{ "text": "字幕志愿者 杨茜茜\n在之前的视频中,我为大家演示了Deepseek Harness的基础用法\n视频发布后,我收到了非常多的私信\n让我继续深入讲解Deepseek Harness的高级用法\n本期视频就为大家演示Deepseek Harness更多高级使用技巧\n最近Deepseek Harness迭代速度非常快\n而且随着Deepseek V4 FlashVision模型的发布\nDeepseek Harness它的多模态能力大幅度增强\n无论是它的OCR能力,还是它的视觉能力\n都得到了显著提升\n在OCR能力方面\n它可以识别这些手写的笔记\n按照原有的格式提取并且输出\n在视觉能力方面\n它能将我们提供的房屋平面图\n用前端技术复刻出来这种3D可探索的模型\n而且它复刻的这些细节\n都遵循了我们给它提供的平面图\n然后它将平面图转换成3D可探索的模型\n下面我们再看一下Deep Sticker Harness中\n他新增了一个非常强大的功能\n也就是workflow功能\n这个功能相当于在执行任务的时候\n召集一批临时专家\n并行完成任务\n将每个专家完成任务的结果\n汇总给主agent\n然后再由主agent整理这些结果\n并且给出最终的结果\n像workflow这个功能\n它非常类似于cloudcode中的dynamic workflow\n在之前的视频中\n我为大家制作过多期cloudcode\ndynamic workflow的视频\n也就是在执行任务之前\n主agent会根据我们的任务\n编写一个脚本\n在编写好的脚本中\n它已经提前定义好了\n使用哪些subagent\n也就是子智能体\n并且每个智能体\n它负责什么任务\n而且还在脚本中明确定义了\n需要执行几个阶段\n每个阶段的任务是什么\n像这种方式\n就相当于有AI coding工具\n根据我们的实际任务\n来自主编写Harness\n所以这个理念还是非常超前的\n当它开始执行这个workflow的时候\n我们就可以直接用香港命令来查看这些workflow执行的状态\n包括这些任务所分的阶段\n然后每个阶段会派生出多少个subagent\n以及这些subagent所调用的模型\n还有实时消耗的token等\n而Deepseekharness它也吸纳了Cloudcode中这种dynamic workflow的功能\n所以在Deepseekharness中我们也可以来启用workflow功能\n尤其是在执行复杂任务的时候\n或者是对复杂的大型代码仓库进行code review的时候\n我们都可以使用workflow功能\n在deepseekharness中我们想使用workflow功能\n我们只需要选中这个标准模式\n因为标准模式它才支持workflow功能\n在这里我选中了一个代码仓库\n然后我们可以使用workflow功能\n让deepseekharness动态派生多个subagent\n对我们的代码仓库进行review\n然后我们就可以来数据提示测\n我们可以简单看一下这个提示测\n要求使用workflow对当前代码库进行一次并行指图的发布前审计\n并且要求启动4个子任务\n包括检查模块边界依赖方向\n检查错误处理\n检查测试覆盖口\n检查输入验证等\n在要求这里告诉他必须使用workflow\n并且禁止修改文件\n每个subagent都返回结构化结果\n然后我们就运行\n看一下deep secret harness\n它执行workflow的效果\n好在这里我们就可以看到\n他在构建这个工作流\n在这里我们可以看到\n他派生了四个并行执行的subagent\n然后我们可以点击一个进入进行查看\n点击之后我们就可以来查看\n这个agent他正在执行的这些任务\n这里会实时显示他执行的任务\n然后我们可以点击这条他执行的命令\n就可以看到他执行的这些命令\n然后我们如果想返回到刚才的这个界面\n我们只需要点击这个任务\n这样的话就回到了刚才\n我们查看到这个状态\n可以看到这里这个任务已经执行完成\n我们再点击看一下\n这个任务它执行的状态\n点击之后我们就看到了这个任务正在执行\n而且还可以看到它读取了哪些文件\n执行了哪些具体的命令\n以及它实时输出的这些状态\n好这里就提示它已经执行完了这个任务\n在这里给出的就是结构化的执行结果\n这是它按照我们要求的JSON格式\n输出的这个最终结果\n这个结构化的结果\n最后就会由主Agent进行整理\n然后我们再继续查看这些状态\n当任务执行完成之后\n这里就会显示绿色\n然后这里还有一个任务还在执行\n我们点击进行查看\n这个任务还在执行中\n然后我们再返回等它执行完成\n好这里这些任务全都执行完成\n现在正在输出最终的这个汇总结果\n首先这里提示这个工作流已经完成\n然后4个subagent已全部成功\n共收集25条原始问题\n然后他按严重程度减轻排序\n这里是给出的这些执行阶段\n这里就是按照严重程度来输出的最终的结果\n他按由高到低减轻的排序\n这里是发现的问题\n这里就是给出的存在这个问题的代码文件\n在这里就给出了证据要点\n然后还给出了修复建议\n他输出的这个结果是非常详细的\n在下面这里就是完整的这些jessing结果\n在最后这里他就给出的总体判断\n然后我们还可以让他来输出\n他构建的这个工作流的代码\n然后我们看一下这个代码是怎样的\n在这里我输入提示次\n让他将这个工作流的代码展示给我\n然后发送\n可以看到这里他开始展示这个代码\n首先这里就是工作流标识与阶段声明\n像这个就和Cloudcode\n它的Dynamic Workflow就非常像\n因为刚才我们也查看了Cloudcode\n它生成的Dynamic Workflow的代码\n然后DeepSig harness它构建的这个代码\n也包含工作流名称\n然后具体的描述\n还有分为这几个阶段\n然后我们再看一下最具体的脚本\n它构建的这个脚本也是纯js\n像Cloudcode构建的workflow的脚本\n也是由js编写的\n在这个脚本中\n它还定义了跳过哪些文件\n然后我们再继续看\n这里构建了一个返回的这些格式\n而且可以看到\n它构建的这个脚本写的非常完善\n在这里就是传入的参数\n也就是这个项目的路径\n然后他给出了这个脚本的结构说明\n在这里就是他定义的这四个subagent\n这里还列出了每个subagent他们的职责\n在最后这里他提到每个subagent独立运行互不可见\n好相见的话我们就可以在deep-sigal harness中\n直接通过自然语言描述\n让deep-sigal harness来动态编写工作流的脚本\n然后按照编写好的工作流脚本来执行任务\nDeepSync Harness它就具备了Cloud Code\n最早实现了动态工作流的功能\n像这个功能还是非常非常强大的\n尤其是针对复杂的项目\n复杂的代码库或者复杂的任务\n刚才我也要演示的是DeepSync Harness\n它的Workflow功能\n而且DeepSync Harness它还支持Agent Teams功能\n这个功能就非常类似于Cloud Code中的Agent Teams功能\n像Agent Teams功能它和Workflow的区别就是\nagent teams它是建立一个可以持续存在的开发团队\n它会先创建任务看板\n然后分配成员\n也就是分配agent\n这些agent之间可以互相发送消息\n而刚才为代言演示的workflow功能\n这些agent之间无法互相发送消息\n所以agent teams的优势\n就是这些成员之间可以互相发送消息\n而且还可以实现多轮推进\n当这些成员的任务都完成之后\n最后会由这个主agent\n也就是领导这个团队的agina对这些内容进行汇总\n想使用agina teams功能也非常简单\n我们只需要保证我们的deep secret harness已经升级到了最新版\n在模式这里我们还是选择这个标准模式\n在这里是我本机上一个已有的代码仓库\n然后我们就可以直接使用自然语言\n让deep secret harness来启动agina teams\n比如说我这里\n比如说我这里直接输入自然语言\n让他使用agent teams对当前代码库进行code review\n然后我们就直接发送\n然后deep secret harness他会先了解一下当前代码库的结构和规模\n然后他才会设计合适的agent teams评审方案\n像这种方式还是非常灵活的\n他并不是收到指令之后就直接构建几个成员的团队\n而是先粗略了解一下我们的代码库\n然后再构建真正的agent teams\n好这里他提示这个代码库已经清楚\n包含ls应用\n还有Chrome拓展\n还有superbase后端以及脚本\n他提到他会快速统计规模\n然后设计agent的评审方案\n这里他提示这个代码库已经明确\n他将按照agent teams模式来组织评审\n包括一个上下文观察的agent\n还有7个专业评审agent\n以及一个主审稿agent\n在这里我们可以看到\n他这里生成了这个任务\n然后现在正在执行agent teams评审方案\n在这里我们就看到他正在执行\n然后我们可以点击进行查看\n这里是他生成的这个项目简报\n包括产品是什么架构总览\n还有就是评审建议\n然后我们点击左上角再返回刚才的这个状态\n然后我们继续查看\n他已经构建好了这个agent teams\n包含7个成员\n他们的状态就是运行中\n然后我们可以点击其中的第一个进入查看\n我们可以看到这里他收到了刚才的上下文件报\n然后下面就是给他的这些具体提示词\n然后这里他正在执行\n可以看到他正在读取代码库\n还没有输出其他内容\n然后我们再返回继续查看\n在左上角我们将鼠标定位到这里\n就可以看到这些agent他们执行的实施状态\n以及他们运行的时间\n还有所消耗的token\n比如说我们可以进入一个进行查看\n然后这个agent他已经执行完成\n他输出了最终的这个报告\n包括总结还有发现的这些问题\n然后我们就可以返回\n继续等待这些agent执行完成\n这个7个成员的agent提姆斯\n已经全部执行完成\n现在再由这个领导agent在执行\n现在他开始输出这个代码评审的最终结果\n他提到已经按照agentteams模式\n对当前代码库进行评审\n共9个agent参与\n包括一个侦查agent\n然后7个专业评审并行审查\n还包含一个主审agent\n这里给出这个完成报告\n已经在这个文件中\n我们点击就可以查看\n点击之后就打开了这个代码审查报告\n这里我们就不再具体去看了\n好这里给出的是这个评审团队的构成\n然后这里就是核心结论\n这里给出的就是这些需要修复的问题的优先级\n这是我们测试的deep-sick harness\n它的aginal tips功能\n在上一期演示deep-sick harness的视频发布后\n在私信中我还收到了好多让我讲解\n如何为deep-sick harness来创建插件\n因为deep-sick harness它的理念就是一切接插件\n然后我们可以将我们的工作流直接封装成插件\n为了方便大家零门槛为Deepseekharness来开发插件\n我这里直接创建了一个skill\n然后使用这个skill就可以为Deepseekharness创建插件\n然后大家可以将这个skill为codex或者cloudcodeDeepseekharness进行安装\n安装方式非常简单\n大家可以直接查看说明文档进行手动安装\n然后也可以直接复制这个项目的仓库链接\n然后在Deepseekharness中我们就可以直接用自然源描述\n让他阅读这个skill的说明文档并且安装\n在后面我们跟上这个仓库链接就可以\n然后点击发送\n当安装完成之后\n我们就用斜杠命令\n就可以来调出这个scale\n然后我们就可以用自然语言描述\n我们所需要创建的这些插件\n我示茂的这个任务是\n让他为我创建一个\n可以为deepstick harness设置定时任务的插件\n然后我们就点击发送\n他就会调用\n刚才我们安装的这个scale\n来为我们创建这个插件\n好 这里创建完成\n他为我们创建了一个专门为DeepseekHarness\n来设置定时任务的一个插件\n在说明文档中\n他给出了这个定时任务的插件\n与官方自带的这个插件有哪些区别\n像这些优势是官方插件所不具备的优势\n而且我还使用这个创建插件的这个skill\n为DeepseekHarness开发了一个\n能够调用PayAgent进行code review的插件\n在DeepseekHarness中我们可以测试一下\n比如说这里我可以用DeepseekHarness\n将这个平面图做成的这个3D可探索项目\n对这个项目进行code review\n我说的提示词就是调用这个插件\n对代码进行review\n然后我们就直接发送\n在这里就可以看到他调用了Pi code review\n在最后这里他就提示Pi审查通过\n然后这里是给出的审查结论\n这里还给出了验证方式\n像这样的话\n我们就可以为Deepseekharness\n零门槛来创建我们所需要的各种插件\n本期视频为大家分享了DeepSig Harness\n它的Workflow\n以及Agent Teams的使用方式\n以及如何零门槛为DeepSig Harness\n来创建插件\n掌握了DeepSig Harness这几个使用技巧\n就能大幅提升我们的开发效率\n本期视频就做到这里\n欢迎大家点赞关注和转发\n谢谢大家观看\n请不吝点赞 订阅 转发 打赏支持明镜与点点栏目", "language": "multilingual", "duration": 797.651875, "segments": [ { "start": 0.0, "end": 1.758, "text": "字幕志愿者 杨茜茜", "language": "zh", "chunk": 0, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 2.874, "end": 6.524, "text": "在之前的视频中,我为大家演示了Deepseek Harness的基础用法", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 6.524, "end": 9.424, "text": "视频发布后,我收到了非常多的私信", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 9.424, "end": 13.024, "text": "让我继续深入讲解Deepseek Harness的高级用法", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 13.024, "end": 17.304, "text": "本期视频就为大家演示Deepseek Harness更多高级使用技巧", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 17.304, "end": 20.404, "text": "最近Deepseek Harness迭代速度非常快", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 20.404, "end": 24.024, "text": "而且随着Deepseek V4 FlashVision模型的发布", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 24.024, "end": 27.304, "text": "Deepseek Harness它的多模态能力大幅度增强", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 27.304, "end": 31.044, "text": "无论是它的OCR能力,还是它的视觉能力", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 31.044, "end": 32.804, "text": "都得到了显著提升", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 32.804, "end": 34.364, "text": "在OCR能力方面", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 34.364, "end": 36.824, "text": "它可以识别这些手写的笔记", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 36.824, "end": 40.124, "text": "按照原有的格式提取并且输出", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 40.124, "end": 41.584, "text": "在视觉能力方面", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 41.584, "end": 44.164, "text": "它能将我们提供的房屋平面图", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 44.164, "end": 48.444, "text": "用前端技术复刻出来这种3D可探索的模型", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 48.444, "end": 50.324, "text": "而且它复刻的这些细节", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 50.324, "end": 52.644, "text": "都遵循了我们给它提供的平面图", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 52.644, "end": 56.104, "text": "然后它将平面图转换成3D可探索的模型", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 56.104, "end": 58.784, "text": "下面我们再看一下Deep Sticker Harness中", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 58.784, "end": 61.404, "text": "他新增了一个非常强大的功能", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 61.404, "end": 62.764, "text": "也就是workflow功能", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 62.764, "end": 65.324, "text": "这个功能相当于在执行任务的时候", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 65.324, "end": 66.944, "text": "召集一批临时专家", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 66.944, "end": 68.284, "text": "并行完成任务", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 68.284, "end": 70.784, "text": "将每个专家完成任务的结果", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 70.784, "end": 72.364, "text": "汇总给主agent", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 72.364, "end": 75.184, "text": "然后再由主agent整理这些结果", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 75.184, "end": 77.344, "text": "并且给出最终的结果", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 77.344, "end": 78.764, "text": "像workflow这个功能", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 78.764, "end": 82.624, "text": "它非常类似于cloudcode中的dynamic workflow", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 82.624, "end": 84.144, "text": "在之前的视频中", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 84.144, "end": 86.304, "text": "我为大家制作过多期cloudcode", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 86.304, "end": 87.964, "text": "dynamic workflow的视频", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 87.964, "end": 89.964, "text": "也就是在执行任务之前", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 89.964, "end": 92.404, "text": "主agent会根据我们的任务", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 92.404, "end": 93.664, "text": "编写一个脚本", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 93.664, "end": 95.264, "text": "在编写好的脚本中", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 95.264, "end": 96.984, "text": "它已经提前定义好了", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 96.984, "end": 98.464, "text": "使用哪些subagent", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 98.464, "end": 100.024, "text": "也就是子智能体", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 100.024, "end": 101.324, "text": "并且每个智能体", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 101.324, "end": 102.704, "text": "它负责什么任务", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 102.704, "end": 105.124, "text": "而且还在脚本中明确定义了", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 105.124, "end": 106.644, "text": "需要执行几个阶段", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 106.644, "end": 108.104, "text": "每个阶段的任务是什么", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 108.104, "end": 109.364, "text": "像这种方式", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 109.364, "end": 111.524, "text": "就相当于有AI coding工具", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 111.524, "end": 113.044, "text": "根据我们的实际任务", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 113.044, "end": 114.624, "text": "来自主编写Harness", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 114.624, "end": 117.224, "text": "所以这个理念还是非常超前的", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 117.224, "end": 119.424, "text": "当它开始执行这个workflow的时候", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 119.424, "end": 123.764, "text": "我们就可以直接用香港命令来查看这些workflow执行的状态", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 123.764, "end": 125.764, "text": "包括这些任务所分的阶段", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 125.764, "end": 128.864, "text": "然后每个阶段会派生出多少个subagent", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 128.864, "end": 131.224, "text": "以及这些subagent所调用的模型", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 131.224, "end": 132.804, "text": "还有实时消耗的token等", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 132.804, "end": 137.824, "text": "而Deepseekharness它也吸纳了Cloudcode中这种dynamic workflow的功能", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 137.824, "end": 142.004, "text": "所以在Deepseekharness中我们也可以来启用workflow功能", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 142.004, "end": 144.244, "text": "尤其是在执行复杂任务的时候", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 144.244, "end": 148.324, "text": "或者是对复杂的大型代码仓库进行code review的时候", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 148.324, "end": 150.024, "text": "我们都可以使用workflow功能", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 150.024, "end": 153.364, "text": "在deepseekharness中我们想使用workflow功能", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 153.364, "end": 155.764, "text": "我们只需要选中这个标准模式", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 155.764, "end": 158.484, "text": "因为标准模式它才支持workflow功能", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 158.484, "end": 160.664, "text": "在这里我选中了一个代码仓库", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 160.664, "end": 162.684, "text": "然后我们可以使用workflow功能", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 162.684, "end": 165.744, "text": "让deepseekharness动态派生多个subagent", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 165.744, "end": 167.924, "text": "对我们的代码仓库进行review", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 167.924, "end": 169.724, "text": "然后我们就可以来数据提示测", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 169.724, "end": 172.224, "text": "我们可以简单看一下这个提示测", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 172.224, "end": 177.024, "text": "要求使用workflow对当前代码库进行一次并行指图的发布前审计", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 177.024, "end": 179.324, "text": "并且要求启动4个子任务", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 179.324, "end": 181.724, "text": "包括检查模块边界依赖方向", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 181.724, "end": 182.924, "text": "检查错误处理", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 182.924, "end": 184.584, "text": "检查测试覆盖口", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 184.584, "end": 186.524, "text": "检查输入验证等", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 186.524, "end": 189.144, "text": "在要求这里告诉他必须使用workflow", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 189.144, "end": 190.804, "text": "并且禁止修改文件", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 190.804, "end": 193.484, "text": "每个subagent都返回结构化结果", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 193.484, "end": 194.624, "text": "然后我们就运行", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 194.624, "end": 196.124, "text": "看一下deep secret harness", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 196.124, "end": 198.484, "text": "它执行workflow的效果", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 198.484, "end": 200.384, "text": "好在这里我们就可以看到", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 200.384, "end": 202.624, "text": "他在构建这个工作流", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 202.624, "end": 203.964, "text": "在这里我们可以看到", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 203.964, "end": 206.884, "text": "他派生了四个并行执行的subagent", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 206.884, "end": 209.444, "text": "然后我们可以点击一个进入进行查看", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 209.444, "end": 211.524, "text": "点击之后我们就可以来查看", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 211.524, "end": 214.684, "text": "这个agent他正在执行的这些任务", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 214.684, "end": 216.924, "text": "这里会实时显示他执行的任务", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 216.924, "end": 219.604, "text": "然后我们可以点击这条他执行的命令", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 219.604, "end": 222.224, "text": "就可以看到他执行的这些命令", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 222.224, "end": 225.244, "text": "然后我们如果想返回到刚才的这个界面", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 225.244, "end": 227.044, "text": "我们只需要点击这个任务", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 227.044, "end": 229.584, "text": "这样的话就回到了刚才", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 229.584, "end": 231.144, "text": "我们查看到这个状态", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 231.144, "end": 233.544, "text": "可以看到这里这个任务已经执行完成", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 233.544, "end": 234.824, "text": "我们再点击看一下", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 234.824, "end": 237.084, "text": "这个任务它执行的状态", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 237.084, "end": 240.324, "text": "点击之后我们就看到了这个任务正在执行", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 240.324, "end": 243.184, "text": "而且还可以看到它读取了哪些文件", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 243.184, "end": 245.864, "text": "执行了哪些具体的命令", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 245.864, "end": 248.584, "text": "以及它实时输出的这些状态", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 248.584, "end": 251.924, "text": "好这里就提示它已经执行完了这个任务", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 251.924, "end": 255.064, "text": "在这里给出的就是结构化的执行结果", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 255.064, "end": 257.824, "text": "这是它按照我们要求的JSON格式", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 257.824, "end": 259.704, "text": "输出的这个最终结果", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 259.704, "end": 261.024, "text": "这个结构化的结果", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 261.024, "end": 263.384, "text": "最后就会由主Agent进行整理", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 263.384, "end": 265.784, "text": "然后我们再继续查看这些状态", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 265.784, "end": 267.344, "text": "当任务执行完成之后", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 267.344, "end": 268.504, "text": "这里就会显示绿色", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 268.504, "end": 270.704, "text": "然后这里还有一个任务还在执行", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 270.704, "end": 272.124, "text": "我们点击进行查看", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 272.124, "end": 273.964, "text": "这个任务还在执行中", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 273.964, "end": 276.484, "text": "然后我们再返回等它执行完成", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 276.484, "end": 279.004, "text": "好这里这些任务全都执行完成", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 279.004, "end": 282.744, "text": "现在正在输出最终的这个汇总结果", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 282.744, "end": 285.564, "text": "首先这里提示这个工作流已经完成", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 285.564, "end": 288.144, "text": "然后4个subagent已全部成功", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 288.144, "end": 290.204, "text": "共收集25条原始问题", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 290.204, "end": 292.524, "text": "然后他按严重程度减轻排序", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 292.524, "end": 294.724, "text": "这里是给出的这些执行阶段", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 294.724, "end": 298.804, "text": "这里就是按照严重程度来输出的最终的结果", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 298.804, "end": 300.744, "text": "他按由高到低减轻的排序", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 300.744, "end": 302.144, "text": "这里是发现的问题", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 302.144, "end": 305.224, "text": "这里就是给出的存在这个问题的代码文件", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 305.224, "end": 307.424, "text": "在这里就给出了证据要点", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 307.424, "end": 309.324, "text": "然后还给出了修复建议", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 309.324, "end": 311.804, "text": "他输出的这个结果是非常详细的", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 311.804, "end": 314.664, "text": "在下面这里就是完整的这些jessing结果", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 314.664, "end": 317.744, "text": "在最后这里他就给出的总体判断", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 317.744, "end": 320.124, "text": "然后我们还可以让他来输出", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 320.124, "end": 322.364, "text": "他构建的这个工作流的代码", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 322.364, "end": 324.824, "text": "然后我们看一下这个代码是怎样的", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 324.824, "end": 326.104, "text": "在这里我输入提示次", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 326.104, "end": 328.764, "text": "让他将这个工作流的代码展示给我", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 328.764, "end": 329.524, "text": "然后发送", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 329.524, "end": 332.304, "text": "可以看到这里他开始展示这个代码", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 332.304, "end": 335.224, "text": "首先这里就是工作流标识与阶段声明", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 335.224, "end": 337.024, "text": "像这个就和Cloudcode", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 337.024, "end": 339.664, "text": "它的Dynamic Workflow就非常像", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 339.664, "end": 341.944, "text": "因为刚才我们也查看了Cloudcode", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 341.944, "end": 344.484, "text": "它生成的Dynamic Workflow的代码", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 344.484, "end": 347.164, "text": "然后DeepSig harness它构建的这个代码", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 347.164, "end": 348.584, "text": "也包含工作流名称", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 348.584, "end": 350.004, "text": "然后具体的描述", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 350.004, "end": 351.744, "text": "还有分为这几个阶段", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 351.744, "end": 354.124, "text": "然后我们再看一下最具体的脚本", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 354.124, "end": 356.464, "text": "它构建的这个脚本也是纯js", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 356.464, "end": 359.124, "text": "像Cloudcode构建的workflow的脚本", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 359.124, "end": 360.764, "text": "也是由js编写的", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 360.764, "end": 361.864, "text": "在这个脚本中", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 361.864, "end": 364.444, "text": "它还定义了跳过哪些文件", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 364.444, "end": 366.064, "text": "然后我们再继续看", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 366.064, "end": 368.704, "text": "这里构建了一个返回的这些格式", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 368.704, "end": 369.724, "text": "而且可以看到", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 369.724, "end": 372.184, "text": "它构建的这个脚本写的非常完善", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 372.184, "end": 374.344, "text": "在这里就是传入的参数", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 374.344, "end": 375.884, "text": "也就是这个项目的路径", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 375.884, "end": 378.664, "text": "然后他给出了这个脚本的结构说明", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 378.664, "end": 381.564, "text": "在这里就是他定义的这四个subagent", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 381.564, "end": 384.584, "text": "这里还列出了每个subagent他们的职责", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 384.584, "end": 388.544, "text": "在最后这里他提到每个subagent独立运行互不可见", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 388.544, "end": 391.764, "text": "好相见的话我们就可以在deep-sigal harness中", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 391.764, "end": 393.564, "text": "直接通过自然语言描述", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 393.564, "end": 396.864, "text": "让deep-sigal harness来动态编写工作流的脚本", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 396.864, "end": 400.424, "text": "然后按照编写好的工作流脚本来执行任务", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 400.424, "end": 403.204, "text": "DeepSync Harness它就具备了Cloud Code", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 403.204, "end": 406.104, "text": "最早实现了动态工作流的功能", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 406.104, "end": 408.844, "text": "像这个功能还是非常非常强大的", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 408.844, "end": 410.844, "text": "尤其是针对复杂的项目", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 410.844, "end": 413.504, "text": "复杂的代码库或者复杂的任务", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 413.504, "end": 416.004, "text": "刚才我也要演示的是DeepSync Harness", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 416.004, "end": 417.604, "text": "它的Workflow功能", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 417.604, "end": 421.084, "text": "而且DeepSync Harness它还支持Agent Teams功能", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 421.084, "end": 425.824, "text": "这个功能就非常类似于Cloud Code中的Agent Teams功能", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 425.824, "end": 428.964, "text": "像Agent Teams功能它和Workflow的区别就是", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 428.964, "end": 432.984, "text": "agent teams它是建立一个可以持续存在的开发团队", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 432.984, "end": 434.984, "text": "它会先创建任务看板", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 434.984, "end": 436.364, "text": "然后分配成员", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 436.364, "end": 437.724, "text": "也就是分配agent", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 437.724, "end": 440.604, "text": "这些agent之间可以互相发送消息", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 440.604, "end": 443.164, "text": "而刚才为代言演示的workflow功能", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 443.164, "end": 446.264, "text": "这些agent之间无法互相发送消息", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 446.264, "end": 448.004, "text": "所以agent teams的优势", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 448.004, "end": 450.884, "text": "就是这些成员之间可以互相发送消息", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 450.884, "end": 453.024, "text": "而且还可以实现多轮推进", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 453.024, "end": 455.544, "text": "当这些成员的任务都完成之后", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 455.544, "end": 457.404, "text": "最后会由这个主agent", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 457.404, "end": 462.164, "text": "也就是领导这个团队的agina对这些内容进行汇总", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 462.164, "end": 464.604, "text": "想使用agina teams功能也非常简单", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 464.604, "end": 468.364, "text": "我们只需要保证我们的deep secret harness已经升级到了最新版", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 468.364, "end": 471.444, "text": "在模式这里我们还是选择这个标准模式", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 471.444, "end": 474.364, "text": "在这里是我本机上一个已有的代码仓库", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 474.364, "end": 476.524, "text": "然后我们就可以直接使用自然语言", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 476.524, "end": 479.244, "text": "让deep secret harness来启动agina teams", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 479.244, "end": 480.0, "text": "比如说我这里", "language": "zh", "chunk": 0, "local_language_run": 1, "language_run": 0, "source": "whisper" }, { "start": 480.0, "end": 481.64, "text": "比如说我这里直接输入自然语言", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 481.9, "end": 485.48, "text": "让他使用agent teams对当前代码库进行code review", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 485.74, "end": 487.02, "text": "然后我们就直接发送", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 487.28, "end": 491.62, "text": "然后deep secret harness他会先了解一下当前代码库的结构和规模", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 491.62, "end": 495.82, "text": "然后他才会设计合适的agent teams评审方案", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 495.82, "end": 498.22, "text": "像这种方式还是非常灵活的", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 498.22, "end": 502.62, "text": "他并不是收到指令之后就直接构建几个成员的团队", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 502.62, "end": 505.22, "text": "而是先粗略了解一下我们的代码库", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 505.22, "end": 507.82, "text": "然后再构建真正的agent teams", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 507.82, "end": 510.42, "text": "好这里他提示这个代码库已经清楚", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 510.42, "end": 511.82, "text": "包含ls应用", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 511.82, "end": 513.02, "text": "还有Chrome拓展", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 513.02, "end": 515.42, "text": "还有superbase后端以及脚本", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 515.42, "end": 517.82, "text": "他提到他会快速统计规模", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 517.82, "end": 519.82, "text": "然后设计agent的评审方案", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 519.82, "end": 522.02, "text": "这里他提示这个代码库已经明确", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 522.02, "end": 525.4, "text": "他将按照agent teams模式来组织评审", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 525.4, "end": 527.82, "text": "包括一个上下文观察的agent", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 527.82, "end": 529.82, "text": "还有7个专业评审agent", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 529.82, "end": 531.9, "text": "以及一个主审稿agent", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 531.9, "end": 533.12, "text": "在这里我们可以看到", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 533.12, "end": 535.04, "text": "他这里生成了这个任务", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 535.04, "end": 538.24, "text": "然后现在正在执行agent teams评审方案", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 538.24, "end": 540.64, "text": "在这里我们就看到他正在执行", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 540.64, "end": 542.46, "text": "然后我们可以点击进行查看", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 542.46, "end": 544.64, "text": "这里是他生成的这个项目简报", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 544.64, "end": 547.06, "text": "包括产品是什么架构总览", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 547.06, "end": 548.58, "text": "还有就是评审建议", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 548.58, "end": 552.88, "text": "然后我们点击左上角再返回刚才的这个状态", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 552.88, "end": 553.98, "text": "然后我们继续查看", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 553.98, "end": 556.36, "text": "他已经构建好了这个agent teams", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 556.36, "end": 557.76, "text": "包含7个成员", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 557.76, "end": 559.84, "text": "他们的状态就是运行中", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 559.84, "end": 563.28, "text": "然后我们可以点击其中的第一个进入查看", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 563.28, "end": 566.74, "text": "我们可以看到这里他收到了刚才的上下文件报", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 566.74, "end": 569.56, "text": "然后下面就是给他的这些具体提示词", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 569.56, "end": 571.46, "text": "然后这里他正在执行", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 571.46, "end": 573.32, "text": "可以看到他正在读取代码库", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 573.32, "end": 574.92, "text": "还没有输出其他内容", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 574.92, "end": 576.76, "text": "然后我们再返回继续查看", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 576.76, "end": 580.02, "text": "在左上角我们将鼠标定位到这里", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 580.02, "end": 583.74, "text": "就可以看到这些agent他们执行的实施状态", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 583.74, "end": 585.14, "text": "以及他们运行的时间", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 585.14, "end": 586.3, "text": "还有所消耗的token", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 586.3, "end": 588.48, "text": "比如说我们可以进入一个进行查看", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 588.48, "end": 590.68, "text": "然后这个agent他已经执行完成", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 590.68, "end": 593.26, "text": "他输出了最终的这个报告", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 593.26, "end": 596.16, "text": "包括总结还有发现的这些问题", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 596.16, "end": 597.7, "text": "然后我们就可以返回", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 597.7, "end": 600.6, "text": "继续等待这些agent执行完成", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 600.6, "end": 602.96, "text": "这个7个成员的agent提姆斯", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 602.96, "end": 604.74, "text": "已经全部执行完成", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 604.74, "end": 607.58, "text": "现在再由这个领导agent在执行", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 607.58, "end": 610.7, "text": "现在他开始输出这个代码评审的最终结果", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 610.7, "end": 613.16, "text": "他提到已经按照agentteams模式", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 613.16, "end": 615.3, "text": "对当前代码库进行评审", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 615.3, "end": 616.76, "text": "共9个agent参与", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 616.76, "end": 618.22, "text": "包括一个侦查agent", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 618.22, "end": 620.78, "text": "然后7个专业评审并行审查", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 620.78, "end": 622.72, "text": "还包含一个主审agent", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 622.72, "end": 624.62, "text": "这里给出这个完成报告", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 624.62, "end": 626.0, "text": "已经在这个文件中", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 626.0, "end": 627.44, "text": "我们点击就可以查看", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 627.44, "end": 630.36, "text": "点击之后就打开了这个代码审查报告", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 630.36, "end": 632.7, "text": "这里我们就不再具体去看了", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 632.7, "end": 635.94, "text": "好这里给出的是这个评审团队的构成", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 636.18, "end": 637.7, "text": "然后这里就是核心结论", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 637.98, "end": 641.58, "text": "这里给出的就是这些需要修复的问题的优先级", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 641.78, "end": 643.46, "text": "这是我们测试的deep-sick harness", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 643.46, "end": 645.02, "text": "它的aginal tips功能", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 645.3, "end": 648.74, "text": "在上一期演示deep-sick harness的视频发布后", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 648.74, "end": 651.26, "text": "在私信中我还收到了好多让我讲解", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 651.26, "end": 653.94, "text": "如何为deep-sick harness来创建插件", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 654.1, "end": 657.54, "text": "因为deep-sick harness它的理念就是一切接插件", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 657.54, "end": 661.22, "text": "然后我们可以将我们的工作流直接封装成插件", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 661.22, "end": 664.76, "text": "为了方便大家零门槛为Deepseekharness来开发插件", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 664.76, "end": 666.72, "text": "我这里直接创建了一个skill", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 666.72, "end": 670.42, "text": "然后使用这个skill就可以为Deepseekharness创建插件", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 670.42, "end": 675.84, "text": "然后大家可以将这个skill为codex或者cloudcodeDeepseekharness进行安装", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 675.84, "end": 677.28, "text": "安装方式非常简单", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 677.28, "end": 680.16, "text": "大家可以直接查看说明文档进行手动安装", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 680.16, "end": 683.54, "text": "然后也可以直接复制这个项目的仓库链接", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 683.54, "end": 686.82, "text": "然后在Deepseekharness中我们就可以直接用自然源描述", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 686.82, "end": 689.64, "text": "让他阅读这个skill的说明文档并且安装", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 689.64, "end": 692.16, "text": "在后面我们跟上这个仓库链接就可以", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 692.16, "end": 693.32, "text": "然后点击发送", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 693.32, "end": 694.7, "text": "当安装完成之后", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 694.7, "end": 695.88, "text": "我们就用斜杠命令", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 695.88, "end": 698.28, "text": "就可以来调出这个scale", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 698.28, "end": 700.6, "text": "然后我们就可以用自然语言描述", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 700.6, "end": 702.78, "text": "我们所需要创建的这些插件", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 702.78, "end": 704.06, "text": "我示茂的这个任务是", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 704.06, "end": 705.7, "text": "让他为我创建一个", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 705.7, "end": 708.96, "text": "可以为deepstick harness设置定时任务的插件", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 708.96, "end": 710.74, "text": "然后我们就点击发送", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 710.74, "end": 711.92, "text": "他就会调用", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 711.92, "end": 713.76, "text": "刚才我们安装的这个scale", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 713.76, "end": 716.22, "text": "来为我们创建这个插件", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 716.22, "end": 717.76, "text": "好 这里创建完成", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 717.76, "end": 720.64, "text": "他为我们创建了一个专门为DeepseekHarness", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 720.64, "end": 722.6, "text": "来设置定时任务的一个插件", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 722.6, "end": 723.9, "text": "在说明文档中", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 723.9, "end": 726.2, "text": "他给出了这个定时任务的插件", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 726.2, "end": 729.24, "text": "与官方自带的这个插件有哪些区别", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 729.24, "end": 732.96, "text": "像这些优势是官方插件所不具备的优势", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 732.96, "end": 735.88, "text": "而且我还使用这个创建插件的这个skill", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 735.88, "end": 738.14, "text": "为DeepseekHarness开发了一个", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 738.14, "end": 741.86, "text": "能够调用PayAgent进行code review的插件", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 741.86, "end": 744.66, "text": "在DeepseekHarness中我们可以测试一下", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 744.66, "end": 747.06, "text": "比如说这里我可以用DeepseekHarness", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 747.06, "end": 750.8, "text": "将这个平面图做成的这个3D可探索项目", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 750.8, "end": 752.82, "text": "对这个项目进行code review", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 752.82, "end": 755.4, "text": "我说的提示词就是调用这个插件", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 755.4, "end": 756.7, "text": "对代码进行review", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 756.7, "end": 758.06, "text": "然后我们就直接发送", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 758.06, "end": 760.96, "text": "在这里就可以看到他调用了Pi code review", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 760.96, "end": 763.9, "text": "在最后这里他就提示Pi审查通过", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 763.9, "end": 766.02, "text": "然后这里是给出的审查结论", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 766.02, "end": 768.02, "text": "这里还给出了验证方式", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 768.02, "end": 768.94, "text": "像这样的话", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 768.94, "end": 770.86, "text": "我们就可以为Deepseekharness", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 770.86, "end": 774.8, "text": "零门槛来创建我们所需要的各种插件", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 774.8, "end": 777.44, "text": "本期视频为大家分享了DeepSig Harness", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 777.44, "end": 778.38, "text": "它的Workflow", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 778.38, "end": 780.56, "text": "以及Agent Teams的使用方式", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 780.56, "end": 783.0, "text": "以及如何零门槛为DeepSig Harness", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 783.0, "end": 784.2, "text": "来创建插件", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 784.2, "end": 787.18, "text": "掌握了DeepSig Harness这几个使用技巧", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 787.18, "end": 789.48, "text": "就能大幅提升我们的开发效率", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 789.48, "end": 791.2, "text": "本期视频就做到这里", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 791.2, "end": 793.7, "text": "欢迎大家点赞关注和转发", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 793.7, "end": 794.9, "text": "谢谢大家观看", "language": "zh", "chunk": 1, "local_language_run": 0, "language_run": 0, "source": "whisper" }, { "start": 796.449, "end": 797.651875, "text": "请不吝点赞 订阅 转发 打赏支持明镜与点点栏目", "language": "zh", "chunk": 1, "local_language_run": 1, "language_run": 0, "source": "whisper" } ], "transcript_selection": { "schema": "videonote.transcript_selection.v1", "status": "SELECTED", "requested_language": "zh", "duration": 797.651875, "reference_path": null, "reference_kind": null, "reference_language": null, "selected_source": "whisper_segmented_multilingual", "decision_reason": "multilingual_segmented_whisper_passed", "whisper_rejected": false, "source_profile": { "segment_count": 0, "duration_coverage": 0.0, "usable_full": false, "usable_partial": false, "hallucination_region_count": 0, "hallucination_regions": [], "language": { "requested": "zh", "matched": false } }, "whisper_profile": { "segment_count": 332, "multilingual": false, "first_start": 0.0, "final_end": 797.652, "duration_coverage": 1.0, "maximum_gap_seconds": 1.549, "inverted_count": 0, "language": { "requested": "zh", "matched": true, "latin": 1407, "han": 3639, "kana": 0, "hangul": 0 }, "hallucination_region_count": 0, "hallucination_regions": [], "usable_full": true, "usable_partial": true }, "comparison": { "window_seconds": 30.0, "compared_window_count": 0, "median_score": 0.0, "divergent_window_count": 0, "divergent_ratio": 0.0, "divergent_windows": [] }, "timestamp_alignment": { "applied": false, "estimated_offset_seconds": 0.0, "matched_segment_count": 0, "match_scores": [] }, "caption_timing_repair": null, "filled_source_profile": null, "whisper_duration_guard": { "media_duration": 797.652, "rejected_segment_count": 0, "rejected_segments": [] } } }