<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Loop-Engineering on All about Raspberry Pi</title><link>https://hugozhu.site/tags/loop-engineering/</link><description>Recent content in Loop-Engineering on All about Raspberry Pi</description><generator>Hugo</generator><language>en</language><lastBuildDate>Wed, 10 Jun 2026 00:00:00 +0000</lastBuildDate><atom:link href="https://hugozhu.site/tags/loop-engineering/index.xml" rel="self" type="application/rss+xml"/><item><title>Loop Engineering：AI Agent 工程的第五层</title><link>https://hugozhu.site/post/2026/263-loop-engineering/</link><pubDate>Wed, 10 Jun 2026 00:00:00 +0000</pubDate><guid>https://hugozhu.site/post/2026/263-loop-engineering/</guid><description>&lt;p&gt;4 月底，OpenAI 发布了 Codex CLI 0.128.0。更新日志里藏着一句话：「Added persisted /goal workflows with app-server APIs, model tools, runtime continuation, and TUI controls.」几乎同一时间，Claude Code 也上线了 &lt;code&gt;/goal&lt;/code&gt; 指令。Greg Brockman 在推特上写了一句：「codex now has a built-in Ralph loop++.」&lt;/p&gt;
&lt;p&gt;大多数人把这条更新当作又一个 feature flag 滑过去了。但如果你仔细拆解 &lt;code&gt;/goal&lt;/code&gt; 的设计，会发现它代表的不是「又一个功能」，而是 AI Agent 工程的一次范式跃迁。&lt;/p&gt;
&lt;p&gt;&lt;a href="https://hugozhu.site/img/2026/loop-engineering.png"&gt;&lt;img src="https://hugozhu.site/img/2026/loop-engineering-thumb.jpg" alt="Loop Engineering 五层演进"&gt;&lt;/a&gt;&lt;/p&gt;</description></item></channel></rss>