<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:content="http://purl.org/rss/1.0/modules/content/"><channel><title>Agentic Loops on mayulu的AI笔记</title><link>https://mayulu.co/tags/agentic-loops/</link><description>Recent content in Agentic Loops on mayulu的AI笔记</description><generator>Hugo</generator><language>zhcn</language><copyright>Copyright © 2025–2026 mayulu All Rights Reserved</copyright><lastBuildDate>Tue, 14 Jul 2026 15:41:33 +0800</lastBuildDate><atom:link href="https://mayulu.co/tags/agentic-loops/index.xml" rel="self" type="application/rss+xml"/><item><title>循环工程：Claude官方入门指南</title><link>https://mayulu.co/posts/loop-engineering-anthropic/</link><pubDate>Tue, 14 Jul 2026 15:41:33 +0800</pubDate><guid>https://mayulu.co/posts/loop-engineering-anthropic/</guid><description>本文解读 Anthropic Claude Code 官方循环工程指南，定义智能体循环为重复迭代至达标，划分回合、目标、时间、主动四类循环，详解`/goal`、`/loop`、`/schedule`指令用法与适配场景，配套技能文件、多智能体评审等代码质量保障方案，提供轮次、调度、模型选型等 Token 成本管控手段，附选型对照表，适合 AI 编程、智能体自动化开发从业者学习落地自主工作流。</description></item></channel></rss>