持续智能体，以及在拉尔夫·维格姆之后会发生什么？
3 分•作者: waynenilsen•6 个月前
有没有人和我一样，在玩具项目上完全放手，让软件生命周期自主运行？我让 Claude 在一个类似 Ralph 的循环中运行了超过 15 个小时，无人监管，创建了 118 多个提交。<p>这个技术是这样的：<p>while true:
如果存在任务 -> 减少一个待办任务，退出
如果不存在 -> 确定接下来添加什么功能比较合理，
创建 PRD 和 ERD，分解成任务，退出<p>它曾经因为在非 tty 环境下运行 playwright 导致 tty 问题而卡住，但除此之外，我没有手动介入。<p>我用 systemd 在一个 droplet 上持续运行它。<p>这个代理正在创建的玩具代码是一个多租户待办事项 kata。这是提示集：<p>https://github.com/waynenilsen/ralph-kata-2/tree/main/prompts<p>任何人都可以制作自己的版本，这些只是对我有效的一组提示。<p>在 15 个小时内，它从头开始创建了一个完整的多租户身份验证系统，以及带有分配人、截止日期、电子邮件提醒、标签和全文搜索的待办事项。我手动创建了第一个 PRD，内容类似于“为多租户待办事项系统创建一个 PRD”。<p>对于任何想做类似事情的人来说，e2e 测试在闭合代理与现实的循环方面发挥了关键作用。<p>用提示进行编程的时代显然正在到来。
查看原文
Is anyone else doing the full software lifecycle for toy projects completely hands off the wheel? I have had Claude running in a Ralph like loop for over 15 hours unsupervised creating over 118 commits.<p>The technique works like this<p>while true:
if tickets exist -> burn down the backlog by one ticket, exit
if not -> figure out what feature would make sense to add next,
create PRD and ERD, break down into tickets, exit<p>It did get stuck once due to tty issues related to running playwright in a non-tty environment but otherwise I have not had to manually step in.<p>I have it running in a droplet using systemd continuously.<p>Toy code the agent is creating is a multi-tenant todo kata. Here is the set of prompts:<p>https://github.com/waynenilsen/ralph-kata-2/tree/main/prompts<p>Anyone could make their own version of the same, these are just the set of prompts that work for me.<p>In 15 hours it created a full multi-tenant auth system from scratch and todos with assignees due dates, email reminders, tags and full text search. I created the first PRD by hand with something like "create a PRD for a multi-tenant todo system".<p>For anyone looking to do something similar, the e2e tests have played a critical role in closing the agent's loop with reality.<p>The age of programming with prompts is clearly arriving.