Welcome to Learn Harness Engineering 欢迎来到“Harness Engineering”工程之道

Learn Harness Engineering is a course dedicated to the engineering of AI coding agents. We have deeply studied and synthesized the most advanced Harness Engineering theories and practices in the industry. Our core references include:本课程专研 AI 编程智能体的工程架构。我们博采众长，深研业内顶尖的 Harness Engineering 理论与实践，核心参考如下：

• OpenAI: Harness engineering: leveraging Codex in an agent-first worldOpenAI：《Harness engineering: leveraging Codex in an agent-first world》
• Anthropic: Effective harnesses for long-running agentsAnthropic：《Effective harnesses for long-running agents》
• Anthropic: Harness design for long-running application developmentAnthropic：《Harness design for long-running application development》
• Awesome Harness EngineeringAwesome Harness Engineering

Through systematic environment design, state management, verification, and control systems, this course teaches you how to make agentic coding tools like Codex and Claude Code truly reliable. It helps you build features, fix bugs, and automate development tasks by constraining your AI coding assistant with explicit rules and boundaries.所谓 Harness，便是通过环境构建、状态管理、校验与控制体系，为智能体筑起一道防线。本课程将教你如何驾驭 Codex、Claude Code 等工具，通过明确的规则与边界约束，让 AI 助手在构建功能、修复 Bug 及自动化开发任务时，变得真正可靠。

Get started 启程

Choose your learning path to get started. The course is divided into theoretical lectures, hands-on projects, and a copy-ready resource library.择一路径，由此启程。课程分为理论研习、实战演练与资源库三部分，助你由浅入深。

Lectures理论研习

Understand why strong models still fail and learn the theory behind effective harnesses.探究为何强力模型亦会失手，参透高效 Harness 工程背后的精妙理论。

Projects实战演练

Hands-on practice building a reliable agentic environment from scratch.亲手操练，从无到有，搭建起一套稳固的智能体开发环境。

Resource Library资源库

Copy-ready templates (AGENTS.md, feature_list.json) to use in your own repositories.内含 AGENTS.md、feature_list.json 等现成模板，信手拈来，即可嵌入你的代码仓库。

The Core Mechanism of a Harness Harness 的核心机理

A harness doesn't "make the model smarter"; rather, it establishes a closed-loop working system for the model. You can understand its core workflow through this simple diagram:Harness 的妙处，不在于“让模型变聪明”，而在于为模型建立一套闭环运行的系统。以此图为引，其精微之处一目了然：

r6njpc

What you will learn 所学为何

Here are some of the key concepts you will master:你将习得以下核心心法：

• Constrain agent behavior with explicit rules and boundaries.以明确的规则与边界，约束智能体的行为。
• Maintain context across long-running, multi-session tasks.在长时运行、多轮会话的任务中，稳固维持上下文。
• Stop agents from declaring victory too early.防范智能体过早自满，草率宣告任务完成。
• Verify work using full-pipeline tests and self-reflection.通过全链路测试与自我反思，校验工作成效。
• Make runtime observable and debuggable.让运行过程清晰可见，便于排查疑难。

Next steps 后续进阶

Once you understand the core concepts, these guides help you go deeper:既已洞悉核心，不妨循此深入，更上一层楼：

• Lecture 01: Why Capable Agents Still Fail: Start with the theory behind harness engineering.第一讲：为何强力智能体亦会失手？从 Harness 工程的理论根基谈起。
• Project 01: Baseline vs Minimal Harness: Walk through your first real task.项目一：基准与极简 Harness。带你走完实战的第一步。
• Templates: Grab the minimal harness pack (AGENTS.md, feature_list.json, claude-progress.md) for your own projects.模板：获取极简 Harness 工具包（含 AGENTS.md、feature_list.json、claude-progress.md），即刻用于你的项目。