenLM/qwen-code/pull/6680)) by @qqqys
- Composer 工具栏现在会将当前 Git 分支显示为标签（chip），让你随时清楚正在处理哪个分支。 ([#6725](https://github.com/QwenLM/qwen-code/pull/6725)) by @han-dreamer
- 前台 Shell 命令现在遵循可配置默认超时设置，防止长时间运行的命令无限期阻塞。 ([#6628](https://github.com/QwenLM/qwen-code/pull/6628)) by @Nas01010101
- TypeScript 和 Python SDK 现在均暴露传输选择与查询选项，以便更精细地控制与守护进程的通信。 ([#6491](https://github.com/QwenLM/qwen-code/pull/6491)) by @juhuan
- Web Shell 会话侧边栏现在支持配置，允许嵌入方选择显示哪些会话元数据和操作。 ([#6750](https://github.com/QwenLM/qwen-code/pull/6750)) by @dreamWB
- 工作区转录阅读器现在按工作区提供持久化的对话历史，支持服务端会话浏览。 ([#6740](https://github.com/QwenLM/qwen-code/pull/6740)) by @doudouOUC
- Web Shell 中的会话组现在支持自定义十六进制颜色代码，便于更直观的视觉分类。 ([#6752](https://github.com/QwenLM/qwen-code/pull/6752)) by @samuelhsin
- Composer 模型选择器已重新设计，提供更清晰的模型元数据并实现更快的切换。 ([#6758](https://github.com/QwenLM/qwen-code/pull/6758)) by @dreamWB
- 代码审查现在包含流程正确性检查、可调节的审查力度，以及发布前验证结果的护栏机制。 ([#6711](https://github.com/QwenLM/qwen-code/pull/6711)) by @wenshao
- 多工作区分屏视图现在支持跨工作区会话、工作区标签，以及用于并排编辑的响应式布局。 ([#6746](https://github.com/QwenLM/qwen-code/pull/6746)) by @wenshao
- 所有工作区的定时任务现在已聚合到 Web Shell 的单一视图中，便于统一监控。 ([#6759](https://github.com/QwenLM/qwen-code/pull/6759)) by @wenshao
- Composer 现在支持自定义占位符文本，允许嵌入方根据上下文向用户展示特定提示。 ([#6765](https://github.com/QwenLM/qwen-code/pull/6765)) by @dreamWB
- 发布说明现在借助 AI 辅助起草，将完整提交范围的变更总结为分类清晰、面向用户的要点列表。 ([#6756](https://github.com/QwenLM/qwen-code/pull/6756)) by @yiliang114
- Web Shell 现在内置基于 shadcn 的 UI 基础架构，提供一致的设计令牌和可访问的组件基元。 ([#6760](https://github.com/QwenLM/qwen-code/pull/6760)) by @ytahdn
- 子代理活动现在以时间顺序转录形式展示，并附带并行代理时间线，便于多任务检查。 ([#6772](https://github.com/QwenLM/qwen-code/pull/6772)) by @wenshao
- CLI 现在支持在运行时启动、停止和列出守护进程通道工作进程，无需重启守护进程。 ([#6741](https://github.com/QwenLM/qwen-code/pull/6741)) by @doudouOUC
- 持久化对话转录记录现在设置了上限，以防止磁盘空间无限增长。 ([#6769](https://github.com/QwenLM/qwen-code/pull/6769)) by @doudouOUC
- 创建新会话时，Web Shell 会触发回调，从而支持自定义集成。 ([#6703](https://github.com/QwenLM/qwen-code/pull/6703)) by @ytahdn
- 技能安装路径现在已公开，用户可查看技能的存储位置。 ([#6811](https://github.com/QwenLM/qwen-code/pull/6811)) by @callmeYe
- 多工作区侧边栏已重新设计，采用更简洁、现代的布局。 ([#6804](https://github.com/QwenLM/qwen-code/pull/6804)) by @ytahdn
- 设置页面已使用现代组件库进行刷新，呈现更精致的外观与交互体验。 ([#6817](https://github.com/QwenLM/qwen-code/pull/6817)) by @ytahdn
- 用户级设置现在可直接在面板中编辑，并支持面板内模型管理。 ([#6768](https://github.com/QwenLM/qwen-code/pull/6768)) by @wenshao
- 回退（Rewind）和 Shell 操作现在支持跨多工作区运行。 ([#6826](https://github.com/QwenLM/qwen-code/pull/6826)) by @doudouOUC
- 工作区现在支持在运行时直接移除，无需重启服务器。 ([#6745](https://github.com/QwenLM/qwen-code/pull/6745)) by @doudouOUC
- 代码审查现在可捕获未跟踪文件，解析代码片段中的锚点，并对代码发布进行拦截控制。 ([#6771](https://github.com/QwenLM/qwen-code/pull/6771)) by @wenshao
- PR 分类评论现在包含置信度评分、时序图、文件概览以及审查页脚。 ([#6789](https://github.com/QwenLM/qwen-code/pull/6789)) by @wenshao
- 探索子代理现在默认继承主模型，而非使用固定的较小模型。 ([#6807](https://github.com/QwenLM/qwen-code/pull/6807)) by @tanzhenxin
- 守护进程状态现在报告模型 API 错误计数与重试指标。 ([#6837](https://github.com/QwenLM/qwen-code/pull/6837)) by @wenshao
- 新增 API 端点支持以编程方式切换工作区技能。 ([#6816](https://github.com/QwenLM/qwen-code/pull/6816)) by @callmeYe
- 新增 xAI Grok 作为内置模型提供商预设。 ([#6805](https://github.com/QwenLM/qwen-code/pull/6805)) by @TianYuan1024
- 扩展管理已升级至 v2 版本，优化了安装、更新和移除流程。 ([#6825](https://github.com/QwenLM/qwen-code/pull/6825)) by @doudouOUC
- 语音输入现在按工作区进行限定，以提供更好的多工作区支持。 ([#6839](https://github.com/QwenLM/qwen-code/pull/6839)) by @doudouOUC
- 会话导出现在按工作区进行限定，确保跨工作区的导出内容保持条理清晰。 ([#6844](https://github.com/QwenLM/qwen-code/pull/6844)) by @doudouOUC
- Web Shell 中的 Markdown 表格现在显示行、列和单元格的选择统计信息。 ([#6838](https://github.com/QwenLM/qwen-code/pull/6838)) by @jifeng
### 错误修复
- YOLO 审批模式在模型进入计划模式时不再被静默禁用。([#6630](https://github.com/QwenLM/qwen-code/pull/6630)) by @Nas01010101
- 审批模式的 UI 标签和提示现已本地化，防止在非英文界面下使用 Shift+Tab 切换模式时出现语言混杂的输出。([#6592](https://github.com/QwenLM/qwen-code/pull/6592)) by @han-dreamer
- 渠道投递现在会过滤掉嵌套子代理的输出，防止在根代理给出最终答案前将中间研究报告发送到消息平台。([#6696](https://github.com/QwenLM/qwen-code/pull/6696)) by @qqqys
- 将预构建 npm 包的大小限制从 80 MB 提升至 96 MB，以适应自然增长并解除 Docker 沙箱构建的阻塞。([#6691](https://github.com/QwenLM/qwen-code/pull/6691)) by @qwen-code-dev-bot
- 修复了协议标签重试测试的 Docker 沙箱网络问题，使 CLI 在容器沙箱模式下可通过 host.docker.internal 访问测试服务器。([#6692](https://github.com/QwenLM/qwen-code/pull/6692)) by @qwen-code-dev-bot
- 修复了在 VS Code 中将光标设置到行首时被错误覆盖的问题。([#2971](https://github.com/QwenLM/qwen-code/pull/2971)) by @chinesepowered
- 通过在无工具调用时仍定期执行内存压力检查，修复了长交互会话退出时的内存溢出崩溃问题。([#6682](https://github.com/QwenLM/qwen-code/pull/6682)) by @yiliang114
- 通过移除部分屏幕外元素的负坐标边界属性，减少了移动端 UI 转储的 Token 消耗。([#6624](https://github.com/QwenLM/qwen-code/pull/6624)) by @chinesepowered
- 通过在最大 Token 升级和恢复过程中丢弃并重试无效的助手回合，修复了响应中协议标签泄露的问题。([#6683](https://github.com/QwenLM/qwen-code/pull/6683)) by @yiliang114
- 更新测试预期以匹配新的自动化 PR 审查工作流超时值，解决持续集成失败问题。([#6720](https://github.com/QwenLM/qwen-code/pull/6720)) by @yiliang114
- 当非只读工具被阻止时返回明确的错误响应，防止模型尝试变通方案，从而提升计划模式的可靠性。([#6667](https://github.com/QwenLM/qwen-code/pull/6667)) by @Alex-ai-future
- 在 /mcp 服务器详情视图中添加了“批准”按钮，允许用户批准待处理或之前被拒绝的工作区服务器。([#6518](https://github.com/QwenLM/qwen-code/pull/6518)) by @LaZzyMan
- 在规范化过程中保留以斜杠结尾的 URL 查询参数和片段，修复了远程 MCP 连接失败的问题。([#6587](https://github.com/QwenLM/qwen-code/pull/6587)) by @VectorPeak
- 将模板占位符限制为有效标识符，将 ${0} 等数字模式视为纯文本，修复了子代理启动失败的问题。([#6672](https://github.com/QwenLM/qwen-code/pull/6672)) by @ksws00684315
- 以 ACP 代理运行时，用户输入的指令现在会放置在附加文件内容之后，以便模型优先处理实际请求。([#6607](https://github.com/QwenLM/qwen-code/pull/6607)) by @kaymeer
- ACP 的 readTextFile 回退机制现在会拒绝小数形式的行号和限制值，为格式错误的读取窗口返回一致的无效参数错误。([#6704](https://github.com/QwenLM/qwen-code/pull/6704)) by @VectorPeak
- “添加工作区”对话框现在支持亮色和暗色主题，多工作区会话列表恢复了完整的行功能和单会话操作。([#6705](https://github.com/QwenLM/qwen-code/pull/6705)) by @wenshao
- 钉钉 Stream 连接在 Socket 打开时即被视为就绪，防止因缺少注册帧而断开健康会话。([#6715](https://github.com/QwenLM/qwen-code/pull/6715)) by @qqqys
- 受信任的非主工作区现在可以通过新的工作区范围组织端点更新会话置顶、分组和颜色状态。([#6724](https://github.com/QwenLM/qwen-code/pull/6724)) by @doudouOUC
- Claude Opus 4.6、4.7 和 4.8 模型现在默认使用 1M 上下文窗口和 128K 输出限制，防止过早压缩。([#6718](https://github.com/QwenLM/qwen-code/pull/6718)) by @yiliang114
- 剪贴板图片粘贴功能现在在独立桌面构建中能正常工作。([#6708](https://github.com/QwenLM/qwen-code/pull/6708)) by @zjunothing
- 使用 remember 命令后，已保存的记忆指令现在会立即刷新。([#6497](https://github.com/QwenLM/qwen-code/pull/6497)) by @han-dreamer
- 目标评估现在具备生命周期安全性，防止在会话清理期间出现错误。([#6681](https://github.com/QwenLM/qwen-code/pull/6681)) by @qqqys
- 托管记忆条目现在在微压缩期间会被保留，而不是被丢弃。([#6714](https://github.com/QwenLM/qwen-code/pull/6714)) by @yiliang114
- Claude 输出 Token 限制现在使用十进制数值，以实现更精确的上下文预算分配。([#6735](https://github.com/QwenLM/qwen-code/pull/6735)) by @yiliang114
- 模型和审批模式的更改现在能正确应用于 Web Shell 中的非主工作区会话。([#6737](https://github.com/QwenLM/qwen-code/pull/6737)) by @wenshao
- 接收到重复的无效模型流时，代理不再崩溃，而是能够优雅恢复。([#6712](https://github.com/QwenLM/qwen-code/pull/6712)) by @yiliang114
- MCP 服务器的 OAuth 认证在收到 HTTP 401 响应后现在能够自动恢复。([#6732](https://github.com/QwenLM/qwen-code/pull/6732)) by @yiliang114
- Web Shell 中的内联标签提示框不再重复显示。([#6729](https://github.com/QwenLM/qwen-code/pull/6729)) by @dreamWB
- 目标评审判决现在会忽略推理内容，仅评估最终答案。([#6738](https://github.com/QwenLM/qwen-code/pull/6738)) by @qqqys
- 定时任务对话框现在显示正确的标签图标，而非过时的引用。 ([#6748](https://github.com/QwenLM/qwen-code/pull/6748)) by @wenshao
- 聊天编辑器中的 Git 分支标签在窗口较窄时不再将工具栏控件挤出屏幕。 ([#6753](https://github.com/QwenLM/qwen-code/pull/6753)) by @wenshao
- 聊天录制失败现在会被持久化并提示给用户，而不是静默失败。 ([#6743](https://github.com/QwenLM/qwen-code/pull/6743)) by @doudouOUC
- 来自模型的格式错误流式响应现在会自动重试。 ([#6754](https://github.com/QwenLM/qwen-code/pull/6754)) by @yiliang114
- 当计划模式阻止写入操作时，代理会被引导使用只读工具，而不是直接停止。 ([#6764](https://github.com/QwenLM/qwen-code/pull/6764)) by @Alex-ai-future
- 飞书凭证现在会在建立 WebSocket 连接前进行验证，以防止静默失败。 ([#6780](https://github.com/QwenLM/qwen-code/pull/6780)) by @BenGuanRan
- 桌面版发布说明现在会被截断，以避免 CI 中生成过大的负载。 ([#6792](https://github.com/QwenLM/qwen-code/pull/6792)) by @DragonnZhang
- 在多工作区设置中，会话操作现在会路由到正确的工作区所有者。 ([#6798](https://github.com/QwenLM/qwen-code/pull/6798)) by @doudouOUC
- Web UI 现在链接到工作区本地的 SDK 包，解除了 SDK 发布的阻塞。 ([#6823](https://github.com/QwenLM/qwen-code/pull/6823)) by @qwen-code-dev-bot
- 代码审查不再丢弃实时阻塞项，并会验证新测试是否确实能覆盖变更代码。 ([#6790](https://github.com/QwenLM/qwen-code/pull/6790)) by @wenshao
- 频道中的斜杠命令输出现在在视觉上与常规消息区分开来。 ([#6818](https://github.com/QwenLM/qwen-code/pull/6818)) by @qqqys
- Web 终端中的 React 18 下已恢复打包对话框样式。 ([#6827](https://github.com/QwenLM/qwen-code/pull/6827)) by @ytahdn
- 推理时长显示现在能展示更准确、更精细的时间信息。 ([#6793](https://github.com/QwenLM/qwen-code/pull/6793)) by @tanzhenxin
- 会话继续、语言选择和工件操作现在会路由到正确的工作区所有者。 ([#6833](https://github.com/QwenLM/qwen-code/pull/6833)) by @doudouOUC
- 即使存储的值为假值，LRU 缓存项在访问时现在也能正确重新排序。 ([#6787](https://github.com/QwenLM/qwen-code/pull/6787)) by @chinesepowered
- 会话的最新活跃时间戳现在基于实际活动时间计算，而非当前系统时钟。 ([#6834](https://github.com/QwenLM/qwen-code/pull/6834)) by @chinesepowered
- 等待待处理重写完成时排队的重写任务，现在会在继续前被完全处理完毕。 ([#6800](https://github.com/QwenLM/qwen-code/pull/6800)) by @chinesepowered
- LLM 重写器现在将其输出历史记录限制在配置的上下文轮次窗口内。 ([#6799](https://github.com/QwenLM/qwen-code/pull/6799)) by @chinesepowered
- 仅从 GitHub 仓库名称中移除末尾的 `.git` 后缀，保留内部的 `.git` 片段。 ([#6797](https://github.com/QwenLM/qwen-code/pull/6797)) by @chinesepowered
- 如 `.eslintrc` 等点文件现在能根据文件路径被正确检测并分配语言。 ([#6785](https://github.com/QwenLM/qwen-code/pull/6785)) by @chinesepowered
- 修复了一个安全问题：洞察报告数据中的尖括号可能导致脚本上下文逃逸。 ([#6802](https://github.com/QwenLM/qwen-code/pull/6802)) by @chinesepowered
- 审查块代理现在会接收正确构造的提示词，从而不再在无上下文的情况下运行。 ([#6840](https://github.com/QwenLM/qwen-code/pull/6840)) by @wenshao
- 重新引入了针对格式错误流式响应的重试逻辑，并采用更精确的检测以避免误报。 ([#6794](https://github.com/QwenLM/qwen-code/pull/6794)) by @yiliang114
- 审查覆盖率现在使用测试框架自身的记录进行验证，以确保结果准确。 ([#6843](https://github.com/QwenLM/qwen-code/pull/6843)) by @wenshao
- 将浏览器守护进程包大小限制提升至 156 KB，以支持更大的构建。 ([#6852](https://github.com/QwenLM/qwen-code/pull/6852)) by @qwen-code-dev-bot
### 性能
- 通过延迟加载 tree-sitter 解析器运行时，提升了启动性能。 ([#6747](https://github.com/QwenLM/qwen-code/pull/6747)) by @dexhunter
- 减少了会话期间生成的 Git 快照进程数量，以提升性能。 ([#6784](https://github.com/QwenLM/qwen-code/pull/6784)) by @dexhunter
### 文档
- 开发工作流现在在构建和审查之间增加了一个明确的自审步骤，要求作者在声明工作完成前，基于完整上下文验证变更。 ([#6685](https://github.com/QwenLM/qwen-code/pull/6685)) by @wenshao
- 修正了不准确的 token 限制注释，并修复了核心包代码注释中的拼写错误，无行为变更。 ([#6698](https://github.com/QwenLM/qwen-code/pull/6698)) by @chinesepowered
### 内部变更
- 将默认自动 PR 审查超时时间延长至 180 分钟，并允许覆盖至 240 分钟，以防止大型 Pull Request 超时。([#6706](https://github.com/QwenLM/qwen-code/pull/6706)) by @wenshao
- 移除了废弃的钉钉（DingTalk）规划集成产物。([#6722](https://github.com/QwenLM/qwen-code/pull/6722)) by @qqqys
- 由于存在问题，回退了格式错误的流式响应重试逻辑。([#6783](https://github.com/QwenLM/qwen-code/pull/6783)) by @wenshao
- 稳定了工具控制与子代理场景的端到端测试用例。([#6803](https://github.com/QwenLM/qwen-code/pull/6803)) by @yiliang114
- 移除了不稳定的无头（headless）子进程录制测试。([#6830](https://github.com/QwenLM/qwen-code/pull/6830)) by @yiliang114
- 审查测试效能探针现运行于隔离的临时工作树（worktrees）中。([#6836](https://github.com/QwenLM/qwen-code/pull/6836)) by @wenshao
- 在审查探针间共享了统一的工作树路径辅助函数，并改进了过期工作树的清理机制。([#6841](https://github.com/QwenLM/qwen-code/pull/6841)) by @wenshao
## 新贡献者
- @juhuan 在 [#6492](https://github.com/QwenLM/qwen-code/pull/6492) 中首次贡献
- @ksws00684315 在 [#6672](https://github.com/QwenLM/qwen-code/pull/6672) 中首次贡献
- @kaymeer 在 [#6607](https://github.com/QwenLM/qwen-code/pull/6607) 中首次贡献
- @destire-mio 在 [#6709](https://github.com/QwenLM/qwen-code/pull/6709) 中首次贡献
- @yanchenko 在 [#6489](https://github.com/QwenLM/qwen-code/pull/6489) 中首次贡献
- @ever-o 在 [#6537](https://github.com/QwenLM/qwen-code/pull/6537) 中首次贡献
**完整更新日志**: https://github.com/QwenLM/qwen-code/compare/v0.19.9...v0.19.10
<!-- qwen-release-notes:v1 -->
## Highlights
- Multi-workspace support now spans ACP transport, daemon workers, split-view sessions, and workspace-aware actions. ([#6621](https://github.com/QwenLM/qwen-code/pull/6621), [#6635](https://github.com/QwenLM/qwen-code/pull/6635), [#6746](https://github.com/QwenLM/qwen-code/pull/6746), [#6724](https://github.com/QwenLM/qwen-code/pull/6724))
- Web Shell gains an artifact panel, a modernized sidebar and settings experience, and editable user-level model configuration. ([#6591](https://github.com/QwenLM/qwen-code/pull/6591), [#6804](https://github.com/QwenLM/qwen-code/pull/6804), [#6817](https://github.com/QwenLM/qwen-code/pull/6817), [#6768](https://github.com/QwenLM/qwen-code/pull/6768))
- Session and model recovery is more resilient to malformed streams, leaked protocol turns, OAuth expiry, and daemon restarts. ([#6794](https://github.com/QwenLM/qwen-code/pull/6794), [#6683](https://github.com/QwenLM/qwen-code/pull/6683), [#6732](https://github.com/QwenLM/qwen-code/pull/6732), [#6680](https://github.com/QwenLM/qwen-code/pull/6680))
- Subagents are easier to follow with richer live command context, transcript paths, a chronological timeline, and inherited Explore models. ([#6580](https://github.com/QwenLM/qwen-code/pull/6580), [#6772](https://github.com/QwenLM/qwen-code/pull/6772), [#6807](https://github.com/QwenLM/qwen-code/pull/6807))
- CLI workflows gain project-scoped prompt stashes, visible Git branches, configurable shell timeouts, and safer memory-pressure handling. ([#6709](https://github.com/QwenLM/qwen-code/pull/6709), [#6725](https://github.com/QwenLM/qwen-code/pull/6725), [#6628](https://github.com/QwenLM/qwen-code/pull/6628), [#6682](https://github.com/QwenLM/qwen-code/pull/6682))
- SDK and daemon integrations can now control effort, models, usage, context, transport options, and runtime channels. ([#6492](https://github.com/QwenLM/qwen-code/pull/6492), [#6491](https://github.com/QwenLM/qwen-code/pull/6491), [#6741](https://github.com/QwenLM/qwen-code/pull/6741))
## Breaking Changes
No known breaking changes.
## Complete Change List
### Features
- SDK-hosted agents now forward user answers from ask_user_question tool approvals back to the model, enabling interactive question flows in TypeScript and Python SDK integrations. ([#6655](https://github.com/QwenLM/qwen-code/pull/6655)) by @TianYuan1024
- DingTalk channels can now optionally mention group message senders in bot replies using visible @ tokens, with Markdown rendering preserved for replies without mentions. ([#6679](https://github.com/QwenLM/qwen-code/pull/6679)) by @qqqys
- Web shell now includes a right-side review panel for inspecting edited files, artifacts, and scheduled tasks with diffs, file navigation, and session-scoped tabs. ([#6591](https://github.com/QwenLM/qwen-code/pull/6591)) by @ytahdn
- Daemon now supports workspace-qualified ACP endpoints, allowing SDK and web shell clients to open sessions scoped to specific workspaces in multi-workspace deployments. ([#6621](https://github.com/QwenLM/qwen-code/pull/6621)) by @doudouOUC
- Adds a composer header render slot and an opt-in mobile welcome footer placement so embedders can show custom content above the composer on small screens. ([#6584](https://github.com/QwenLM/qwen-code/pull/6584)) by @yuanyuanAli
- Adds SDK methods to control reasoning effort, list available models, and query account and context usage statistics at runtime. ([#6492](https://github.com/QwenLM/qwen-code/pull/6492)) by @juhuan
- Expanding thinking blocks with Alt+T during streaming now displays the full reasoning content in real-time instead of a truncated preview. ([#6678](https://github.com/QwenLM/qwen-code/pull/6678)) by @huww98
- Scheduled task prompts now support inline extension, skill, and MCP reference tags via floating pickers that serialize cleanly into the saved prompt text. ([#6589](https://github.com/QwenLM/qwen-code/pull/6589)) by @BZ-D
- Pressing Ctrl+S with a non-empty prompt now stashes it for the current project and restores it on the next launch, clearing the stash upon submission. ([#6709](https://github.com/QwenLM/qwen-code/pull/6709)) by @destire-mio
- Session recovery now uses a unified plan across all entrypoints, consistently handling interrupted prompts, dangling tool calls, and degraded history. ([#6731](https://github.com/QwenLM/qwen-code/pull/6731)) by @water-in-stone
- Subagent views now show untruncated live commands, deeper progress history, execution transcript paths, and recent tool context on inline approvals. ([#6580](https://github.com/QwenLM/qwen-code/pull/6580)) by @TianYuan1024
- The qwen serve daemon can now host channel workers for multiple trusted workspaces, binding each worker to its respective workspace directory. ([#6635](https://github.com/QwenLM/qwen-code/pull/6635)) by @doudouOUC
- A new MessageDisplay hook fires incrementally during assistant streaming, allowing observers to track reply text before the turn completes. ([#6489](https://github.com/QwenLM/qwen-code/pull/6489)) by @yanchenko
- Double-clicking a cell in a rendered Markdown table now opens a dialog with the full cell value for easy reading and copying. ([#6530](https://github.com/QwenLM/qwen-code/pull/6530)) by @jifeng
- The daemon now exposes read-only session catalogs for untrusted workspaces, letting clients browse sessions without write access. ([#6717](https://github.com/QwenLM/qwen-code/pull/6717)) by @doudouOUC
- Workspace registrations added at runtime are now persisted so they survive daemon restarts without manual reconfiguration. ([#6716](https://github.com/QwenLM/qwen-code/pull/6716)) by @doudouOUC
- User messages in the web shell now render extension, skill, and MCP reference tags inline, matching the composer preview. ([#6537](https://github.com/QwenLM/qwen-code/pull/6537)) by @ever-o
- Channel workers now reconnect automatically when the daemon restarts, restoring DingTalk and Feishu sessions without manual intervention. ([#6680](https://github.com/QwenLM/qwen-code/pull/6680)) by @qqqys
- The composer toolbar now displays the current Git branch as a chip so you always know which branch you're working on. ([#6725](https://github.com/QwenLM/qwen-code/pull/6725)) by @han-dreamer
- Foreground shell commands now respect a configurable default timeout, preventing long-running commands from blocking indefinitely. ([#6628](https://github.com/QwenLM/qwen-code/pull/6628)) by @Nas01010101
- Both the TypeScript and Python SDKs now expose transport selection and query options for finer control over daemon communication. ([#6491](https://github.com/QwenLM/qwen-code/pull/6491)) by @juhuan
- The web shell session sidebar is now configurable, letting embedders choose which session metadata and actions to display. ([#6750](https://github.com/QwenLM/qwen-code/pull/6750)) by @dreamWB
- Workspace transcript reader now provides persisted conversation history per workspace for server-side session browsing. ([#6740](https://github.com/QwenLM/qwen-code/pull/6740)) by @doudouOUC
- Session groups in the web shell now support custom Hex color codes for easier visual organization. ([#6752](https://github.com/QwenLM/qwen-code/pull/6752)) by @samuelhsin
- The composer model selector has been redesigned with clearer model metadata and faster switching. ([#6758](https://github.com/QwenLM/qwen-code/pull/6758)) by @dreamWB
- Code review now includes procedural correctness checks, adjustable effort levels, and guardrails that verify findings before posting. ([#6711](https://github.com/QwenLM/qwen-code/pull/6711)) by @wenshao
- Multi-workspace split view now supports cross-workspace sessions, workspace labels, and a responsive layout for side-by-side editing. ([#6746](https://github.com/QwenLM/qwen-code/pull/6746)) by @wenshao
- Scheduled tasks from all workspaces are now aggregated into a single view in the web shell for easier monitoring. ([#6759](https://github.com/QwenLM/qwen-code/pull/6759)) by @wenshao
- The composer now accepts custom placeholder text, allowing embedders to show context-specific prompts to users. ([#6765](https://github.com/QwenLM/qwen-code/pull/6765)) by @dreamWB
- Release notes are now drafted with AI assistance, summarizing changes from the full commit range into categorized, user-facing bullets. ([#6756](https://github.com/QwenLM/qwen-code/pull/6756)) by @yiliang114
- Web shell now ships a shadcn-based UI foundation with consistent design tokens and accessible component primitives. ([#6760](https://github.com/QwenLM/qwen-code/pull/6760)) by @ytahdn
- Sub-agent activity is now shown as a chronological transcript with a parallel-agent timeline for easier multi-task inspection. ([#6772](https://github.com/QwenLM/qwen-code/pull/6772)) by @wenshao
- The CLI can now start, stop, and list daemon channel workers at runtime without restarting the daemon process. ([#6741](https://github.com/QwenLM/qwen-code/pull/6741)) by @doudouOUC
- Persisted conversation transcripts are now bounded to prevent unbounded disk growth. ([#6769](https://github.com/QwenLM/qwen-code/pull/6769)) by @doudouOUC
- Web shell fires a callback when a new session is created, enabling custom integrations. ([#6703](https://github.com/QwenLM/qwen-code/pull/6703)) by @ytahdn
- Skill installation paths are now exposed so users can see where skills are stored. ([#6811](https://github.com/QwenLM/qwen-code/pull/6811)) by @callmeYe
- Multi-workspace sidebar redesigned with a cleaner, more modern layout. ([#6804](https://github.com/QwenLM/qwen-code/pull/6804)) by @ytahdn
- Settings page refreshed with a modern component library for a polished look and feel. ([#6817](https://github.com/QwenLM/qwen-code/pull/6817)) by @ytahdn
- User-scope settings can now be edited directly in the panel, with in-panel model management. ([#6768](https://github.com/QwenLM/qwen-code/pull/6768)) by @wenshao
- Rewind and shell actions now work across multiple workspaces. ([#6826](https://github.com/QwenLM/qwen-code/pull/6826)) by @doudouOUC
- Workspaces can now be removed at runtime without restarting the server. ([#6745](https://github.com/QwenLM/qwen-code/pull/6745)) by @doudouOUC
- Code review now captures untracked files, resolves anchors from snippets, and gates posting in code. ([#6771](https://github.com/QwenLM/qwen-code/pull/6771)) by @wenshao
- PR triage comments now include a confidence score, sequence diagram, files overview, and review footer. ([#6789](https://github.com/QwenLM/qwen-code/pull/6789)) by @wenshao
- Explore subagent now inherits the main model by default instead of using a fixed smaller model. ([#6807](https://github.com/QwenLM/qwen-code/pull/6807)) by @tanzhenxin
- Daemon status now reports model API error counts and retry metrics. ([#6837](https://github.com/QwenLM/qwen-code/pull/6837)) by @wenshao
- New API endpoint lets you toggle workspace skills programmatically. ([#6816](https://github.com/QwenLM/qwen-code/pull/6816)) by @callmeYe
- Added xAI Grok as a built-in model provider preset. ([#6805](https://github.com/QwenLM/qwen-code/pull/6805)) by @TianYuan1024
- Extension management upgraded to v2 with improved install, update, and removal flows. ([#6825](https://github.com/QwenLM/qwen-code/pull/6825)) by @doudouOUC
- Voice input is now qualified per workspace for better multi-workspace support. ([#6839](https://github.com/QwenLM/qwen-code/pull/6839)) by @doudouOUC
- Session export is now workspace-qualified so exports stay organized across workspaces. ([#6844](https://github.com/QwenLM/qwen-code/pull/6844)) by @doudouOUC
- Markdown tables in web shell now show row, column, and cell selection statistics. ([#6838](https://github.com/QwenLM/qwen-code/pull/6838)) by @jifeng
### Bug Fixes
- YOLO approval mode is no longer silently disabled when the model enters plan mode. ([#6630](https://github.com/QwenLM/qwen-code/pull/6630)) by @Nas01010101
- Approval mode UI labels and notices are now localized, preventing mixed-language output when cycling modes with Shift+Tab in non-English interfaces. ([#6592](https://github.com/QwenLM/qwen-code/pull/6592)) by @han-dreamer
- Channel delivery now filters out nested subagent output, preventing intermediate research reports from being sent to messaging platforms before the root agent's final answer. ([#6696](https://github.com/QwenLM/qwen-code/pull/6696)) by @qqqys
- Raised prepared npm package size limit from 80 MB to 96 MB to accommodate natural growth and unblock Docker sandbox builds. ([#6691](https://github.com/QwenLM/qwen-code/pull/6691)) by @qwen-code-dev-bot
- Fixed Docker sandbox networking for protocol tag retry tests, enabling CLI to reach test servers via host.docker.internal when running in container sandbox mode. ([#6692](https://github.com/QwenLM/qwen-code/pull/6692)) by @qwen-code-dev-bot
- Fixed cursor position being incorrectly overridden when set to the start of a line in VS Code. ([#2971](https://github.com/QwenLM/qwen-code/pull/2971)) by @chinesepowered
- Fixes out-of-memory crashes on exit during long interactive sessions by running periodic memory-pressure checks even when no tool calls occur. ([#6682](https://github.com/QwenLM/qwen-code/pull/6682)) by @yiliang114
- Reduces token usage in mobile UI dumps by stripping bounds attributes with negative coordinates for partially off-screen elements. ([#6624](https://github.com/QwenLM/qwen-code/pull/6624)) by @chinesepowered
- Fixes leaked protocol tags in responses by discarding and retrying invalid assistant turns during max-token escalation and continuation recovery. ([#6683](https://github.com/QwenLM/qwen-code/pull/6683)) by @yiliang114
- Updates test expectations to match the new automated PR review workflow timeout values, resolving continuous integration failures. ([#6720](https://github.com/QwenLM/qwen-code/pull/6720)) by @yiliang114
- Improves plan mode reliability by returning a clear error response when non-read-only tools are blocked, preventing the model from attempting workarounds. ([#6667](https://github.com/QwenLM/qwen-code/pull/6667)) by @Alex-ai-future
- Adds an Approve button to the /mcp server detail view, allowing users to approve pending or previously rejected workspace servers. ([#6518](https://github.com/QwenLM/qwen-code/pull/6518)) by @LaZzyMan
- Fixes remote MCP connection failures by preserving URL query parameters and fragments that end with a trailing slash during normalization. ([#6587](https://github.com/QwenLM/qwen-code/pull/6587)) by @VectorPeak
- Fixes subagent startup failures by restricting template placeholders to valid identifiers, treating numeric patterns like ${0} as literal text. ([#6672](https://github.com/QwenLM/qwen-code/pull/6672)) by @ksws00684315
- When running as an ACP agent, the user's typed instruction is now placed after attached file content so models prioritize the actual request. ([#6607](https://github.com/QwenLM/qwen-code/pull/6607)) by @kaymeer
- The ACP readTextFile fallback now rejects fractional line and limit values, returning consistent invalid parameter errors for malformed read windows. ([#6704](https://github.com/QwenLM/qwen-code/pull/6704)) by @VectorPeak
- The Add Workspace dialog now respects light and dark themes, and multi-workspace session lists restore full row features and per-session actions. ([#6705](https://github.com/QwenLM/qwen-code/pull/6705)) by @wenshao
- DingTalk Stream connections are now treated as ready when the socket is open, preventing healthy sessions from being disconnected due to missing registration frames. ([#6715](https://github.com/QwenLM/qwen-code/pull/6715)) by @qqqys
- Trusted non-primary workspaces can now update session pin, group, and color state via a new workspace-scoped organization endpoint. ([#6724](https://github.com/QwenLM/qwen-code/pull/6724)) by @doudouOUC
- Claude Opus 4.6, 4.7, and 4.8 models now default to a 1M context window and 128K output limit, preventing premature compaction. ([#6718](https://github.com/QwenLM/qwen-code/pull/6718)) by @yiliang114
- Clipboard image pasting now works correctly in standalone desktop builds. ([#6708](https://github.com/QwenLM/qwen-code/pull/6708)) by @zjunothing
- Saved memory instructions are now refreshed immediately after using the remember command. ([#6497](https://github.com/QwenLM/qwen-code/pull/6497)) by @han-dreamer
- Goal evaluation is now lifecycle-safe, preventing errors during session teardown. ([#6681](https://github.com/QwenLM/qwen-code/pull/6681)) by @qqqys
- Managed memory entries are now preserved during microcompaction instead of being dropped. ([#6714](https://github.com/QwenLM/qwen-code/pull/6714)) by @yiliang114
- Claude output token limits now use decimal values for more accurate context budgeting. ([#6735](https://github.com/QwenLM/qwen-code/pull/6735)) by @yiliang114
- Model and approval-mode changes now apply correctly to non-primary workspace sessions in the web shell. ([#6737](https://github.com/QwenLM/qwen-code/pull/6737)) by @wenshao
- The agent no longer crashes when receiving repeated invalid model streams and recovers gracefully. ([#6712](https://github.com/QwenLM/qwen-code/pull/6712)) by @yiliang114
- OAuth authentication for MCP servers now recovers automatically after an HTTP 401 response. ([#6732](https://github.com/QwenLM/qwen-code/pull/6732)) by @yiliang114
- Inline tag tooltips in the web shell no longer appear as duplicates. ([#6729](https://github.com/QwenLM/qwen-code/pull/6729)) by @dreamWB
- Goal judge verdicts now ignore reasoning content and evaluate only the final answer. ([#6738](https://github.com/QwenLM/qwen-code/pull/6738)) by @qqqys
- The Scheduled Tasks dialog now shows correct tag icons instead of stale references. ([#6748](https://github.com/QwenLM/qwen-code/pull/6748)) by @wenshao
- The Git branch chip in the composer no longer pushes toolbar controls off-screen on narrow windows. ([#6753](https://github.com/QwenLM/qwen-code/pull/6753)) by @wenshao
- Chat recording failures are now persisted and surfaced to the user instead of failing silently. ([#6743](https://github.com/QwenLM/qwen-code/pull/6743)) by @doudouOUC
- Malformed streamed responses from the model are now retried automatically. ([#6754](https://github.com/QwenLM/qwen-code/pull/6754)) by @yiliang114
- When plan mode blocks write actions the agent is guided to use read-only tools instead of stopping. ([#6764](https://github.com/QwenLM/qwen-code/pull/6764)) by @Alex-ai-future
- Feishu credentials are now validated before starting the WebSocket connection, preventing silent failures. ([#6780](https://github.com/QwenLM/qwen-code/pull/6780)) by @BenGuanRan
- Desktop release notes are now truncated to avoid oversized payloads in CI. ([#6792](https://github.com/QwenLM/qwen-code/pull/6792)) by @DragonnZhang
- Session actions are now routed to the correct owning workspace in multi-workspace setups. ([#6798](https://github.com/QwenLM/qwen-code/pull/6798)) by @doudouOUC
- The web UI now links to the workspace-local SDK package, unblocking SDK publishing. ([#6823](https://github.com/QwenLM/qwen-code/pull/6823)) by @qwen-code-dev-bot
- Code review no longer drops live blockers and now verifies that new tests actually gate the changed code. ([#6790](https://github.com/QwenLM/qwen-code/pull/6790)) by @wenshao
- Slash-command output in channels is now visually distinguished from regular messages. ([#6818](https://github.com/QwenLM/qwen-code/pull/6818)) by @qqqys
- Packaged dialog styles are restored under React 18 in the web shell. ([#6827](https://github.com/QwenLM/qwen-code/pull/6827)) by @ytahdn
- Reasoning duration displays now show more accurate and refined timing information. ([#6793](https://github.com/QwenLM/qwen-code/pull/6793)) by @tanzhenxin
- Session continue, language selection, and artifact actions are now routed to the correct workspace owner. ([#6833](https://github.com/QwenLM/qwen-code/pull/6833)) by @doudouOUC
- LRU cache entries are now correctly reordered on access even when the stored value is falsy. ([#6787](https://github.com/QwenLM/qwen-code/pull/6787)) by @chinesepowered
- The latest-active timestamp for sessions is now computed from the real activity time, not the current wall clock. ([#6834](https://github.com/QwenLM/qwen-code/pull/6834)) by @chinesepowered
- Rewrites queued while waiting for pending rewrites to finish are now fully drained before proceeding. ([#6800](https://github.com/QwenLM/qwen-code/pull/6800)) by @chinesepowered
- The LLM rewriter now limits its output history to the configured context turn window. ([#6799](https://github.com/QwenLM/qwen-code/pull/6799)) by @chinesepowered
- Only a trailing .git suffix is stripped from GitHub repo names, preserving internal .git segments. ([#6797](https://github.com/QwenLM/qwen-code/pull/6797)) by @chinesepowered
- Dotfiles such as .eslintrc are now detected and assigned the correct language from their file path. ([#6785](https://github.com/QwenLM/qwen-code/pull/6785)) by @chinesepowered
- Fixed a security issue where angle brackets in insight report data could break out of script contexts. ([#6802](https://github.com/QwenLM/qwen-code/pull/6802)) by @chinesepowered
- Review chunk agents now receive properly constructed prompts so they no longer run without context. ([#6840](https://github.com/QwenLM/qwen-code/pull/6840)) by @wenshao
- Reintroduced retry logic for malformed streaming responses with more precise detection to avoid false positives. ([#6794](https://github.com/QwenLM/qwen-code/pull/6794)) by @yiliang114
- Review coverage is now verified using the test harness's own records for accurate results. ([#6843](https://github.com/QwenLM/qwen-code/pull/6843)) by @wenshao
- Increased the browser daemon bundle size limit to 156 KB to accommodate larger builds. ([#6852](https://github.com/QwenLM/qwen-code/pull/6852)) by @qwen-code-dev-bot
### Performance
- Improved startup performance by lazy-loading the tree-sitter parser runtime. ([#6747](https://github.com/QwenLM/qwen-code/pull/6747)) by @dexhunter
- Reduced the number of Git snapshot processes spawned during sessions for better performance. ([#6784](https://github.com/QwenLM/qwen-code/pull/6784)) by @dexhunter
### Documentation
- Development workflow now includes a bounded self-audit step between build and review, requiring authors to verify changes with full context before declaring work complete. ([#6685](https://github.com/QwenLM/qwen-code/pull/6685)) by @wenshao
- Corrected inaccurate token limit comment and fixed typos in core package code comments with no behavior changes. ([#6698](https://github.com/QwenLM/qwen-code/pull/6698)) by @chinesepowered
### Internal Changes
- Increases the default automated PR review timeout to 180 minutes and allows overrides up to 240 minutes to prevent timeouts on large pull requests. ([#6706](https://github.com/QwenLM/qwen-code/pull/6706)) by @wenshao
- Removed obsolete DingTalk planning integration artifacts. ([#6722](https://github.com/QwenLM/qwen-code/pull/6722)) by @qqqys
- Reverted the malformed streamed response retry logic due to issues. ([#6783](https://github.com/QwenLM/qwen-code/pull/6783)) by @wenshao
- Stabilized end-to-end test cases for tool control and subagent scenarios. ([#6803](https://github.com/QwenLM/qwen-code/pull/6803)) by @yiliang114
- Removed a flaky headless child-process recording test. ([#6830](https://github.com/QwenLM/qwen-code/pull/6830)) by @yiliang114
- Review test-efficacy probes now run in isolated disposable worktrees. ([#6836](https://github.com/QwenLM/qwen-code/pull/6836)) by @wenshao
- Shared a common worktree path helper across review probes and improved stale worktree cleanup. ([#6841](https://github.com/QwenLM/qwen-code/pull/6841)) by @wenshao
## New Contributors
- @juhuan made their first contribution in [#6492](https://github.com/QwenLM/qwen-code/pull/6492)
- @ksws00684315 made their first contribution in [#6672](https://github.com/QwenLM/qwen-code/pull/6672)
- @kaymeer made their first contribution in [#6607](https://github.com/QwenLM/qwen-code/pull/6607)
- @destire-mio made their first contribution in [#6709](https://github.com/QwenLM/qwen-code/pull/6709)
- @yanchenko made their first contribution in [#6489](https://github.com/QwenLM/qwen-code/pull/6489)
- @ever-o made their first contribution in [#6537](https://github.com/QwenLM/qwen-code/pull/6537)
**Full Changelog**: https://github.com/QwenLM/qwen-code/compare/v0.19.9...v0.19.10