## 更新内容
* chore: 将版本升级至 0.13.0 by @qwen-code-ci-bot in https://github.com/QwenLM/qwen-code/pull/2451
* fix(pipeline): 处理来自 OpenRouter 的重复 finish_reason 块 by @simon100500 in https://github.com/QwenLM/qwen-code/pull/2403
* feat: 在 SDK 和 CLI 中添加系统提示词自定义选项 by @DragonnZhang in https://github.com/QwenLM/qwen-code/pull/2400
* fix(cli): 在 Ctrl+R 命令搜索中显示最新优先的历史记录 by @kkhomej33-netizen in https://github.com/QwenLM/qwen-code/pull/2425
* feat(hooks): 实现 hooks 扩展机制 by @DennisYu07 in https://github.com/QwenLM/qwen-code/pull/2352
* Fix: 确保 SDK 流式传输中 message_start 和 message_stop 事件成对出现 by @Mingholy in https://github.com/QwenLM/qwen-code/pull/2448
* fix(core): 为 MCP 工具输出添加截断支持 by @tanzhenxin in https://github.com/QwenLM/qwen-code/pull/2446
* feat(core): 并发执行任务工具以提升性能 by @tanzhenxin in https://github.com/QwenLM/qwen-code/pull/2434
* refactor(core): 改进错误处理和配额检测 by @tanzhenxin in https://github.com/QwenLM/qwen-code/pull/2458
* feat(arena): 新增支持多模型竞争执行的智能体协作竞技场 by @tanzhenxin in https://github.com/QwenLM/qwen-code/pull/1912
* refactor: 重构 VS Code 文件补全系统以使用模糊搜索 by @qqqys in https://github.com/QwenLM/qwen-code/pull/2437
* feat(ui): 在加载/进度指示器中显示 Token 使用情况 by @qqqys in https://github.com/QwenLM/qwen-code/pull/2445
* feat(vscode-ide-companion): 为补全功能添加仅按 Tab 键填充的行为 by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/2431
* fix(vscode-ide-companion): 更新 Windows 路径的 URI 处理 by @qqqys in https://github.com/QwenLM/qwen-code/pull/2457
* feat: 新增 /context 命令以显示上下文窗口 Token 使用明细 by @pomelo-nwu in https://github.com/QwenLM/qwen-code/pull/1835
* fix(test): 更新 LoadingIndicator 快照以确保输出对齐正确 by @qqqys in https://github.com/QwenLM/qwen-code/pull/2469
* feat: 支持在 .agents 目录及其他提供程序目录中使用技能 by @LaZzyMan in https://github.com/QwenLM/qwen-code/pull/2202
* feat: 新增 `auth` CLI 命令及 Qwen Code Claw 技能 by @LaZzyMan in https://github.com/QwenLM/qwen-code/pull/2440
* feat(export): 添加元数据和统计信息追踪 by @Mingholy in https://github.com/QwenLM/qwen-code/pull/2328
* fix: 修正 MiniMax-M2.5 和 GLM 模型的 Token 限制 by @Mingholy in https://github.com/QwenLM/qwen-code/pull/2470
* feat(hooks): 为会话生命周期和工具执行实现 10 个核心事件钩子 by @DennisYu07 in https://github.com/QwenLM/qwen-code/pull/2203
* fix: 更新 VS Code 扩展 README 中的 TOS 链接 by @Br1an67 in https://github.com/QwenLM/qwen-code/pull/2495
* fix: 在 OpenAI 日志请求转换期间保留模态信息 by @kkhomej33-netizen in https://github.com/QwenLM/qwen-code/pull/2473
* Feat: 支持权限控制 by @LaZzyMan in https://github.com/QwenLM/qwen-code/pull/2283
## 新贡献者
* @simon100500 在 https://github.com/QwenLM/qwen-code/pull/2403 中首次贡献
* @Br1an67 在 https://github.com/QwenLM/qwen-code/pull/2495 中首次贡献
**完整更新日志**: https://github.com/QwenLM/qwen-code/compare/v0.12.6...v0.13.0-nightly.20260320.b8dda154d
## What's Changed
* chore: bump version to 0.13.0 by @qwen-code-ci-bot in https://github.com/QwenLM/qwen-code/pull/2451
* fix(pipeline): handle duplicate finish_reason chunks from OpenRouter by @simon100500 in https://github.com/QwenLM/qwen-code/pull/2403
* feat: add system prompt customization options in SDK and CLI by @DragonnZhang in https://github.com/QwenLM/qwen-code/pull/2400
* fix(cli): show newest-first history for Ctrl+R command search by @kkhomej33-netizen in https://github.com/QwenLM/qwen-code/pull/2425
* feat(hooks): implement hooks extension mechanism by @DennisYu07 in https://github.com/QwenLM/qwen-code/pull/2352
* Fix: Ensure message_start and message_stop events are paired in SDK streaming by @Mingholy in https://github.com/QwenLM/qwen-code/pull/2448
* fix(core): add truncation support for MCP tool output by @tanzhenxin in https://github.com/QwenLM/qwen-code/pull/2446
* feat(core): execute task tools concurrently for improved performance by @tanzhenxin in https://github.com/QwenLM/qwen-code/pull/2434
* refactor(core): improve error handling and quota detection by @tanzhenxin in https://github.com/QwenLM/qwen-code/pull/2458
* feat(arena): Add agent collaboration arena with multi-model competitive execution by @tanzhenxin in https://github.com/QwenLM/qwen-code/pull/1912
* refactor: Refactors the VS Code file completion system to use fuzzy search by @qqqys in https://github.com/QwenLM/qwen-code/pull/2437
* feat(ui): Display token usage in the loading/progress indicator by @qqqys in https://github.com/QwenLM/qwen-code/pull/2445
* feat(vscode-ide-companion): add Tab key fill-only behavior for completions by @yiliang114 in https://github.com/QwenLM/qwen-code/pull/2431
* fix(vscode-ide-companion): update URI handling for Windows paths by @qqqys in https://github.com/QwenLM/qwen-code/pull/2457
* feat: add /context command to display context window token usage breakdown by @pomelo-nwu in https://github.com/QwenLM/qwen-code/pull/1835
* fix(test): update LoadingIndicator snapshot for correct output alignment by @qqqys in https://github.com/QwenLM/qwen-code/pull/2469
* feat: support skills in .agents directory and other provider directories by @LaZzyMan in https://github.com/QwenLM/qwen-code/pull/2202
* feat: add `auth` CLI command and Qwen Code Claw skill by @LaZzyMan in https://github.com/QwenLM/qwen-code/pull/2440
* feat(export): add metadata and statistics tracking by @Mingholy in https://github.com/QwenLM/qwen-code/pull/2328
* fix: correct token limits for MiniMax-M2.5 and GLM models by @Mingholy in https://github.com/QwenLM/qwen-code/pull/2470
* feat(hooks): Implement 10 core event hooks for session lifecycle and tool execution by @DennisYu07 in https://github.com/QwenLM/qwen-code/pull/2203
* fix: update TOS link in VS Code extension README by @Br1an67 in https://github.com/QwenLM/qwen-code/pull/2495
* fix: preserve modalities during OpenAI logging request conversion by @kkhomej33-netizen in https://github.com/QwenLM/qwen-code/pull/2473
* Feat: support permission by @LaZzyMan in https://github.com/QwenLM/qwen-code/pull/2283
## New Contributors
* @simon100500 made their first contribution in https://github.com/QwenLM/qwen-code/pull/2403
* @Br1an67 made their first contribution in https://github.com/QwenLM/qwen-code/pull/2495
**Full Changelog**: https://github.com/QwenLM/qwen-code/compare/v0.12.6...v0.13.0-nightly.20260320.b8dda154d