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/.claude/skills/curate-research/SKILL.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/skills/deusyu/harness-engineering/curate-research)<a href="https://agentmods.dev/skills/deusyu/harness-engineering/curate-research"><img src="https://agentmods.dev/badge/skills/deusyu/harness-engineering/curate-research/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/skills/deusyu/harness-engineering/curate-research"><img src="https://agentmods.dev/badge/skills/deusyu/harness-engineering/curate-research.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
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 | $0.00116 | $0.02434 |
| Opus 5.5 | $0.00046 | $0.00974 |
| Sonnet 5.5 | $0.00023 | $0.00487 |
| Haiku 4.5 | $0.00012 | $0.00243 |
Grade A, and why
curate-research scanned grade A with 1 finding 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.
Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
- [ ] **数图要抓原文 HTML 数,不能只看渲染出来的正文。** 2026-07-27 的教训:一篇 claude.com 译文声明 `sourceFigureCount: 0`,而原文 HTML 里有 4 个 `<figure>`——抓取工具吐出的 markdown 把它们丢了,人只看那份 markdown 就会以为没有图。可用 `curl <url> | grep -c '<figure'` 交叉验证 How it starts
The opening of the file, as written. The whole thing — 98 lines — stays where its author put it; the contents beside it link to each section on GitHub.
curate-research —— 仓库自我策展 harness
一个讲 Harness Engineering 的仓库,用一个 harness 来策展自己。本 skill 把外部调研候选可控地整合进 tracked 档案。
核心约束:评审全自动;收录是人类闸门,必须停下来和用户来回讨论后才动
works/与articles.md。
何时用
- 用户给出一批候选 URL(文章 / 论文 / 工具 / 项目),希望整合进仓库
- 用户说「处理 / 收录 / 整理」
translate/暂存区的翻译 - 已有
translate/<batch>/works-ready/候选译文,要决定去向
输入与产物路径
暂存区(gitignored,本地过程区):translate/<batch>/
sources/<slug>/source.md(原文快照;论文额外抓source-full.md)translations/<slug>/{01-analysis,02-prompt,translation}.md(过程三件套)works-ready/<slug>-translation.md(发布候选)+works-ready/README.md(状态表,本批权威)
正式档案(tracked):works/<slug>-translation.md + references/articles.md 条目。
流水线 6 阶段
①抓取 ──②翻译 ──③评审[全自动] ──🚧人类闸门🚧──④收录 ──⑤校验 ──⑥清理
① 抓取
用 baoyu-url-to-markdown 把每个 URL 存到 sources/<slug>/source.md。论文/长文必须额外抓 HTML 全文到 source-full.md(否则只有摘要页,C8 会拦谎报)。
② 翻译
按 baoyu-translate 配置生成 translations/<slug>/ 三件套 → works-ready/<slug>-translation.md。01-analysis.md 要含收录建议;source-full.md 存在时,分析稿不得声称「仅摘要页」。
③ 评审(全自动并行扇出)
对每篇候选派一个评审 agent(一批 3–4 篇,并行多个 agent)。统一打分维度,吐结构化定性。标准评审 prompt 模板:
你是 Harness Engineering 中文知识库的内容评审。仓库主题:人类设计约束与反馈回路、AI agent 写代码。 读
sources/<slug>/source.md(+source-full.md)、works-ready/<slug>-translation.md、translations/<slug>/01-analysis.md,逐篇回答:
- 原文价值:原创洞察密度 / 长文实质 vs 产品页·发布稿·摘要。高/中/低
- 翻译质量:完整逐译 / 压缩摘要 / 首轮粗稿;通顺度、术语到位度。精品/合格/需返工
- 与仓库契合度:补薄弱环节还是重复
- 一句话定性 + 建议去向:works/ 正式收录 / articles.md 观察项一行 / tools/ / 暂不收录 基于实际内容,紧凑中文,结构化输出即为最终产出。
汇总成一张「候选 × 定性 × 去向」表。
🚧 人类闸门(不可跳过)
把评审表交给用户,来回讨论收录决策。精品 vs 边角的边界判断、是否返工、哪些合并成专题——都由人定。未经用户确认,不得进入 ④。
④ 收录(按减熵分流规则)
- 实质原创长文 / 论文 →
works/:cp到works/<slug>-translation.md,在articles.md加一个### N.编号条目(脉络一末尾,旧编号顺移),并同步全部计数缓存(见下)。 - 边角材料(产品页 / README / 发布稿 / 短 bliki)→ articles.md「观察项 / 候选材料(不计入文章数)」表一行,链上游 URL,不做编号条目(文章数不变)。
- 工具类 → 只在用户实际用过后才进
tools/;未实测前在观察项表标 🔵「待实测后入 tools/」。 - 需返工的长文:先润色(事实核查加译者注 / 术语统一)再进
works/。
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 · 98 lines · 116 tokens per session scan A f098d6151ae8
curate-research is a skill published in the GitHub repository deusyu/harness-engineering (6,250 stars, last pushed 6d ago), licensed MIT. It adds 116 tokens to every session and 2,434 once invoked, about $0.0005 per session on Opus 5.5. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
curate-research compared with skills from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| curate-researchdeusyu/harness-engineering | Skill | 6.3k | 6d ago | MIT | A | 116 | Claude Code |
| imagegenopenai/codex | Skill | 133k | yesterday | Apache-2.0 | A | 113 | Codex |
| playwright-climicrosoft/playwright | Skill | 99k | yesterday | Apache-2.0 | A | 19 | Claude Code |
| local-ai-agentsmicrosoft/ai-agents-for-beginners | Skill | 79k | 4d ago | MIT | A | 200 | Claude Code, Codex |
| systematic-debuggingobra/superpowers | Skill | 313k | 3d ago | MIT | A | 21 | Claude Code |
| next-cache-components-adoptionvercel/next.js | Skill | 143k | today | MIT | A | 95 | Claude Code |
| next-partial-prefetching-adoptionvercel/next.js | Skill | 143k | today | MIT | A | 103 | Claude Code |