<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Claude-Code on Lost Temple</title><link>https://cloudcold.ai/tags/claude-code/</link><description>Recent content in Claude-Code on Lost Temple</description><generator>Hugo</generator><language>en</language><lastBuildDate>Mon, 08 Jun 2026 00:00:00 +0000</lastBuildDate><atom:link href="https://cloudcold.ai/tags/claude-code/index.xml" rel="self" type="application/rss+xml"/><item><title>我所知道的所有 Agentic Engineering 技巧（2026 年 6 月）</title><link>https://cloudcold.ai/posts/2026-06-08-agentic-engineering-hacks-2026/</link><pubDate>Mon, 08 Jun 2026 00:00:00 +0000</pubDate><guid>https://cloudcold.ai/posts/2026-06-08-agentic-engineering-hacks-2026/</guid><description>Matt Van Horn 的 22 个 Agentic Engineering hack 全解析：从 plan.md 到双引擎并行构建。</description></item><item><title>Claude Code Slash Commands 完全指南 — 从入门到生产级命令库</title><link>https://cloudcold.ai/posts/2026-06-07-claude-code-slash-commands-deep-dive/</link><pubDate>Sun, 07 Jun 2026 00:00:00 +0000</pubDate><guid>https://cloudcold.ai/posts/2026-06-07-claude-code-slash-commands-deep-dive/</guid><description>7个开箱即用命令模板 + 深度分析推文缺失的Skills系统、动态注入、Hooks等高级玩法，附行动建议。</description></item><item><title>4-Agent Pipeline: The Seductive Trap of 'Ship While You Sleep'</title><link>https://cloudcold.ai/posts/2026-06-06-4-agent-pipeline-seductive-trap/</link><pubDate>Sat, 06 Jun 2026 00:00:00 +0000</pubDate><guid>https://cloudcold.ai/posts/2026-06-06-4-agent-pipeline-seductive-trap/</guid><description>四个 AI Agent 的 Planner-Coder-Tester-Reviewer 流水线真正擅长的是小任务。Token 成本 4-7x 基线、无 feedback loop、silent failure 系统性存在。Anthropic 官方立场直接矛盾。学术背刺：14 种 multi-agent 失败模式。</description></item><item><title>每个任务都需要一个 Harness：Claude Code 动态工作流深度解析</title><link>https://cloudcold.ai/posts/2026-06-06-harness-claude-code/</link><pubDate>Sat, 06 Jun 2026 00:00:00 +0000</pubDate><guid>https://cloudcold.ai/posts/2026-06-06-harness-claude-code/</guid><description>Anthropic 把 agent 编排计划从 context window 搬进 JavaScript 脚本，用代码确定性替代模型实时判断。深度解析运行时拓扑、三个强化机制、竞品横向对比、Multi-Agent 性能陷阱（-70%）、~15x token 成本、历史类比。</description></item><item><title>Claude Code Dynamic Workflows: 从单兵到千军万马的编排革命</title><link>https://cloudcold.ai/posts/2026-06-04-claude-code-dynamic-workflows/</link><pubDate>Thu, 04 Jun 2026 00:00:00 +0000</pubDate><guid>https://cloudcold.ai/posts/2026-06-04-claude-code-dynamic-workflows/</guid><description>Anthropic Dynamic Workflows 全面解读：脚本驱动千军万马的编排架构，性能提升 90.2%，token 成本暴涨 15 倍。三层编排对比、竞品横向分析、历史类比、成本现实与关键风险。</description></item><item><title>Claude Code 动态工作流深度分析：当 AI 学会给自己写调度器</title><link>https://cloudcold.ai/posts/2026-06-03-claude-code-dynamic-workflows-analysis-guide/</link><pubDate>Wed, 03 Jun 2026 00:00:00 +0000</pubDate><guid>https://cloudcold.ai/posts/2026-06-03-claude-code-dynamic-workflows-analysis-guide/</guid><description>Anthropic 发布 Claude Code 动态工作流的全面深度分析：技术架构、平台战略、竞争格局、历史类比和个性化实操建议。</description></item></channel></rss>