[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"tag-claudemd":3},{"tag":4,"articles":11},{"id":5,"name":6,"slug":7,"article_count":8,"description_zh":9,"description_en":10},"4013e340-8e53-47fb-b4d4-4491916b1ed3","CLAUDE.md","claudemd",3,"CLAUDE.md 是 Claude Code 專案層級的操作與上下文檔案，用來固定團隊規則、工作流程、目錄慣例與代理行為。它影響記憶、派工與工具使用，常和 Skills、Hooks、Subagents 一起設計，決定 AI 在真實開發中是否可控。","CLAUDE.md is the project-level instruction file for Claude Code. It defines repo conventions, workflow rules, and agent behavior, shaping memory, tool use, and task delegation. In practice it sits alongside Skills, Hooks, and Subagents to keep AI work predictable in real codebases.",[12],{"id":13,"slug":14,"title":15,"summary":16,"category":17,"image_url":18,"cover_image":18,"language":19,"created_at":20},"cbfdbb12-d79f-47c3-8a9a-df1443ff0d74","claude-code-advanced-patterns-six-months","Claude Code Advanced Patterns: 5 CLI Agent Design Tips","Most people use Claude Code for a week, come out with \"it's basically ChatGPT with a terminal,\" and miss its real differentiator: persistent memory and a system that grows with use. This article compresses six months of production use into five patterns that matter — what CLAUDE.md is really for, what to store (and not store) in MEMORY, the composition logic behind Skills, three ways people misuse Subagent dispatch, and the trade-offs between tmux, MCP servers, and file buses for cross-CLI orchestration.","ai-agent","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1776738623373-dy9r.png","en","2026-04-20T08:26:01.139966+00:00"]