## 更新内容
* feat(acp): 为 ACP 集成添加完整的 hooks 支持 by @DennisYu07 in https://github.com/QwenLM/qwen-code/pull/3248
* feat: 优化紧凑模式用户体验 — 快捷键、设置同步与安全性 by @chiga0 in https://github.com/QwenLM/qwen-code/pull/3100
* feat(hooks): 添加 HTTP Hook、Function Hook 和 Async Hook 支持 by @DennisYu07 in https://github.com/QwenLM/qwen-code/pull/2827
* fix(sdk): 避免在 ProcessTransport 中泄漏进程退出监听器 by @reidliu41 in https://github.com/QwenLM/qwen-code/pull/3295
* fix(cli): 防止 statusline spawn EBADF 导致 CLI 崩溃 (#3264) by @wenshao in https://github.com/QwenLM/qwen-code/pull/3310
* fix(cli): 在摘要更改前保持“开始新聊天会话”状态 by @reidliu41 in https://github.com/QwenLM/qwen-code/pull/3308
* feat(memory): 管理自动记忆与自动梦境系统 by @LaZzyMan in https://github.com/QwenLM/qwen-code/pull/3087
* fix(cli): 延迟更新通知直至模型响应完成 by @pic4xiu in https://github.com/QwenLM/qwen-code/pull/3321
* test(core): 稳定化 glob 截断测试 by @reidliu41 in https://github.com/QwenLM/qwen-code/pull/3322
* docs: 修复 Windows 安装命令以同时兼容 CMD 和 PowerShell by @Skyline-9 in https://github.com/QwenLM/qwen-code/pull/3252
* feat(cli): 支持多行状态栏输出 by @wenshao in https://github.com/QwenLM/qwen-code/pull/3311
* fix(core): 限制 skill watcher 深度以防止文件描述符耗尽 by @tanzhenxin in https://github.com/QwenLM/qwen-code/pull/3320
* fix(core): 切换模型时从历史记录中移除思考块 by @tanzhenxin in https://github.com/QwenLM/qwen-code/pull/3315
* docs: 更新认证方法以反映 OAuth 停用 by @joeytoday in https://github.com/QwenLM/qwen-code/pull/3325
* fix(core): 添加 shell 参数引号使用指南以防止特殊字符错误 by @LaZzyMan in https://github.com/QwenLM/qwen-code/pull/3327
* refactor(core): 将 fork 子代理参数从 execute() 移至构造阶段 by @tanzhenxin in https://github.com/QwenLM/qwen-code/pull/3255
* feat(skills): 添加用于并行批量操作的 /batch 技能 by @doudouOUC in https://github.com/QwenLM/qwen-code/pull/3079
* feat: 新增后台子代理，支持无头模式与 SDK by @tanzhenxin in https://github.com/QwenLM/qwen-code/pull/3076
* feat(core): 从 .qwen/rules/ 添加基于路径的上下文规则注入 by @wenshao in https://github.com/QwenLM/qwen-code/pull/3339
* test(integration): 在交互测试中匹配新的 cron 通知格式 by @tanzhenxin in https://github.com/QwenLM/qwen-code/pull/3402
* feat(cli): 为 TUI 添加双输出 sidecar 模式 by @chiga0 in https://github.com/QwenLM/qwen-code/pull/3352
* feat: 将 `M-d` 绑定至合理的默认值（类 Emacs 风格） by @pedrormjunior in https://github.com/QwenLM/qwen-code/pull/3358
* fix(cli): 减少终端重绘时的光标移动 by @reidliu41 in https://github.com/QwenLM/qwen-code/pull/3381
## 新贡献者
* @Skyline-9 在 https://github.com/QwenLM/qwen-code/pull/3252 中完成了首次贡献
* @pedrormjunior 在 https://github.com/QwenLM/qwen-code/pull/3358 中完成了首次贡献
**完整更新日志**：https://github.com/QwenLM/qwen-code/compare/v0.14.5...v0.14.5-nightly.20260418.418acc548
## What's Changed
* feat(acp): add complete hooks support for ACP integration by @DennisYu07 in https://github.com/QwenLM/qwen-code/pull/3248
* feat: optimize compact mode UX — shortcuts, settings sync, and safety by @chiga0 in https://github.com/QwenLM/qwen-code/pull/3100
* feat(hooks): Add HTTP Hook, Function Hook and Async Hook support by @DennisYu07 in https://github.com/QwenLM/qwen-code/pull/2827
* fix(sdk): avoid leaking process exit listeners in ProcessTransport by @reidliu41 in https://github.com/QwenLM/qwen-code/pull/3295
* fix(cli): prevent statusline spawn EBADF from crashing CLI (#3264) by @wenshao in https://github.com/QwenLM/qwen-code/pull/3310
* fix(cli): remember "Start new chat session" until summary changes by @reidliu41 in https://github.com/QwenLM/qwen-code/pull/3308
* feat(memory): managed auto-memory and auto-dream system by @LaZzyMan in https://github.com/QwenLM/qwen-code/pull/3087
* fix(cli): defer update notifications until model response completes by @pic4xiu in https://github.com/QwenLM/qwen-code/pull/3321
* test(core): stabilize glob truncation tests by @reidliu41 in https://github.com/QwenLM/qwen-code/pull/3322
* docs: fix Windows install command to work in both CMD and PowerShell by @Skyline-9 in https://github.com/QwenLM/qwen-code/pull/3252
* feat(cli): support multi-line status line output by @wenshao in https://github.com/QwenLM/qwen-code/pull/3311
* fix(core): limit skill watcher depth to prevent FD exhaustion by @tanzhenxin in https://github.com/QwenLM/qwen-code/pull/3320
* fix(core): strip thinking blocks from history on model switch by @tanzhenxin in https://github.com/QwenLM/qwen-code/pull/3315
* docs: update authentication methods to reflect OAuth discontinuation by @joeytoday in https://github.com/QwenLM/qwen-code/pull/3325
* fix(core): add shell argument quoting guidance to prevent special char errors by @LaZzyMan in https://github.com/QwenLM/qwen-code/pull/3327
* refactor(core): move fork subagent params from execute() to construction time by @tanzhenxin in https://github.com/QwenLM/qwen-code/pull/3255
* feat(skills): add /batch skill for parallel batch operations by @doudouOUC in https://github.com/QwenLM/qwen-code/pull/3079
* feat: background subagents with headless and SDK support by @tanzhenxin in https://github.com/QwenLM/qwen-code/pull/3076
* feat(core): add path-based context rule injection from .qwen/rules/ by @wenshao in https://github.com/QwenLM/qwen-code/pull/3339
* test(integration): match new cron notification format in interactive tests by @tanzhenxin in https://github.com/QwenLM/qwen-code/pull/3402
* feat(cli): add dual-output sidecar mode for TUI by @chiga0 in https://github.com/QwenLM/qwen-code/pull/3352
* feat: bind `M-d` to a reasonable (Emacs-like) default by @pedrormjunior in https://github.com/QwenLM/qwen-code/pull/3358
* fix(cli): reduce terminal redraw cursor movement by @reidliu41 in https://github.com/QwenLM/qwen-code/pull/3381
## New Contributors
* @Skyline-9 made their first contribution in https://github.com/QwenLM/qwen-code/pull/3252
* @pedrormjunior made their first contribution in https://github.com/QwenLM/qwen-code/pull/3358
**Full Changelog**: https://github.com/QwenLM/qwen-code/compare/v0.14.5...v0.14.5-nightly.20260418.418acc548