## 更新内容
* feat(skills): 添加内置的 /review 技能，提供开箱即用的代码审查功能 by @wenshao in https://github.com/QwenLM/qwen-code/pull/2348
* fix(insight): 处理定性洞察中单个 LLM 的失败情况 (#2341) by @netbrah in https://github.com/QwenLM/qwen-code/pull/2361
* fix(core): 将 deepseek-r1 添加到输出令牌限制模式 by @netbrah in https://github.com/QwenLM/qwen-code/pull/2362
* chore(release): 将版本提升至 0.12.4 by @qwen-code-ci-bot in https://github.com/QwenLM/qwen-code/pull/2393
* fix(i18n): 本地化斜杠命令描述 by @chen893 in https://github.com/QwenLM/qwen-code/pull/2333
* fix(core): 在 convertOpenAIResponseToGemini 中防范空选项 by @netbrah in https://github.com/QwenLM/qwen-code/pull/2364
* feat(skills): 添加文档审计和更新辅助工具 by @DragonnZhang in https://github.com/QwenLM/qwen-code/pull/2397
* 文档：子代理系统提示词限制 by @hs-ye in https://github.com/QwenLM/qwen-code/pull/2001
* fix(extension): 在 Windows 上进行 git clone 时禁用符号链接，以修复安装失败 by @LaZzyMan in https://github.com/QwenLM/qwen-code/pull/2286
* fix(core): 拒绝 PDF 文件以防止会话损坏（修复 #2020） by @drewd789 in https://github.com/QwenLM/qwen-code/pull/2024
* 在计划模式下保持被拒绝的计划内容可见 by @DragonnZhang in https://github.com/QwenLM/qwen-code/pull/2157
* fix(cli): 在 macOS 宽松沙箱中允许 /dev/ptmx 和 /dev/ttys* by @kkhomej33-netizen in https://github.com/QwenLM/qwen-code/pull/2391
* fix: 修正 hooks JSON schema 类型定义 by @xuewenjie123 in https://github.com/QwenLM/qwen-code/pull/2280
* fix(core): 在重试前移除孤立的用户条目以防止 API 错误 by @tanzhenxin in https://github.com/QwenLM/qwen-code/pull/2367
* fix(core): 在交互式 shell 模式下正确捕获快速的 pty 输出 by @tanzhenxin in https://github.com/QwenLM/qwen-code/pull/2389
* fix(vscode): 防止提示词取消和流式传输中的竞态条件 by @tanzhenxin in https://github.com/QwenLM/qwen-code/pull/2374
* fix(core): 改进 shell 工具截断，简化工具输出处理，并移除摘要功能 by @tanzhenxin in https://github.com/QwenLM/qwen-code/pull/2388
* fix:移除冗余的计划文件 by @LaZzyMan in https://github.com/QwenLM/qwen-code/pull/2407
* chore(CODEOWNERS): 移除 vscode-ide-companion 和 webui 包的必需审查者 by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/2408
* fix(core): 规范化用于 shell 执行的 Windows PATH 风格环境变量键 by @Sakuranda in https://github.com/QwenLM/qwen-code/pull/1904
* docs: 解释 Docker 沙箱运行时和 Java 用法 by @kiri-chenchen in https://github.com/QwenLM/qwen-code/pull/1642
* docs(integration): 为 Zed 和 JetBrains 集成文档添加 ACP Registry by @ShihaoShenDev in https://github.com/QwenLM/qwen-code/pull/2372
* 将 DEFAULT_OUTPUT_TOKEN_LIMIT 从 8K 增加到 16K by @Mingholy in https://github.com/QwenLM/qwen-code/pull/2411
* fix: 当 provider 未设置时，从模型自动检测 max_tokens by @netbrah in https://github.com/QwenLM/qwen-code/pull/2356
## 新贡献者
* @netbrah 在 https://github.com/QwenLM/qwen-code/pull/2361 中首次贡献
* @chen893 在 https://github.com/QwenLM/qwen-code/pull/2333 中首次贡献
* @hs-ye 在 https://github.com/QwenLM/qwen-code/pull/2001 中首次贡献
* @drewd789 在 https://github.com/QwenLM/qwen-code/pull/2024 中首次贡献
* @Sakuranda 在 https://github.com/QwenLM/qwen-code/pull/1904 中首次贡献
* @kiri-chenchen 在 https://github.com/QwenLM/qwen-code/pull/1642 中首次贡献
* @ShihaoShenDev 在 https://github.com/QwenLM/qwen-code/pull/2372 中首次贡献
**完整更新日志**: https://github.com/QwenLM/qwen-code/compare/v0.12.3...v0.12.4
## What's Changed
* feat(skills): add bundled /review skill for out-of-the-box code review by @wenshao in https://github.com/QwenLM/qwen-code/pull/2348
* fix(insight): handle individual LLM failures in qualitative insights (#2341) by @netbrah in https://github.com/QwenLM/qwen-code/pull/2361
* fix(core): add deepseek-r1 to output token limit patterns by @netbrah in https://github.com/QwenLM/qwen-code/pull/2362
* chore(release): bump version to 0.12.4 by @qwen-code-ci-bot in https://github.com/QwenLM/qwen-code/pull/2393
* fix(i18n): localize slash command descriptions by @chen893 in https://github.com/QwenLM/qwen-code/pull/2333
* fix(core): guard against empty choices in convertOpenAIResponseToGemini by @netbrah in https://github.com/QwenLM/qwen-code/pull/2364
* feat(skills): add docs audit and update helpers by @DragonnZhang in https://github.com/QwenLM/qwen-code/pull/2397
* Docs/subagent system prompt limits by @hs-ye in https://github.com/QwenLM/qwen-code/pull/2001
* fix(extension): disable symlinks on Windows during git clone to fix install failure by @LaZzyMan in https://github.com/QwenLM/qwen-code/pull/2286
* fix(core): reject PDF files to prevent session corruption (fixes #2020) by @drewd789 in https://github.com/QwenLM/qwen-code/pull/2024
* Keep rejected plan content visible in plan mode by @DragonnZhang in https://github.com/QwenLM/qwen-code/pull/2157
* fix(cli): allow /dev/ptmx and /dev/ttys* in macOS permissive sandbox by @kkhomej33-netizen in https://github.com/QwenLM/qwen-code/pull/2391
* fix: correct hooks JSON schema type definition by @xuewenjie123 in https://github.com/QwenLM/qwen-code/pull/2280
* fix(core): strip orphaned user entries before retry to prevent API errors by @tanzhenxin in https://github.com/QwenLM/qwen-code/pull/2367
* fix(core): correctly capture rapid pty outputs in interactive shell mode by @tanzhenxin in https://github.com/QwenLM/qwen-code/pull/2389
* fix(vscode): prevent race conditions in prompt cancellation and streaming by @tanzhenxin in https://github.com/QwenLM/qwen-code/pull/2374
* fix(core): improve shell tool truncation, simplify tool output handling, and remove summarization by @tanzhenxin in https://github.com/QwenLM/qwen-code/pull/2388
* fix:remove redundant plan files by @LaZzyMan in https://github.com/QwenLM/qwen-code/pull/2407
* chore(CODEOWNERS): remove required reviewers for vscode-ide-companion and webui packages by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/2408
* fix(core): normalize Windows PATH-like env keys for shell execution by @Sakuranda in https://github.com/QwenLM/qwen-code/pull/1904
* docs: explain Docker sandbox runtime and Java usage by @kiri-chenchen in https://github.com/QwenLM/qwen-code/pull/1642
* docs(integration): add ACP Registry for Zed and JetBrains integration docs by @ShihaoShenDev in https://github.com/QwenLM/qwen-code/pull/2372
* Increase DEFAULT_OUTPUT_TOKEN_LIMIT from 8K to 16K by @Mingholy in https://github.com/QwenLM/qwen-code/pull/2411
* fix: auto-detect max_tokens from model when not set by provider by @netbrah in https://github.com/QwenLM/qwen-code/pull/2356
## New Contributors
* @netbrah made their first contribution in https://github.com/QwenLM/qwen-code/pull/2361
* @chen893 made their first contribution in https://github.com/QwenLM/qwen-code/pull/2333
* @hs-ye made their first contribution in https://github.com/QwenLM/qwen-code/pull/2001
* @drewd789 made their first contribution in https://github.com/QwenLM/qwen-code/pull/2024
* @Sakuranda made their first contribution in https://github.com/QwenLM/qwen-code/pull/1904
* @kiri-chenchen made their first contribution in https://github.com/QwenLM/qwen-code/pull/1642
* @ShihaoShenDev made their first contribution in https://github.com/QwenLM/qwen-code/pull/2372
**Full Changelog**: https://github.com/QwenLM/qwen-code/compare/v0.12.3...v0.12.4