本课程教你用Claude Code、子智能体与MCP服务器搭建AI协作工作流，涵盖CLAUDE.md、代码审查、规范驱动开发与全栈毕业项目，掌握工业级AI软件开发。
原始标题：Agentic Engineering: Build Software with AI Coding Agents
《智能体工程：用 AI 编码智能体构建软件》是一门前沿实战课程，专注于教授开发者如何驾驭 Claude Code、子智能体（Subagents） 和 MCP 服务器（如 Playwright），建立起一套包含“探索、计划、实现、验证、提交”的高效 AI 协作循环工作流。
课程不仅涵盖编写 CLAUDE.md 内存文件、定制自动化代码审查员以及为 React、Laravel 安全和数据库创建复用技能（Skills）等核心技巧，还深度融合了基于 OpenSpec 的规范驱动开发（Spec-driven Development）、GitHub PR、CI 以及 git worktrees 并行开发等现代化工程交付手段。学员最终将通过一个全栈毕业项目（使用 Laravel 13、React 和 PostgreSQL 开发的服务台应用），在 AI 辅助下完成从需求到数据建模、再到通过 133 个自动化测试并完美达成 12 项核心需求 的完整闭环，从而真正掌握构建安全、可测试且具备工业级交付标准的软件开发能力。
Published 10/2026
Created by Sahinur Islam
MP4 | Video: h264, 1920×1080 | Audio: AAC, 44.1 KHz, 2 Ch
Level: All Levels | Genre: eLearning | Language: English | Duration: 36 Lectures ( 7h 44m ) | Size: 1.8 GB
Learn to plan, build, test and ship real software with Claude Code, subagents, skills, MCP and spec-driven development
What you’ll learn
⚡ Plan, build and test real software with AI coding agents using the explore, plan, implement, verify, commit loop
⚡ Write a CLAUDE memory file, build subagents and a code reviewer, and create reusable skills for React, Laravel security and databases
⚡ Add hooks and MCP servers such as Playwright, and build and change features with spec-driven development using OpenSpec
⚡ Build a full-stack Support Desk app with Laravel 13, React and PostgreSQL, using GitHub pull requests, CI and git worktrees
Requirements
❗ No prior AI coding agent experience needed: the course starts with how coding agents work and how to run a real Claude Code session
Description
This course contains the use of artificial intelligence.
AI coding agents can write a lot of code fast. The hard part is steering them so the code is correct, tested and something you would actually ship. This course teaches you that skill, step by step, from your first session with an agent to a finished full-stack app.
You will start with how coding agents work and how to run a real Claude Code session. Then you will learn to manage context, write prompts that work, and follow a simple loop: explore, plan, implement, verify, commit. You will write a CLAUDE memory file that keeps the agent on track, build your own subagents and a code reviewer, and create reusable skills for React, Laravel security and databases.
Next you will add hooks, connect MCP servers such as Playwright for browser testing, and use spec-driven development with OpenSpec to build and change features from a written spec. You will also see how to work with GitHub pull requests and CI, and how to run parallel work with git worktrees and Conductor.
The course ends with a capstone project: a Support Desk app built with Laravel 13, React and PostgreSQL. You will go from requirements to data modeling to core features and an admin area, with 133 passing tests and all 12 requirements done.
What you get
✨ 36 lectures, from a short trailer to the course finale
✨ Every prompt, config file, skill, subagent and spec shown in the course
✨ The full capstone source code in a public GitHub repository, linked in the first lecture
Who this course is for
⭐ Developers who want to use AI coding agents on real projects, not just demos
⭐ Full-stack and web developers who want a repeatable workflow with Claude Code
⭐ Anyone who wants AI-written code that is planned, tested and reviewed
此处内容需要权限查看
会员免费查看
