lerd-env/lerd：README 來源編輯指南

lerd-env/lerd 的 README 將專案描述為「Open-source, Herd-like local PHP development environment for Linux and macOS. Automatic .test domains, per-project PHP/Node isolation, one-command TLS. Podman-native, rootless.」。本文只整理倉庫可直接核對的內容，不把 star、Fork 或宣傳語當成品質證明。README 在「Lerd」下寫到：> Open-source Herd-like local PHP development environment for Linux and macOS, > with Windows supported via WSL2 (beta). Podman-native, rootless, with a > built-in Web UI.。這說明的是專案邊界，不是已完成的生產驗證。

從 README 的「Sites, domains and TLS」與相關條目，可以先判斷它是否處理你的實際問題：🔗 Site groups. Group related sites so a main site owns a base domain and the rest occupy its subdomains, with a shared or separate database per secondary.。若需求不同，不應只因專案熱度就採用。本文保留原始專案名、命令與元件名，方便回到一手來源核對。 README 另外列出一項可核對的資訊：🌐 Automatic .test domains. One command gives a project a hostname and TLS that reissues before it expires, with no dnsmasq, no system resolver tweak and no sudo for the DNS bits.。這類原文條目可用來設計試跑步驟，但不能取代實際環境測試。

README 將運作方式分散在「Built for Linux PHP developers」等段落。可確認的線索包括：Lerd is built for PHP developers on Linux who want frictionless local development: automatic .test domains, per-project PHP versions, one-click HTTPS, zero Docker.。本文不把未寫出的架構、效能或安全邊界補成結論；真正的執行鏈仍要配合目錄、設定檔與版本標籤檢查。

第一次安裝應從 README 指出的入口開始。目前可核對的命令是： You: set up the project I just cloned AI: → site(action: "link") → exec(action: "composer", args: ["install"]) → env(action: "setup") # detects MySQL + Redis, starts them, creates DB, generates APP_KEY → framework(action: "setup") # storage:link + migrate for Laravel, doctrine:migrations:migrate for Symfony ✓ myapp → https://myapp.test ready 如果倉庫沒有命令，本文不會自行編造步驟，而是建議先閱讀「Built for Linux PHP developers」，確認系統依賴、預設埠與首次初始化。

日常使用取決於專案文件。README 的「AI Integration (MCP)」段落提到：Lerd ships a built-in Model Context Protocol server. Connect it to Claude Code, Cursor, JetBrains Junie, Codex CLI, Gemini CLI, GitHub Copilot, Google Antigravity, Windsurf, or any MCP-compatible AI assistant and manage your dev。設定檔、環境變數、權限與資料目錄只在來源明確時才會記錄；沒有寫出的預設值，應在測試環境驗證並保留回滾副本。 同一部分也提到：🧱 Host-proxy sites. Run a Node, Python, Go or any non-PHP dev server on the host and have nginx serve it at a .test domain with HTTPS, git worktrees included.。