Qwen CUA Driver 预编译二进制文件
（托管于 `packages/cua-driver` 下）。
- **macOS**：代码签名 + 公证通用二进制文件 + `QwenCuaDriver.app`
- **Linux**：未签名（x86_64 + arm64，glibc 2.31 最低版本）
- **Windows**：未签名（x86_64 + arm64）
- **Node.js**：相同的工作流在针对这些发布资源进行干净安装后，发布单个
`@qwen-code/cua-sdk` npm 包
启用相对坐标：`CUA_DRIVER_RS_COORDINATE_SPACE=1`
（默认 `0` = 关闭；可选 `CUA_DRIVER_RS_COORDINATE_SCALE=1000`）。
为受影响的模型 API 路由启用文本 MCP 负载过滤：
`MCP_MODEL_PAYLOAD_FILTER=1`（默认关闭）。
<!-- 使用 .github/release.yml 中的配置（提交哈希 5cff52c6e856e6528408863d83eb36eba6058b14）生成发布说明 -->
## 更新内容
* feat(web-shell)：将审批和询问用户对话框作为流程内工作表；修复后台代理误报失败 by @ytahdn in https://github.com/QwenLM/qwen-code/pull/9351
* feat(serve)：添加实时状态会话活动水印 by @doudouOUC in https://github.com/QwenLM/qwen-code/pull/9396
* perf(web-shell)：保持流式输出响应灵敏 by @ytahdn in https://github.com/QwenLM/qwen-code/pull/9405
* feat(serve)：测量 ACP 子进程峰值老年代堆内存 by @doudouOUC in https://github.com/QwenLM/qwen-code/pull/9380
* feat(review)：通过 /review 命令、审查运行和 CI 重试连接 --resume 参数 by @wenshao in https://github.com/QwenLM/qwen-code/pull/9153
* feat(review)：在账本标记中记录每轮的发布量 by @wenshao in https://github.com/QwenLM/qwen-code/pull/9413
* feat(review)：使验证器从实时双臂验证中学习四种运行规范 by @wenshao in https://github.com/QwenLM/qwen-code/pull/9447
* feat(daemon)：添加批量扩展激活 API by @callmeYe in https://github.com/QwenLM/qwen-code/pull/8788
* fix(triage)：将 flake-gate 文件列表与固定的基础 OID 进行差异比对 by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/9464
* fix(cli)：将守护进程重复工具调用中断器暴露为可见的循环检测停止 by @doudouOUC in https://github.com/QwenLM/qwen-code/pull/9435
* feat(providers)：在 Token Plan 模型列表中添加 qwen3.8-max by @qqqys in https://github.com/QwenLM/qwen-code/pull/9383
* fix(cli)：防止 /rewind 在 /compress-fast 之后丢弃对话历史 by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/9331
* fix(core)：拒绝为命名队友设置 run_in_background: false by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/9433
* fix(review)：在源头限制发布量，而不仅是在写入位置 by @wenshao in https://github.com/QwenLM/qwen-code/pull/9460
* feat(extensions)：支持经过身份验证的 HTTPS Git 安装 by @callmeYe in https://github.com/QwenLM/qwen-code/pull/9458
* fix(core)：明确 list_agents 排除 Agent Team 队友 by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/9432
* fix(core)：拒绝上游快速失败占位符响应 by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/8938
* fix(core)：隔离图像负载驱逐状态 by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/9423
* fix(core)：仅在参数匹配时将重复的提供商工具调用 ID 视为重放 by @doudouOUC in https://github.com/QwenLM/qwen-code/pull/9436
* feat(telemetry)：将守护进程 HTTP 请求跨度链接到入站 W3C traceparent by @chiga0 in https://github.com/QwenLM/qwen-code/pull/9391
* feat(web-shell)：添加转录合同预验证 by @water-in-stone in https://github.com/QwenLM/qwen-code/pull/9388
* fix(triage)：在 env -i 重新执行前计算 flake-gate 差异 by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/9468
* fix(sdk)：将未识别的诊断信息路由到受限的转录侧信道 by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/9202
* fix(ci)：将 qwen-autofix 克隆到恢复工作流实体中 by @wenshao in https://github.com/QwenLM/qwen-code/pull/9482
* refactor(cli)：在遗留审计技能之前整合共享辅助函数 by @wenshao in https://github.com/QwenLM/qwen-code/pull/9345
* fix(ci)：执行无操作 touch 以重新注册 autofix 工作流触发器 by @wenshao in https://github.com/QwenLM/qwen-code/pull/9479
* fix(review)：确保阻塞项保留在 COMMENT 正文中，以便所有软化路径都能访问 by @wenshao in https://github.com/QwenLM/qwen-code/pull/9416
* feat(web-shell)：消费实时状态会话活动时间戳 by @doudouOUC in https://github.com/QwenLM/qwen-code/pull/9476
* refactor(review)：仅定义一次每个认证条原子 by @wenshao in https://github.com/QwenLM/qwen-code/pull/9473
* feat(cli)：在首个提示词旁显示已加载的上下文文件 by @ZijianZhang989 in https://github.com/QwenLM/qwen-code/pull/8855
* feat(review)：添加 runtime-axis、table-sweep 和隔离见证表单 by @wenshao in https://github.com/QwenLM/qwen-code/pull/9445
* fix(ci)：将 qwen-autofix.yml 保持在 GitHub 500 KB 启动运行限制之下 by @wenshao in https://github.com/QwenLM/qwen-code/pull/9517
* feat(cli)：添加代理视图 PTY 工作进程 by @ZijianZhang989 in https://github.com/QwenLM/qwen-code/pull/7800
* fix(dingtalk)：附加引用消息中的媒体 by @qqqys in https://github.com/QwenLM/qwen-code/pull/9347
* feat(core)：暴露工作流执行状态 by @qqqys in https://github.com/QwenLM/qwen-code/pull/9034
* refactor(review)：通过单一针脚路由认证路径原子 by @wenshao in https://github.com/QwenLM/qwen-code/pull/9484
* chore(ci)：为 npm publish 添加 --provenance 和 id-token 权限 by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/9532
* fix(cli)：当 brew 没有更新时抑制 Homebrew 更新通知 (#9493) by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/9502
* feat(cli)：添加独立的对话隔离原语 by @doudouOUC in https://github.com/QwenLM/qwen-code/pull/9341
* feat(web-shell)：通过 @ytahdn 在 https://github.com/QwenLM/qwen-code/pull/9477 统一文件上传与引用
* fix(web-shell)：通过 @ytahdn 在 https://github.com/QwenLM/qwen-code/pull/9533 保持侧边栏会话同步
* refactor：通过 @yiliang114 在 https://github.com/QwenLM/qwen-code/pull/9497 集中管理跨包契约
* fix(cli)：通过 @wenshao 在 https://github.com/QwenLM/qwen-code/pull/9551 为变体删除测试提供附件 Mock
* feat(serve)：通过 @chiga0 在 https://github.com/QwenLM/qwen-code/pull/9426 持久化提示词终端账本以进行冷启动对账
* fix(review)：通过 @wenshao 在 https://github.com/QwenLM/qwen-code/pull/9221 在私有临时工作树中运行验证探针 (#9207)
* fix(serve)：通过 @doudouOUC 在 https://github.com/QwenLM/qwen-code/pull/9512 强化独立对话原语
* fix(tests)：通过 @yiliang114 在 https://github.com/QwenLM/qwen-code/pull/9559 在临时 ENOTEMPTY 错误时重试 acp-cron 清理 rm
* chore(ci)：通过 @yiliang114 在 https://github.com/QwenLM/qwen-code/pull/9552 弃用 NPM_TOKEN，改用 npm 可信发布
* fix(autofix)：通过 @wenshao 在 https://github.com/QwenLM/qwen-code/pull/9297 将制动器的 BLOCKED 交接提升为一级轮次结果
* fix(web-shell)：通过 @doudouOUC 在 https://github.com/QwenLM/qwen-code/pull/9563 停止重复的会话标题目录刷新
* fix(daemon)：通过 @ytahdn 在 https://github.com/QwenLM/qwen-code/pull/9567 避免使用自定义附件上传头
* feat(review)：通过 @wenshao 在 https://github.com/QwenLM/qwen-code/pull/9448 规范契约文档，并实现矩阵分层守卫
* feat(web-shell)：通过 @ytahdn 在 https://github.com/QwenLM/qwen-code/pull/9570 支持对话中途消息的文件附件
* feat：通过 @callmeYe 在 https://github.com/QwenLM/qwen-code/pull/9574 为 Qwen 混合模型注册仅切换推理
* feat(web-shell)：通过 @qqqys 在 https://github.com/QwenLM/qwen-code/pull/9393 采用标准 Goal v3 控件
* test(review)：通过 @wenshao 在 https://github.com/QwenLM/qwen-code/pull/9539 单编码 diff-read 夹具并固定失败读取门控
* refactor(review)：通过 @wenshao 在 https://github.com/QwenLM/qwen-code/pull/9267 基于 PR 的 diff 而非检查项构建增量范围
* fix(ci)：通过 @wenshao 在 https://github.com/QwenLM/qwen-code/pull/9518 在 shepherd 中停止将卡住的排队运行计为进行中
* fix(ci)：通过 @wenshao 在 https://github.com/QwenLM/qwen-code/pull/9498 修复符号链接工作区，避免运行器在其上卡死
* fix(autofix)：通过 @wenshao 在 https://github.com/QwenLM/qwen-code/pull/9578 在轮次提交前探测新守卫的变异
* test(ci)：通过 @wenshao 在 https://github.com/QwenLM/qwen-code/pull/9444 在 serve A/B 测试中暂存磁盘会话状态
* fix(ci)：通过 @wenshao 在 https://github.com/QwenLM/qwen-code/pull/9463 使 autofix 查找回复具备幂等性
* fix(ci)：通过 @yiliang114 在 https://github.com/QwenLM/qwen-code/pull/9592 使用第三方机器人 PAT 签署发布 PR
* fix(cli)：通过 @yiliang114 在 https://github.com/QwenLM/qwen-code/pull/9593 持久化跳过的重复工具结果
* fix(web-shell)：通过 @yiliang114 在 https://github.com/QwenLM/qwen-code/pull/9540 在非安全上下文中回退至 execCommand 复制
* fix(ui)：通过 @qwen-code-dev-bot 在 https://github.com/QwenLM/qwen-code/pull/9421 折叠由历史+待处理渲染的重复进行中 tool_group
* feat(review)：通过 @wenshao 在 https://github.com/QwenLM/qwen-code/pull/9491 通过 a1 CLI 将 --comment 评论发布至 Aone Code
* feat(review)：通过 @wenshao 在 https://github.com/QwenLM/qwen-code/pull/9461 告知作者为何审查循环未收敛
* fix(ci)：通过 @wenshao 在 https://github.com/QwenLM/qwen-code/pull/9462 防止回退评论拒绝其已发布的审查
* fix(web-shell)：通过 @wenshao 在 https://github.com/QwenLM/qwen-code/pull/9303 限制守护进程转录保留量以阻止渲染器 OOM 崩溃
* chore(ci)：通过 @yiliang114 在 https://github.com/QwenLM/qwen-code/pull/9577 禁用发布 CI 中的安装脚本并保护安全检查工作流
* feat(web-shell)：通过 @ytahdn 在 https://github.com/QwenLM/qwen-code/pull/9632 在后台 shell 运行时保持当前轮次展开
* feat(autofix)：通过 @wenshao 在 https://github.com/QwenLM/qwen-code/pull/9262 审计方法而非在超出增长预算时停止
* docs(ci)：通过 @wenshao 在 https://github.com/QwenLM/qwen-code/pull/9575 纠正两处错误注释，明确 ECS 池确实运行容器
* fix(autofix)：通过 @wenshao 在 https://github.com/QwenLM/qwen-code/pull/9527 将沙箱镜像绑定至其拉取的摘要
* docs(autofix)：通过 @wenshao 在 https://github.com/QwenLM/qwen-code/pull/9525 设计携带 PAT 步骤的运行器级隔离
* fix(web-shell)：通过 @ytahdn 在 https://github.com/QwenLM/qwen-code/pull/9640 解决并行代理折叠问题并统一代理详情转录
* fix(web-shell)：通过 @deggs7 在 https://github.com/QwenLM/qwen-code/pull/9400 显示创建组提示框
* feat(review)：通过 @wenshao 在 https://github.com/QwenLM/qwen-code/pull/9332 将单跳导入扩展折叠至 `fetch-pr --since` 中
* docs：通过 @DragonnZhang 在 https://github.com/QwenLM/qwen-code/pull/8656 记录内联终端图像预览
* feat(core)：通过 @DragonnZhang 在 https://github.com/QwenLM/qwen-code/pull/9424 将 list_directory 设为可选（默认禁用）
* docs：通过 @DragonnZhang 在 https://github.com/QwenLM/qwen-code/pull/8554 记录会话工作流设置
* feat(cli)：通过 @DragonnZhang 在 https://github.com/QwenLM/qwen-code/pull/9495 将非阻塞斜杠命令扩展至更多内置项
* docs(autofix)：通过 @wenshao 在 https://github.com/QwenLM/qwen-code/pull/9652 将发布端检查固定至补丁文本
* chore(deps)：通过 @yiliang114 在 https://github.com/QwenLM/qwen-code/pull/9584 清除高危 CVE 基线并加固安全网关
* docs(autofix): 为 `/takeover from N` 添加操作员指南 by @wenshao in https://github.com/QwenLM/qwen-code/pull/9622
* fix(artifacts): 将记录的目录扩展为按文件划分的制品 by @zjgzx1988 in https://github.com/QwenLM/qwen-code/pull/9395
* feat(review): 说明 Aone 帖子仅参与讨论门控 by @wenshao in https://github.com/QwenLM/qwen-code/pull/9625
* fix(review): 使增量缓存适用于 Aone AGit-Flow CR by @wenshao in https://github.com/QwenLM/qwen-code/pull/9630
* feat(review): 在预提交阶段检测 Aone 目标上的自提 MR by @wenshao in https://github.com/QwenLM/qwen-code/pull/9629
* fix(autofix): 在繁忙 PR 枚举中包含待处理运行 by @wenshao in https://github.com/QwenLM/qwen-code/pull/9662
* docs(autofix): 更正关于前导空格的 round-seed 指南 by @wenshao in https://github.com/QwenLM/qwen-code/pull/9663
* fix(web-shell): 使用守护进程实时状态 hasActivePrompt 作为加载指示器 (#9487) by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/9631
* fix(cli): 在响应流式传输期间保持斜杠菜单选择稳定 by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/9508
* fix(web-shell): 用户输入时不抢占审批焦点 by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/9609
* fix(core): 支持旧版 Git 的公共 GitHub 扩展 by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/9680
* feat(web-shell): 通过侧边栏徽标和搜索将 GitHub PR 绑定到会话 by @wenshao in https://github.com/QwenLM/qwen-code/pull/9543
* fix(ci): 升级存在漏洞的依赖以解除 CVE 审计阻塞 by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/9703
* fix(review): 在清理的绕过触发器中审计 Aone 目标 by @wenshao in https://github.com/QwenLM/qwen-code/pull/9633
* docs: 按执行机制对架构不变量进行分类 (#9152) by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/9689
* fix(autofix): 停止将空闲超时计入超时上限 by @wenshao in https://github.com/QwenLM/qwen-code/pull/9673
* feat(review): 为收敛观察提供机器可读部分 by @wenshao in https://github.com/QwenLM/qwen-code/pull/9623
* perf(web-shell): 优化流式转录本渲染 by @ytahdn in https://github.com/QwenLM/qwen-code/pull/9672
* fix(web-shell): 按会话活动路由目标消息 by @ytahdn in https://github.com/QwenLM/qwen-code/pull/9667
* i18n(cli): 翻译设置对话框中的 Unset by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/9714
* perf(review): 为审查代理分配独立的子代理类型 by @wenshao in https://github.com/QwenLM/qwen-code/pull/9678
* refactor(autofix): 将推送与报告逻辑移出工作流文件 by @wenshao in https://github.com/QwenLM/qwen-code/pull/9653
* feat(review): 报告驱动服务实际绑定的地址 by @wenshao in https://github.com/QwenLM/qwen-code/pull/9655
* fix(autofix): 通过门控的 env -i 启动传递 CI=true by @wenshao in https://github.com/QwenLM/qwen-code/pull/9649
* refactor(cli): 保持 acp-integration 在 serve 内部处于关闭状态 (#8084) by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/9144
* feat(review): 填补 Aone 剩余空白 — composeUrl、测试计划路由、a1 版本下限 by @wenshao in https://github.com/QwenLM/qwen-code/pull/9624
* fix(review): 停止优先丢弃收敛观察结果 by @wenshao in https://github.com/QwenLM/qwen-code/pull/9715
* fix(review): 从 Aone 写入路径中清除延迟的第 5 轮发现项 by @wenshao in https://github.com/QwenLM/qwen-code/pull/9604
* refactor(core): 移除根目录 barrel 自引用并强化边界 by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/9635
* fix(cli): 跨归档竞态条件恢复会话 by @doudouOUC in https://github.com/QwenLM/qwen-code/pull/9513
* feat(serve): 在会话加载/恢复时还原 ask_user_question 人机协同(HITL) by @doudouOUC in https://github.com/QwenLM/qwen-code/pull/9665
* fix(sdk): 支持 "auto" 权限模式 by @shenyankm in https://github.com/QwenLM/qwen-code/pull/9003
* fix(web-shell): 从概览面板打开时传递会话工作区 cwd by @wenshao in https://github.com/QwenLM/qwen-code/pull/9730
* fix(web-shell): 在过时的中途拒绝后回退 by @ytahdn in https://github.com/QwenLM/qwen-code/pull/9732
* fix(review): 停止在已替代的运行中触发回退评论 by @wenshao in https://github.com/QwenLM/qwen-code/pull/9716
* feat(review): 添加 temporal-reachability 和 incident-replay 透镜 by @wenshao in https://github.com/QwenLM/qwen-code/pull/9708
* fix(web-shell): 稳定固定区域顺序并提供即时固定反馈 by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/9560
* feat(review): 要求每个修复提供测试，并对不收敛做出裁决 by @wenshao in https://github.com/QwenLM/qwen-code/pull/9596
* chore(ci): 将 autofix 说明迁移至设计记录并控制增长 by @wenshao in https://github.com/QwenLM/qwen-code/pull/9677
* feat(cli): 将每个守护进程会话恢复至上次选择的模型 by @doudouOUC in https://github.com/QwenLM/qwen-code/pull/9687
* feat(review): 在发布前根据捕获的差异验证 Aone 行内锚点 by @wenshao in https://github.com/QwenLM/qwen-code/pull/9634
* fix(ci): 在 workflow 尺寸基线中记录 qwen-autofix.yml 的发布尺寸 by @wenshao in https://github.com/QwenLM/qwen-code/pull/9747
* fix(ci): 基于审查载荷提供的字段限制 fork 信号 by @qqqys in https://github.com/QwenLM/qwen-code/pull/9469
* feat(goal): 统计 Goal 消耗的 token by @qqqys in https://github.com/QwenLM/qwen-code/pull/9583
* fix(core): 拒绝回显提示词自身示例的会话标题 by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/9709
* fix(core): 解析 workflow 元字面量而非直接求值 by @qqqys in https://github.com/QwenLM/qwen-code/pull/9325
* fix(sdk): 支持相对工件下载 URL by @ytahdn in https://github.com/QwenLM/qwen-code/pull/9734
* fix(autofix): 为修复流程分配其可完成的预算 by @qqqys in https://github.com/QwenLM/qwen-code/pull/9691
* feat(cli): 通过设置键启用动态工作流 by @qqqys in https://github.com/QwenLM/qwen-code/pull/9098
* fix(core): 将工作流停滞看门狗的范围扩展至传输重试链之外 by @qqqys in https://github.com/QwenLM/qwen-code/pull/9397
* test(core): 确保子代理注册表不包含 MCP 指令 by @wenshao in https://github.com/QwenLM/qwen-code/pull/9720
* fix(core): 根据模型声明内容限制技能公告 by @wenshao in https://github.com/QwenLM/qwen-code/pull/9718
* feat(review): 在 Aone Code 目标上回退 PR 上下文 by @wenshao in https://github.com/QwenLM/qwen-code/pull/9621
* fix(review): 将修复引发的重新报告计为首次工作 by @wenshao in https://github.com/QwenLM/qwen-code/pull/9744
* docs(review): 当两个轴都称为“rank 3”时为其命名 by @wenshao in https://github.com/QwenLM/qwen-code/pull/9759
* fix(web-shell): 预览按文档分类的工件 by @ytahdn in https://github.com/QwenLM/qwen-code/pull/9760
* feat(web-shell): 切换后增量刷新 Composer 技能 by @samuelhsin in https://github.com/QwenLM/qwen-code/pull/9131
* feat(review): 为 Aone Code 目标回退评论状态和预提交 by @wenshao in https://github.com/QwenLM/qwen-code/pull/9627
* perf(cli): 将 VP 滚动渲染提升至 60 FPS by @DragonnZhang in https://github.com/QwenLM/qwen-code/pull/9681
* feat(models): 支持双角色图像生成模型 by @callmeYe in https://github.com/QwenLM/qwen-code/pull/9650
* fix(daemon): 加载后保持恢复的 ask_user_question 有效 by @doudouOUC in https://github.com/QwenLM/qwen-code/pull/9763
* feat(review): 添加持久关键收敛建议（允许带残留风险合并） by @wenshao in https://github.com/QwenLM/qwen-code/pull/9526
* fix(core): 支持旧版 Git 的公共 GitHub 扩展 by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/9690
* feat(cua-driver): 添加版本化的 Computer Use SDK 及发布流水线 by @LaZzyMan in https://github.com/QwenLM/qwen-code/pull/9587
* refactor(node-repl)!: 将持久化 Node REPL 作为独立 MCP 服务器交付 by @LaZzyMan in https://github.com/QwenLM/qwen-code/pull/9499
* fix(review): 在放弃清理 worktree 前修复权限 by @wenshao in https://github.com/QwenLM/qwen-code/pull/9748
* fix(core): 规范化会话标题回显比较的两端 by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/9809
* refactor(cli): 强制实施 utils 叶层依赖方向 (#9146) by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/9737
* fix(web-shell): 限制 React dev performance.measure 累积以防止渲染器内存溢出 by @wenshao in https://github.com/QwenLM/qwen-code/pull/9770
* fix(dingtalk): 解析转发聊天记录 by @qqqys in https://github.com/QwenLM/qwen-code/pull/9339
* fix(web-shell): 添加 remark-cjk-friendly 以支持 CJK 标点符号旁的粗体… by @cactuser-Lu in https://github.com/QwenLM/qwen-code/pull/9457
* feat(mcp): 添加 MCP 2026 核心及 WebShell Apps 主机 by @samuelhsin in https://github.com/QwenLM/qwen-code/pull/8992
* fix(config): 在设置模式中允许提示词钩子 by @zjunothing in https://github.com/QwenLM/qwen-code/pull/8779
* fix(web-shell): 在创建会话前显示推理力度 by @callmeYe in https://github.com/QwenLM/qwen-code/pull/9599
* fix(auth): 使 Vertex AI 能够通过应用默认凭据进行认证 by @axiom-of-choice in https://github.com/QwenLM/qwen-code/pull/9017
* fix(ci): 在流水线尺寸基线中记录 cd-cua-driver.yml 的发布尺寸 by @wenshao in https://github.com/QwenLM/qwen-code/pull/9822
* feat(review): 为验证器提供不可反驳列表及可构建的拒绝阈值 by @wenshao in https://github.com/QwenLM/qwen-code/pull/9799
* fix(core): 在等待完成回调前清空工具显示列表 by @qwen-code-dev-bot in https://github.com/QwenLM/qwen-code/pull/9602
* refactor(cli): 提取 ACP 技能管理 by @callmeYe in https://github.com/QwenLM/qwen-code/pull/8865
* fix(cli): 规范化 MCP 审批键中的 Win32 盘符大小写 by @zhou2024NAU in https://github.com/QwenLM/qwen-code/pull/9779
支持 MiniMax 图像生成模式 by @octo-patch in https://github.com/QwenLM/qwen-code/pull/8322
* fix(core): 读取前检测图像内容 by @rbalachandar in https://github.com/QwenLM/qwen-code/pull/9113
* fix(vscode): 在会话导出中保留 Windows 文件链接 by @yefuyou in https://github.com/QwenLM/qwen-code/pull/8953
* fix(cli): 防止调整大小时输入框边框溢出 by @tlysanhuo in https://github.com/QwenLM/qwen-code/pull/8991
* feat(core): 仅在 qwen serve 下声明 create_sub_session by @DragonnZhang in https://github.com/QwenLM/qwen-code/pull/9425
* fix(core): 报告剩余的后台代理 by @tao943 in https://github.com/QwenLM/qwen-code/pull/9018
* fix(core): 将推理力度层级限制在各端点接受的范围 by @harjothkhara in https://github.com/QwenLM/qwen-code/pull/9501
* fix(cli): 在开始录音时探测麦克风权限，而非语音预热阶段 by @Nas01010101 in https://github.com/QwenLM/qwen-code/pull/8912
* fix(core): 检测内容与推理流中的长段逐字重复循环 by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/9668
* feat(scheduled-tasks): 允许使用现有会话创建任务 by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/9361
* docs(agent): 澄清参数前置条件 by @tao943 in https://github.com/QwenLM/qwen-code/pull/9580
* fix(review): 在残留探测中固定已验证的 Git 身份 (#9557) by @wenshao in https://github.com/QwenLM/qwen-code/pull/9742
* fix(core): 在后台父级下展示嵌套子代理的审批 by @wenshao in https://github.com/QwenLM/qwen-code/pull/9793
* fix(core): 将代理启动失败标记为失败的 tool calls by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/9519
* refactor(goal): 通过单一核心渲染器渲染 Goal 继续提示 by @qqqys in https://github.com/QwenLM/qwen-code/pull/9581
* fix(core): 避免幽灵式关机挂起状态 by @AaronZ345 in https://github.com/QwenLM/qwen-code/pull/9550
* fix(core): 按会话隔离 fork 缓存读取器 by @tomsen02 in https://github.com/QwenLM/qwen-code/pull/9471
* refactor(review): 将 SKILL.md 拆分为核心正文及由判定结果控制的参考文件 by @wenshao in https://github.com/QwenLM/qwen-code/pull/9804
* feat(web-shell): 压缩代理活动摘要 by @ytahdn in https://github.com/QwenLM/qwen-code/pull/9657
* feat(auth): 将 Kimi (Moonshot AI) 添加为内置第三方提供商 by @TianYuan1024 in https://github.com/QwenLM/qwen-code/pull/9814
* fix(goal): 按预算单位计算目录预览数量 by @qqqys in https://github.com/QwenLM/qwen-code/pull/9835
* feat(core): 展示工作流将执行的操作，并限定批准该操作的授权范围 by @qqqys in https://github.com/QwenLM/qwen-code/pull/9837
## 新贡献者
* @qwen-code-review-bot 在 https://github.com/QwenLM/qwen-code/pull/9594 首次贡献
* @deggs7 在 https://github.com/QwenLM/qwen-code/pull/9400 首次贡献
* @cactuser-Lu 在 https://github.com/QwenLM/qwen-code/pull/9457 首次贡献
* @axiom-of-choice 在 https://github.com/QwenLM/qwen-code/pull/9017 首次贡献
* @zhou2024NAU 在 https://github.com/QwenLM/qwen-code/pull/9779 首次贡献
* @octo-patch 在 https://github.com/QwenLM/qwen-code/pull/8322 首次贡献
* @yefuyou 在 https://github.com/QwenLM/qwen-code/pull/8953 首次贡献
* @tao943 在 https://github.com/QwenLM/qwen-code/pull/9018 首次贡献
* @AaronZ345 在 https://github.com/QwenLM/qwen-code/pull/9550 首次贡献
**完整更新日志**: https://github.com/QwenLM/qwen-code/compare/dsw-eas-net-smoke-20260819-r1...cua-driver-rs-v0.20.0
Qwen CUA Driver prebuilt binaries
(vendored under `packages/cua-driver`).
- **macOS**: codesigned + notarized universal binary + `QwenCuaDriver.app`
- **Linux**: unsigned (x86_64 + arm64, glibc 2.31 floor)
- **Windows**: unsigned (x86_64 + arm64)
- **Node.js**: the same workflow publishes the single
`@qwen-code/cua-sdk` npm package after a clean install against
these release assets
Enable relative coordinates: `CUA_DRIVER_RS_COORDINATE_SPACE=1`
(default `0` = off; optional `CUA_DRIVER_RS_COORDINATE_SCALE=1000`).
Enable textual MCP payload filtering for affected model API routes:
`MCP_MODEL_PAYLOAD_FILTER=1` (default off).
<!-- Release notes generated using configuration in .github/release.yml at 5cff52c6e856e6528408863d83eb36eba6058b14 -->
## What's Changed
* feat(web-shell): approval and ask-user dialogs as in-flow sheets; fix background-agent false failure by @ytahdn in https://github.com/QwenLM/qwen-code/pull/9351
* feat(serve): Add live-state session activity watermark by @doudouOUC in https://github.com/QwenLM/qwen-code/pull/9396
* perf(web-shell): keep streaming output responsive by @ytahdn in https://github.com/QwenLM/qwen-code/pull/9405
* feat(serve): measure ACP child peak old-generation heap by @doudouOUC in https://github.com/QwenLM/qwen-code/pull/9380
* feat(review): wire --resume through /review, review run and the CI retry by @wenshao in https://github.com/QwenLM/qwen-code/pull/9153
* feat(review): record each round's posting volume in the ledger marker by @wenshao in https://github.com/QwenLM/qwen-code/pull/9413
* feat(review): teach verifiers four run disciplines from a live two-arm verification by @wenshao in https://github.com/QwenLM/qwen-code/pull/9447
* feat(daemon): add batch extension activation APIs by @callmeYe in https://github.com/QwenLM/qwen-code/pull/8788
* fix(triage): diff the flake-gate file list against the pinned base OID by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/9464
* fix(cli): surface the daemon duplicate tool-call breaker as a visible loop-detected stop by @doudouOUC in https://github.com/QwenLM/qwen-code/pull/9435
* feat(providers): add qwen3.8-max to Token Plan model list by @qqqys in https://github.com/QwenLM/qwen-code/pull/9383
* fix(cli): prevent /rewind from dropping conversation history after /compress-fast by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/9331
* fix(core): reject run_in_background: false for named teammates by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/9433
* fix(review): clamp the posting volume at its origin, not only where it is written by @wenshao in https://github.com/QwenLM/qwen-code/pull/9460
* feat(extensions): support authenticated HTTPS Git installs by @callmeYe in https://github.com/QwenLM/qwen-code/pull/9458
* fix(core): clarify list_agents excludes Agent Team teammates by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/9432
* fix(core): reject upstream fail-fast placeholder responses by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/8938
* fix(core): isolate image payload eviction state by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/9423
* fix(core): treat duplicate provider tool-call ids as replays only when arguments match by @doudouOUC in https://github.com/QwenLM/qwen-code/pull/9436
* feat(telemetry): link daemon HTTP request spans to inbound W3C traceparent by @chiga0 in https://github.com/QwenLM/qwen-code/pull/9391
* feat(web-shell): add transcript contract prevalidation by @water-in-stone in https://github.com/QwenLM/qwen-code/pull/9388
* fix(triage): compute the flake-gate diff before the env -i re-exec by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/9468
* fix(sdk): route unrecognized diagnostics onto a bounded transcript sidechannel by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/9202
* fix(ci): clone qwen-autofix into a recovery workflow entity by @wenshao in https://github.com/QwenLM/qwen-code/pull/9482
* refactor(cli): consolidate shared helpers ahead of the legacy audit skill by @wenshao in https://github.com/QwenLM/qwen-code/pull/9345
* fix(ci): no-op touch to re-register the autofix workflow triggers by @wenshao in https://github.com/QwenLM/qwen-code/pull/9479
* fix(review): keep the blocker in a COMMENT body every softening path reaches by @wenshao in https://github.com/QwenLM/qwen-code/pull/9416
* feat(web-shell): Consume live-state session activity timestamps by @doudouOUC in https://github.com/QwenLM/qwen-code/pull/9476
* refactor(review): define each certification-bar atom exactly once by @wenshao in https://github.com/QwenLM/qwen-code/pull/9473
* feat(cli): show loaded context files alongside the first prompt by @ZijianZhang989 in https://github.com/QwenLM/qwen-code/pull/8855
* feat(review): add runtime-axis, table-sweep and isolation witness forms by @wenshao in https://github.com/QwenLM/qwen-code/pull/9445
* fix(ci): keep qwen-autofix.yml under GitHub's 500 KB start-runs limit by @wenshao in https://github.com/QwenLM/qwen-code/pull/9517
* feat(cli): Add agent view PTY workers by @ZijianZhang989 in https://github.com/QwenLM/qwen-code/pull/7800
* fix(dingtalk): attach media from quoted messages by @qqqys in https://github.com/QwenLM/qwen-code/pull/9347
* feat(core): expose workflow execution state by @qqqys in https://github.com/QwenLM/qwen-code/pull/9034
* refactor(review): route the certification path atoms through one needle by @wenshao in https://github.com/QwenLM/qwen-code/pull/9484
* chore(ci): Add --provenance to npm publish and id-token permission by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/9532
* fix(cli): suppress Homebrew update notification when brew has nothing newer (#9493) by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/9502
* feat(cli): Add standalone conversation isolation primitives by @doudouOUC in https://github.com/QwenLM/qwen-code/pull/9341
* feat(web-shell): unify file uploads and references by @ytahdn in https://github.com/QwenLM/qwen-code/pull/9477
* fix(web-shell): keep sidebar sessions synchronized by @ytahdn in https://github.com/QwenLM/qwen-code/pull/9533
* refactor: centralize cross-package contracts by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/9497
* fix(cli): give the case-variant delete test the attachments mock by @wenshao in https://github.com/QwenLM/qwen-code/pull/9551
* feat(serve): persist prompt terminal ledger for cold-load reconciliation by @chiga0 in https://github.com/QwenLM/qwen-code/pull/9426
* fix(review): run verifier probes in a private scratch worktree (#9207) by @wenshao in https://github.com/QwenLM/qwen-code/pull/9221
* fix(serve): Harden standalone conversation primitives by @doudouOUC in https://github.com/QwenLM/qwen-code/pull/9512
* fix(tests): retry acp-cron cleanup rm on transient ENOTEMPTY by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/9559
* chore(ci): Drop NPM_TOKEN in favor of npm Trusted Publishing by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/9552
* fix(autofix): make the brake's BLOCKED handoff a first-class round outcome by @wenshao in https://github.com/QwenLM/qwen-code/pull/9297
* fix(web-shell): Stop repeated session title catalog refreshes by @doudouOUC in https://github.com/QwenLM/qwen-code/pull/9563
* fix(daemon): avoid custom attachment upload header by @ytahdn in https://github.com/QwenLM/qwen-code/pull/9567
* feat(review): rule on contract documentation, and matrix layered guards by @wenshao in https://github.com/QwenLM/qwen-code/pull/9448
* feat(web-shell): support file attachments in mid-turn messages by @ytahdn in https://github.com/QwenLM/qwen-code/pull/9570
* feat: register toggle-only reasoning for Qwen hybrid models by @callmeYe in https://github.com/QwenLM/qwen-code/pull/9574
* feat(web-shell): adopt canonical Goal v3 controls by @qqqys in https://github.com/QwenLM/qwen-code/pull/9393
* test(review): single-encode the diff-read fixture and pin the failed-read gate by @wenshao in https://github.com/QwenLM/qwen-code/pull/9539
* refactor(review): build the incremental scope from the PR's diff, not a check by @wenshao in https://github.com/QwenLM/qwen-code/pull/9267
* fix(ci): stop counting wedged queued runs as in-flight in the shepherd by @wenshao in https://github.com/QwenLM/qwen-code/pull/9518
* fix(ci): heal a symlinked workspace instead of wedging the runner on it by @wenshao in https://github.com/QwenLM/qwen-code/pull/9498
* fix(autofix): mutation-probe new guards before a round commits by @wenshao in https://github.com/QwenLM/qwen-code/pull/9578
* test(ci): stage on-disk session state in the serve A/B by @wenshao in https://github.com/QwenLM/qwen-code/pull/9444
* fix(ci): make autofix finding replies idempotent by @wenshao in https://github.com/QwenLM/qwen-code/pull/9463
* fix(ci): author the release PR with a third bot PAT by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/9592
* fix(cli): persist skipped duplicate tool results by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/9593
* fix(web-shell): fall back to execCommand copy in non-secure contexts by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/9540
* fix(ui): collapse duplicate in-flight tool_group rendered from history + pending by @qwen-code-dev-bot in https://github.com/QwenLM/qwen-code/pull/9421
* feat(review): post --comment reviews to Aone Code via the a1 CLI by @wenshao in https://github.com/QwenLM/qwen-code/pull/9491
* feat(review): tell the author why a review loop is not settling by @wenshao in https://github.com/QwenLM/qwen-code/pull/9461
* fix(ci): stop the fallback comment from denying a review it already posted by @wenshao in https://github.com/QwenLM/qwen-code/pull/9462
* fix(web-shell): bound daemon transcript retention to stop renderer OOM crashes by @wenshao in https://github.com/QwenLM/qwen-code/pull/9303
* chore(ci): Disable install scripts in release CI and guard security-checks workflow by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/9577
* feat(web-shell): keep a turn expanded while its background shell runs by @ytahdn in https://github.com/QwenLM/qwen-code/pull/9632
* feat(autofix): audit the approach instead of stopping on growth-budget breach by @wenshao in https://github.com/QwenLM/qwen-code/pull/9262
* docs(ci): the ECS pool does run containers — correct two comments that say it does not by @wenshao in https://github.com/QwenLM/qwen-code/pull/9575
* fix(autofix): bind the sandbox image to its pulled digest by @wenshao in https://github.com/QwenLM/qwen-code/pull/9527
* docs(autofix): design runner-level isolation for PAT-bearing steps by @wenshao in https://github.com/QwenLM/qwen-code/pull/9525
* fix(web-shell): settle parallel-agents collapse and unify agent detail transcript by @ytahdn in https://github.com/QwenLM/qwen-code/pull/9640
* fix(web-shell): show create-group tooltip by @deggs7 in https://github.com/QwenLM/qwen-code/pull/9400
* feat(review): fold the one-hop import widening into `fetch-pr --since` by @wenshao in https://github.com/QwenLM/qwen-code/pull/9332
* docs: document inline terminal image previews by @DragonnZhang in https://github.com/QwenLM/qwen-code/pull/8656
* feat(core): make list_directory opt-in (disabled by default) by @DragonnZhang in https://github.com/QwenLM/qwen-code/pull/9424
* docs: document the Session Workflow setting by @DragonnZhang in https://github.com/QwenLM/qwen-code/pull/8554
* feat(cli): extend non-blocking slash commands to more builtins by @DragonnZhang in https://github.com/QwenLM/qwen-code/pull/9495
* docs(autofix): pin the publish-side checks to the patch text by @wenshao in https://github.com/QwenLM/qwen-code/pull/9652
* chore(deps): Clear high-severity CVE baseline and harden the security gate by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/9584
* docs(autofix): add an operator guide for `/takeover from N` by @wenshao in https://github.com/QwenLM/qwen-code/pull/9622
* fix(artifacts): expand recorded directories into per-file artifacts by @zjgzx1988 in https://github.com/QwenLM/qwen-code/pull/9395
* feat(review): disclose that Aone posts join the discussion gate only by @wenshao in https://github.com/QwenLM/qwen-code/pull/9625
* fix(review): make the incremental cache work for Aone AGit-Flow CRs by @wenshao in https://github.com/QwenLM/qwen-code/pull/9630
* feat(review): detect self-MR on Aone targets in presubmit by @wenshao in https://github.com/QwenLM/qwen-code/pull/9629
* fix(autofix): include pending runs in the busy-PR enumeration by @wenshao in https://github.com/QwenLM/qwen-code/pull/9662
* docs(autofix): correct the round-seed guide on leading whitespace by @wenshao in https://github.com/QwenLM/qwen-code/pull/9663
* fix(web-shell): use daemon live-state hasActivePrompt for loading indicator (#9487) by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/9631
* fix(cli): keep slash menu selection stable while a response streams by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/9508
* fix(web-shell): don't steal approval focus while the user is typing by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/9609
* fix(core): support public GitHub extensions with older Git by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/9680
* feat(web-shell): Bind GitHub PRs to sessions with sidebar badge and search by @wenshao in https://github.com/QwenLM/qwen-code/pull/9543
* fix(ci): bump vulnerable dependencies to unblock CVE audit by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/9703
* fix(review): audit Aone targets in cleanup's bypass tripwire by @wenshao in https://github.com/QwenLM/qwen-code/pull/9633
* docs: classify architecture invariants by enforcement mechanism (#9152) by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/9689
* fix(autofix): stop counting idle timeouts toward the timeout cap by @wenshao in https://github.com/QwenLM/qwen-code/pull/9673
* feat(review): give the convergence observation a machine-readable half by @wenshao in https://github.com/QwenLM/qwen-code/pull/9623
* perf(web-shell): optimize streaming transcript rendering by @ytahdn in https://github.com/QwenLM/qwen-code/pull/9672
* fix(web-shell): route goal messages by session activity by @ytahdn in https://github.com/QwenLM/qwen-code/pull/9667
* i18n(cli): translate Unset in the settings dialog by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/9714
* perf(review): give review agents their own subagent type by @wenshao in https://github.com/QwenLM/qwen-code/pull/9678
* refactor(autofix): move the push-and-report body out of the workflow file by @wenshao in https://github.com/QwenLM/qwen-code/pull/9653
* feat(review): report the address a drive's service actually bound by @wenshao in https://github.com/QwenLM/qwen-code/pull/9655
* fix(autofix): pass CI=true through the gate's env -i launches by @wenshao in https://github.com/QwenLM/qwen-code/pull/9649
* refactor(cli): keep acp-integration off serve internals (#8084) by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/9144
* feat(review): close Aone residual gaps — composeUrl, test-plan routing, a1 version floor by @wenshao in https://github.com/QwenLM/qwen-code/pull/9624
* fix(review): stop shedding the convergence observation first by @wenshao in https://github.com/QwenLM/qwen-code/pull/9715
* fix(review): clear the deferred Round-5 findings from the Aone write path by @wenshao in https://github.com/QwenLM/qwen-code/pull/9604
* refactor(core): remove root barrel self-imports and enforce the boundary by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/9635
* fix(cli): Recover sessions across archive races by @doudouOUC in https://github.com/QwenLM/qwen-code/pull/9513
* feat(serve): restore ask_user_question HITL on session load/resume by @doudouOUC in https://github.com/QwenLM/qwen-code/pull/9665
* fix(sdk): support "auto" permission mode by @shenyankm in https://github.com/QwenLM/qwen-code/pull/9003
* fix(web-shell): pass session workspace cwd when opening from overview panel by @wenshao in https://github.com/QwenLM/qwen-code/pull/9730
* fix(web-shell): fall back after stale mid-turn rejection by @ytahdn in https://github.com/QwenLM/qwen-code/pull/9732
* fix(review): stop the fallback comment from firing on superseded runs by @wenshao in https://github.com/QwenLM/qwen-code/pull/9716
* feat(review): add temporal-reachability and incident-replay lenses by @wenshao in https://github.com/QwenLM/qwen-code/pull/9708
* fix(web-shell): stable pinned-section order and instant pin feedback by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/9560
* feat(review): ask each fix for its test, and rule on non-convergence by @wenshao in https://github.com/QwenLM/qwen-code/pull/9596
* chore(ci): migrate autofix prose to the design record and ratchet growth by @wenshao in https://github.com/QwenLM/qwen-code/pull/9677
* feat(cli): restore each daemon session onto its last selected model by @doudouOUC in https://github.com/QwenLM/qwen-code/pull/9687
* feat(review): validate Aone inline anchors against the captured diff before posting by @wenshao in https://github.com/QwenLM/qwen-code/pull/9634
* fix(ci): record qwen-autofix.yml's shipped size in the workflow size baseline by @wenshao in https://github.com/QwenLM/qwen-code/pull/9747
* fix(ci): gate the fork signal on fields the review payload delivers by @qqqys in https://github.com/QwenLM/qwen-code/pull/9469
* feat(goal): account the tokens a Goal spends by @qqqys in https://github.com/QwenLM/qwen-code/pull/9583
* fix(core): reject session titles that echo the prompt's own examples by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/9709
* fix(core): parse the workflow meta literal instead of evaluating it by @qqqys in https://github.com/QwenLM/qwen-code/pull/9325
* fix(sdk): support relative artifact download URLs by @ytahdn in https://github.com/QwenLM/qwen-code/pull/9734
* fix(autofix): give the repair pass a budget it can finish in by @qqqys in https://github.com/QwenLM/qwen-code/pull/9691
* feat(cli): enable dynamic workflows from a settings key by @qqqys in https://github.com/QwenLM/qwen-code/pull/9098
* fix(core): widen the workflow stall watchdog past the transport retry ladder by @qqqys in https://github.com/QwenLM/qwen-code/pull/9397
* test(core): pin that subagent registries carry no MCP instructions by @wenshao in https://github.com/QwenLM/qwen-code/pull/9720
* fix(core): gate skill announcements on what the model was declared by @wenshao in https://github.com/QwenLM/qwen-code/pull/9718
* feat(review): back pr-context on Aone Code targets by @wenshao in https://github.com/QwenLM/qwen-code/pull/9621
* fix(review): count a fix-induced re-report as first-time work by @wenshao in https://github.com/QwenLM/qwen-code/pull/9744
* docs(review): name the axis when two of them both call it "rank 3" by @wenshao in https://github.com/QwenLM/qwen-code/pull/9759
* fix(web-shell): preview document-classified artifacts by @ytahdn in https://github.com/QwenLM/qwen-code/pull/9760
* feat(web-shell): refresh composer skills incrementally after toggles by @samuelhsin in https://github.com/QwenLM/qwen-code/pull/9131
* feat(review): back comment-status and presubmit for Aone Code targets by @wenshao in https://github.com/QwenLM/qwen-code/pull/9627
* perf(cli): raise VP scroll rendering to 60 FPS by @DragonnZhang in https://github.com/QwenLM/qwen-code/pull/9681
* feat(models): support dual-role image generation models by @callmeYe in https://github.com/QwenLM/qwen-code/pull/9650
* fix(daemon): keep restored ask_user_question valid after load by @doudouOUC in https://github.com/QwenLM/qwen-code/pull/9763
* feat(review): add the persistently-critical convergence advisory (land-with-residual-risk) by @wenshao in https://github.com/QwenLM/qwen-code/pull/9526
* fix(core): support public GitHub extensions with older Git by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/9690
* feat(cua-driver): add versioned Computer Use SDK and release pipeline by @LaZzyMan in https://github.com/QwenLM/qwen-code/pull/9587
* refactor(node-repl)!: deliver the persistent Node REPL as a standalone MCP server by @LaZzyMan in https://github.com/QwenLM/qwen-code/pull/9499
* fix(review): repair permissions before giving up on worktree cleanup by @wenshao in https://github.com/QwenLM/qwen-code/pull/9748
* fix(core): Normalize both sides of the session title echo comparison by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/9809
* refactor(cli): enforce utils leaf-layer dependency direction (#9146) by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/9737
* fix(web-shell): cap React dev performance.measure accumulation to stop renderer OOM by @wenshao in https://github.com/QwenLM/qwen-code/pull/9770
* fix(dingtalk): parse forwarded chat records by @qqqys in https://github.com/QwenLM/qwen-code/pull/9339
* fix(web-shell): add remark-cjk-friendly so bold adjacent to CJK punct… by @cactuser-Lu in https://github.com/QwenLM/qwen-code/pull/9457
* feat(mcp): add MCP 2026 core and WebShell Apps host by @samuelhsin in https://github.com/QwenLM/qwen-code/pull/8992
* fix(config): allow prompt hooks in settings schema by @zjunothing in https://github.com/QwenLM/qwen-code/pull/8779
* fix(web-shell): show reasoning effort before session creation by @callmeYe in https://github.com/QwenLM/qwen-code/pull/9599
* fix(auth): let Vertex AI authenticate with Application Default Credentials by @axiom-of-choice in https://github.com/QwenLM/qwen-code/pull/9017
* fix(ci): record cd-cua-driver.yml's shipped size in the workflow size baseline by @wenshao in https://github.com/QwenLM/qwen-code/pull/9822
* feat(review): give verifiers a do-not-refute list and a constructible rejection bar by @wenshao in https://github.com/QwenLM/qwen-code/pull/9799
* fix(core): clear tool display list before awaiting completion callback by @qwen-code-dev-bot in https://github.com/QwenLM/qwen-code/pull/9602
* refactor(cli): extract ACP skill management by @callmeYe in https://github.com/QwenLM/qwen-code/pull/8865
* fix(cli): normalize win32 drive-letter casing in MCP approval keys by @zhou2024NAU in https://github.com/QwenLM/qwen-code/pull/9779
* Support MiniMax image generation schema by @octo-patch in https://github.com/QwenLM/qwen-code/pull/8322
* fix(core): sniff image content before read by @rbalachandar in https://github.com/QwenLM/qwen-code/pull/9113
* fix(vscode): preserve Windows file links in session exports by @yefuyou in https://github.com/QwenLM/qwen-code/pull/8953
* fix(cli): prevent input border overflow on resize by @tlysanhuo in https://github.com/QwenLM/qwen-code/pull/8991
* feat(core): declare create_sub_session only under qwen serve by @DragonnZhang in https://github.com/QwenLM/qwen-code/pull/9425
* fix(core): report remaining background agents by @tao943 in https://github.com/QwenLM/qwen-code/pull/9018
* fix(core): cap the effort tier at what each endpoint accepts by @harjothkhara in https://github.com/QwenLM/qwen-code/pull/9501
* fix(cli): probe microphone permission on recording start, not voice warmup by @Nas01010101 in https://github.com/QwenLM/qwen-code/pull/8912
* fix(core): detect long verbatim repetition loops in content and reasoning streams by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/9668
* feat(scheduled-tasks): allow creating a task with an existing session by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/9361
* docs(agent): clarify parameter preconditions by @tao943 in https://github.com/QwenLM/qwen-code/pull/9580
* fix(review): pin the verified git identity across the residue probe (#9557) by @wenshao in https://github.com/QwenLM/qwen-code/pull/9742
* fix(core): surface nested sub-agent approvals under background parents by @wenshao in https://github.com/QwenLM/qwen-code/pull/9793
* fix(core): mark agent launch failures as failed tool calls by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/9519
* refactor(goal): render Goal continuation prompts from one core renderer by @qqqys in https://github.com/QwenLM/qwen-code/pull/9581
* fix(core): avoid ghost shutdown pending state by @AaronZ345 in https://github.com/QwenLM/qwen-code/pull/9550
* fix(core): isolate fork cache readers by session by @tomsen02 in https://github.com/QwenLM/qwen-code/pull/9471
* refactor(review): split SKILL.md into a core body plus verdict-gated reference files by @wenshao in https://github.com/QwenLM/qwen-code/pull/9804
* feat(web-shell): compact agent activity summaries by @ytahdn in https://github.com/QwenLM/qwen-code/pull/9657
* feat(auth): add Kimi (Moonshot AI) as a built-in third-party provider by @TianYuan1024 in https://github.com/QwenLM/qwen-code/pull/9814
* fix(goal): count catalog previews in the unit their budget is written in by @qqqys in https://github.com/QwenLM/qwen-code/pull/9835
* feat(core): show what a workflow will do, and scope the grant that approves it by @qqqys in https://github.com/QwenLM/qwen-code/pull/9837
## New Contributors
* @qwen-code-review-bot made their first contribution in https://github.com/QwenLM/qwen-code/pull/9594
* @deggs7 made their first contribution in https://github.com/QwenLM/qwen-code/pull/9400
* @cactuser-Lu made their first contribution in https://github.com/QwenLM/qwen-code/pull/9457
* @axiom-of-choice made their first contribution in https://github.com/QwenLM/qwen-code/pull/9017
* @zhou2024NAU made their first contribution in https://github.com/QwenLM/qwen-code/pull/9779
* @octo-patch made their first contribution in https://github.com/QwenLM/qwen-code/pull/8322
* @yefuyou made their first contribution in https://github.com/QwenLM/qwen-code/pull/8953
* @tao943 made their first contribution in https://github.com/QwenLM/qwen-code/pull/9018
* @AaronZ345 made their first contribution in https://github.com/QwenLM/qwen-code/pull/9550
**Full Changelog**: https://github.com/QwenLM/qwen-code/compare/dsw-eas-net-smoke-20260819-r1...cua-driver-rs-v0.20.0