deusyu/harness-engineering is a learning archive about designing repositories, rules, tools, and feedback loops so AI coding agents can work reliably. It is for engineers studying and practicing this approach to agent-assisted software development. Its catalogue entries package parts of the project’s guidance as an agent skill and instruction.
Borrowing it
Nothing to install: this file belongs to deusyu/harness-engineering. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/deusyu/harness-engineering/main/AGENTS.mdgit clone --depth 1 https://github.com/deusyu/harness-engineeringWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/instructions/deusyu/harness-engineering/agents-md)<a href="https://agentmods.dev/instructions/deusyu/harness-engineering/agents-md"><img src="https://agentmods.dev/badge/instructions/deusyu/harness-engineering/agents-md/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/instructions/deusyu/harness-engineering/agents-md"><img src="https://agentmods.dev/badge/instructions/deusyu/harness-engineering/agents-md.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.01815 | $0.01815 |
| Opus 5.5 | $0.00726 | $0.00726 |
| Sonnet 5.5 | $0.00363 | $0.00363 |
| Haiku 4.5 | $0.00181 | $0.00181 |
Grade A, and why
harness-engineering AGENTS.md scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2mo ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 59 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Harness Engineering 学习档案
记录我学习「Harness Engineering」的完整过程：从概念理解到独立实践。
来源：OpenAI — Harness Engineering: Harnessing Codex in an Agent-First World
仓库结构
| 目录 | 内容 | 说明 |
|---|---|---|
concepts/ |
概念笔记 | 原文核心概念的拆解与整理 |
thinking/ |
独立思考 | 自己的理解、质疑、延伸思考 |
practice/ |
动手实践 | 小项目实验，验证文章中的方法论 |
feedback/ |
反馈记录 | 实践中的踩坑、修正、迭代心得 |
works/ |
作品输出 | 可展示的成果（文章、工具、模板等） |
tools/ |
工具具像化 | 降低 6 维复杂度的杠杆库（带主张，不是 awesome-list） |
prompts/ |
提示词积累 | 学习过程中验证有效的提示词 |
references/ |
外部资源 | 相关文章、仓库、工具的索引 |
学习路线（进度）
- Phase 1：理解核心概念（concepts/，8 篇）
- Phase 2：形成自己的观点（thinking/，11 篇，持续中）
- Phase 3：选一个小项目实践（practice/，1 个 Ralph Demo）
- Phase 4：记录反馈迭代（feedback/，1 篇，持续中）
- Phase 5：输出可展示的作品（works/，40 篇翻译 + 1 篇原创 + 2 篇外部中文收录）
进度详情以人类向 README.md 的"学习路线"段为准；本节是给智能体的快照。
导航
每个子目录都有自己的 AGENTS.md，说明该目录的用途、内容组织方式和写作约定。 从任何一个目录开始，都能找到下一步该看什么。
机械化检查
scripts/check-consistency.sh 守护"漂移"问题：
- C1 —
references/articles.md编号 1..N 连续 - C2 — N 与下游 4 处声明同步（README badge × 2、
prompts/deep-research-tracker.md头部、references/AGENTS.md概览）。文件含独立行<!-- check-consistency: skip-count -->时豁免 - C3 —
concepts/、thinking/、feedback/的*.md实际数与 README 中"X 篇"声明一致 - C4 —
works/*-translation.md文件数 ≡ 翻译计数所有声明（badges、<details>摘要、Phase 5 注释、本文件 Phase 5 快照、READMEs 表格行数） - C5 — README 结构树中
concepts/子树的 item 行数 ≡concepts/*.md文件数（防止"计数对了但树漏了"） - C6 —
references/articles.md末尾"不计入 N 篇"中的 N ≡ C1 权威值 - C7 — 三脉络 per-track 计数（脉络一/二/三）在 4 处下游声明保持一致：READMEs 资料库表、
references/AGENTS.md三脉络小标题、prompts/deep-research-tracker.md三脉络明细 - C8 — 翻译流水线本地守卫：
translate/<...>/sources/<slug>/source-full.md存在时，对应01-analysis.md不得再声称"仅摘要页 / 建议补抓全文"。translate/已 gitignore，CI 与干净 clone 自动 SKIP，仅本地有过程稿时触发 - C9 —
concepts//thinking//feedback/正文不得裸写文库计数（"N 篇文章 / N 篇翻译 / N 大概念"）；历史性提法须带"写作时点 / 当时 / 此前 / 首批 / 首轮 / 截至 / 快照"限定词，否则去数字改链references/articles.md - C10 — 图片保真（纯本地、零网络）：每篇
works/*-translation.md的 frontmatter 必须声明sourceFigureCount（缺失即 FAIL；null= 原文不可得、未审计 → SKIP；数字 N → 正文嵌图数须 ≥ N），且所有本地嵌图路径（imgs/...）必须在磁盘上存在 - C11 — markdown 表格形状：README ×2、
references/AGENTS.md、references/articles.md、works/AGENTS.md里每一行表格的单元格数须与表头一致 - C12 — 条目字段完整性：
references/articles.md每个### N.编号条目必须带 作者： 与 日期： 字段 - C13 — 零插图声明须留痕：C10 只能证伪"多报"（嵌图数 < 声明数才 FAIL），因此
sourceFigureCount: 0在本地永远无法被证伪——不管你有没有真去核对原文，它都是绿的。2026-07-27 就是这个洞放行了一个假 0（原文实有 4 张配图）。C10 刻意零网络、无法回查原文，所以改为要求留痕：声明 0 的译文必须同时带sourceFigureAudit字段，值里要有YYYY-MM-DD核对日期，写清怎么核对的、结论是什么。null仍然 SKIP——它本来就自陈未审计 - C14 — 文档站防漂移：VitePress 站点（部署到 harness.dyu.sh）的侧边栏与首页统计一律由
.vitepress/sidebar.mjs在构建时从文件系统生成，不手写。三条不变量：站点源码（index.md、.vitepress/）不得出现裸计数（"N 篇"），数字只能来自computeStats()；每个一等内容文件必须在生成的侧边栏中恰好出现一次；仓库内禁止任何 symlink——Vite 会解引用public/下的软链、图片管线会读取链接目标，一条软链即可把仓库外文件带进公开产物（后两条经node .vitepress/sidebar.mjs --verify执行）。构建产物侧另有scripts/verify-dist.mjs断言页面与.md副本一一对应、副本链接可达、dist 无 symlink。站点脚手架不存在时 SKIP；node 不可用时 verify 半边 SKIP、裸计数 grep 照常执行
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 2mo ago First seen · 59 lines · 1,815 tokens per session scan A 10691915a665
harness-engineering AGENTS.md is an instructions file published in the GitHub repository deusyu/harness-engineering (6,250 stars, last pushed 6d ago), licensed MIT. It adds 1,815 tokens to every session, about $0.0073 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
harness-engineering AGENTS.md compared with instructions from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| harness-engineering AGENTS.mddeusyu/harness-engineering | Instructions file | 6.3k | 6d ago | MIT | A | 1,815 | Codex, OpenCode |
| vscode buildNext.instructions.mdmicrosoft/vscode | Instructions file | 195k | today | MIT | A | 9,719 | GitHub Copilot |
| next.js AGENTS.mdvercel/next.js | Instructions file | 143k | today | MIT | A | 7,461 | Codex, OpenCode |
| langchain AGENTS.mdlangchain-ai/langchain | Instructions file | 149k | today | MIT | A | 4,680 | Codex, OpenCode |
| vscode oss-third-party-notices.instructions.mdmicrosoft/vscode | Instructions file | 195k | today | MIT | A | 5,516 | GitHub Copilot |
| vscode copilot-instructions.mdmicrosoft/vscode | Instructions file | 195k | today | MIT | A | 3,093 | GitHub Copilot |
| deepseek-harness AGENTS.mddeepseek-ai/deepseek-harness | Instructions file | 274k | 2d ago | MIT | A | 4,313 | Codex, OpenCode |