Top AI Repos — open-source AI, indexed and scored

Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.

cnfjlhj/ai-collab-playbook

Practical AI collaboration playbook for research, writing, reading, and coding: article, prompts, agent rules, and reusable skills.

51No licencePython

View on GitHub Website

Stars: 419
Forks: 37
Watchers: 1
Open issues: 0
Contributors: —

Star history, last 90 days— over 2 snapshots

2026-07-31 · 4192026-08-06 · 419

Star history, last 90 days
Date	Stars
2026-07-31	419
2026-08-06	419

Today

— stars today

This week

— stars this week

This month

— stars this month

Momentum

0.0

growth rate 0.00%/day

From the README

<p align="center">
  <img src="docs/figs/hero-banner.png" width="100%" alt="AI Collab Playbook" />
</p>

<p align="center">
  <a href="docs/phd-ai-collab.md"><strong>读主文章</strong></a> · <a href="skills/full/README.md">Skills 目录</a> · <a href="prompts">Prompts</a> · <a href="README.en.md">English</a>
</p>

---

我是一名人工智能方向的在读博士生。从 GPT-3.5 时代开始，我就比较重度地使用 AI。几年下来，AI 已经渗透到我工作的大部分环节——科研、写作、编程、日常学习，甚至和朋友聊天时的零碎想法。这份手册就是这些经历的沉淀。

它不是"装哪些工具"的清单，也不是"怎么写 Prompt"的速成教程。我更想回答的是：当 AI 已经能进入这些场景以后，人应该怎样继续主导问题、判断质量、沉淀经验，并避免把理解能力外包出去。

**把 AI 当同事，不当工具；但人仍然是主变量。** AI 可以帮你探索、生成、执行，但问题表述、验收标准、必要取舍和最终判断不能交出去。否则效率越高，越可能只是更快地制造一种"我好像在推进"的幻觉。

## 先读什么

- **想看完整文章**：从 [`docs/phd-ai-collab.md`](docs/phd-ai-collab.md) 开始。
- **想快速抓主线**：先看下面三张图，再回到文章对应章节。
- **想复用工作流**：看 [`skills/full/README.md`](skills/full/README.md) 和 [`prompts/`](prompts)。
- **想看我怎么约束 Agent**：看 [`AGENTS.md`](AGENTS.md) / [`CLAUDE.md`](CLAUDE.md)。

## 这份手册在讲什么

**人是主变量。** AI 可以放大能力，但不能替代问题意识和判断力。你仍然需要知道自己要什么、什么算好、什么该放弃。

**当同事协作。** 让 AI 进入真实工作流，而不是停留在一次性问答窗口。划词、IM、远程 Agent——入口越贴近任务，越容易形成高频使用。

**低摩擦入口。** 把划词工具栏、IM 消息、远程 Agent 和知识库接到你日常的材料流里。不是每件事都值得拉起本地 Agent 走完整工作流，轻量任务就该轻量入口。

**上下文优先。** 先准备好目标、材料、偏好和验收标准，再让模型执行。没有上下文的 AI 输出，和掷骰子差别不大。

**经验沉淀。** 把有效流程固化为 Skill 或 Workflow，越用越好，越用越快——但也要定期做减法，不然 skill 膨胀本身就是一种噪音。

**反效率幻觉。** 警惕把理解、审美、取舍和学习过程一起外包给 AI。效率很高但不理解自己在做什么，比低效更危险。

[![AI 协作框架图](docs/figs/phd-ai-agent-framework.png)](docs/phd-ai-collab.md#code-agent-framework)

## 仓库内容

| 类别 | 入口 | 说明 |
|------|------|------|
| 主文章 | [`docs/phd-ai-collab.md`](docs/phd-ai-collab.md) | 完整方法论，2026-06-08 版 |
| 协作守则 | [`AGENTS.md`](AGENTS.md) / [`CLAUDE.md`](CLAUDE.md) | 我平时真在用的 AI 协作规则 |
| Prompts | [`prompts/`](prompts) | 提示词优化器、概念解释器、论文精读等模板 |
| 完整 Skills | [`skills/full/README.md`](skills/full/README.md) | 仓内所有 Skill 总目录 |
| 配图 | [`docs/figs/`](docs/figs) | 总览图、学习指南、框架图、路线图 |

## 独立维护的 Skills

这些 Skills 已经拆出去单独维护。它们不是全部推荐一次性安装，而是我在不同阶段沉淀过、可以按需参考的工作流：

| Skill | 用途 |
|-------|------|
| [paper-review-pipeline](https://github.com/cnfjlhj/paper-review-pipeline) | 论文审稿流水线 |
| [paperreview](https://github.com/cnfjlhj/paperreview) | 论文评审 |
| [skills-governance](https://github.com/cnfjlhj/skills-governance) | Skills 治理 |
| [session-recovery-codex](https://github.com/cnfjlhj/session-recovery-codex) | 会话恢复 |
| [collaborating-with-codex](https://github.com/cnfjlhj/collaborating-with-codex) | Codex 协作 |
| [completion-learn](https://github.com/cnfjlhj/completion-learn) | 任务完成后的三轴复盘：self → collaboration → tool |
| [xhs-note-creator](https://github.com/cnfjlhj/xhs-note-creator) | 小红书笔记创作 |
| [prompt-polisher](https://github.com/cnfjlhj/prompt-polisher) | 提示词润色 |
| [writing-anti-ai](https://github.com/cnfjlhj/writing-anti-ai) | 去 AI 味写作 |
| [xhs-longform-private-publisher](https://github.com/cnfjlhj/xhs-longform-private-publisher) | 小红书长文发布 |

## 反馈

- 随手留言 / 读后反馈：[Discussions](https://github.com/cnfjlhj/ai-collab-playbook/discussions/1)
- 勘误 / 结构建议：[提 Issue](https://github.com/cnfjlhj/ai-collab-playbook/issues/new/choose)
- 小红书转发版：[链接](https://www.xiaohongshu.com/discovery/item/69ab040f000000001a02d99e)

## 友链

- [linux.do](https://linux.do/)

---

<details>
<summary>Star History</summary>

<a href="https://www.star-history.com/?repos=cnfjlhj%2Fai-collab-playbook&type=date&legend=top-left">
  <picture>
    <source
      media="(prefers-color-scheme: dark)"
      srcset="https://api.star-history.com/image?repos=cnfjlhj/ai-collab-playbook&type=date&theme=dark&legend=top-left"
    />
    <source
      media="(prefers-color-scheme: light)"
      srcset="https://api.star-history.com/image?repos=cnfjlhj/ai-collab-playbook&type=date&legend=top-left"
    />
    <img
      alt="Star History Chart"
      src="https://api.star-history.com/image?repos=cnfjlhj/ai-collab-playbook&type=date&legend=top-left"
    />
  </picture>
</a>

</details>

Excerpt of 3,811 characters

Read on GitHub

Top contributors

No contributor data yet — it is fetched during the contributor sync pass.

Quality score

Would you bet a product on this? Bounded 0–100 and slow moving.

51/10051

Maintenance: 70%; How recently the default branch was pushed to
Releases: 25%; Recency and cadence of tagged releases
Community: 20%; Contributor breadth and bus factor
Issue backlog: 100%; Open issues measured against the size of the audience
Documentation: 81%; README depth, plus a homepage and topics
Licensing: 0%; How freely the code can be adopted

What to know before adopting

No tagged releases
No license — all rights reserved by default, not safe to reuse
No licence file — default copyright applies, all rights reserved.

Details

Created: 5 months ago
Last push: last month
Last release: none
Release tag: —
Releases in the last year: 0
Default branch: main
Size: 50k KB
Owner type: User
Owner location: —
Discovered via: phrase:agentic
First indexed: 7 days ago
Last synced: 6 days ago

cnfjlhj/ai-collab-playbook

Practical AI collaboration playbook for research, writing, reading, and coding: article, prompts, agent rules, and reusable skills.

51No licencePython

View on GitHub Website

Stars: 419
Forks: 37
Watchers: 1
Open issues: 0
Contributors: —

Star history, last 90 days— over 2 snapshots

2026-07-31 · 4192026-08-06 · 419

Star history, last 90 days
Date	Stars
2026-07-31	419
2026-08-06	419

Today

— stars today

This week

— stars this week

This month

— stars this month

Momentum

0.0

growth rate 0.00%/day

From the README

<p align="center">
  <img src="docs/figs/hero-banner.png" width="100%" alt="AI Collab Playbook" />
</p>

<p align="center">
  <a href="docs/phd-ai-collab.md"><strong>读主文章</strong></a> · <a href="skills/full/README.md">Skills 目录</a> · <a href="prompts">Prompts</a> · <a href="README.en.md">English</a>
</p>

---

我是一名人工智能方向的在读博士生。从 GPT-3.5 时代开始，我就比较重度地使用 AI。几年下来，AI 已经渗透到我工作的大部分环节——科研、写作、编程、日常学习，甚至和朋友聊天时的零碎想法。这份手册就是这些经历的沉淀。

它不是"装哪些工具"的清单，也不是"怎么写 Prompt"的速成教程。我更想回答的是：当 AI 已经能进入这些场景以后，人应该怎样继续主导问题、判断质量、沉淀经验，并避免把理解能力外包出去。

**把 AI 当同事，不当工具；但人仍然是主变量。** AI 可以帮你探索、生成、执行，但问题表述、验收标准、必要取舍和最终判断不能交出去。否则效率越高，越可能只是更快地制造一种"我好像在推进"的幻觉。

## 先读什么

- **想看完整文章**：从 [`docs/phd-ai-collab.md`](docs/phd-ai-collab.md) 开始。
- **想快速抓主线**：先看下面三张图，再回到文章对应章节。
- **想复用工作流**：看 [`skills/full/README.md`](skills/full/README.md) 和 [`prompts/`](prompts)。
- **想看我怎么约束 Agent**：看 [`AGENTS.md`](AGENTS.md) / [`CLAUDE.md`](CLAUDE.md)。

## 这份手册在讲什么

**人是主变量。** AI 可以放大能力，但不能替代问题意识和判断力。你仍然需要知道自己要什么、什么算好、什么该放弃。

**当同事协作。** 让 AI 进入真实工作流，而不是停留在一次性问答窗口。划词、IM、远程 Agent——入口越贴近任务，越容易形成高频使用。

**低摩擦入口。** 把划词工具栏、IM 消息、远程 Agent 和知识库接到你日常的材料流里。不是每件事都值得拉起本地 Agent 走完整工作流，轻量任务就该轻量入口。

**上下文优先。** 先准备好目标、材料、偏好和验收标准，再让模型执行。没有上下文的 AI 输出，和掷骰子差别不大。

**经验沉淀。** 把有效流程固化为 Skill 或 Workflow，越用越好，越用越快——但也要定期做减法，不然 skill 膨胀本身就是一种噪音。

**反效率幻觉。** 警惕把理解、审美、取舍和学习过程一起外包给 AI。效率很高但不理解自己在做什么，比低效更危险。

[![AI 协作框架图](docs/figs/phd-ai-agent-framework.png)](docs/phd-ai-collab.md#code-agent-framework)

## 仓库内容

| 类别 | 入口 | 说明 |
|------|------|------|
| 主文章 | [`docs/phd-ai-collab.md`](docs/phd-ai-collab.md) | 完整方法论，2026-06-08 版 |
| 协作守则 | [`AGENTS.md`](AGENTS.md) / [`CLAUDE.md`](CLAUDE.md) | 我平时真在用的 AI 协作规则 |
| Prompts | [`prompts/`](prompts) | 提示词优化器、概念解释器、论文精读等模板 |
| 完整 Skills | [`skills/full/README.md`](skills/full/README.md) | 仓内所有 Skill 总目录 |
| 配图 | [`docs/figs/`](docs/figs) | 总览图、学习指南、框架图、路线图 |

## 独立维护的 Skills

这些 Skills 已经拆出去单独维护。它们不是全部推荐一次性安装，而是我在不同阶段沉淀过、可以按需参考的工作流：

| Skill | 用途 |
|-------|------|
| [paper-review-pipeline](https://github.com/cnfjlhj/paper-review-pipeline) | 论文审稿流水线 |
| [paperreview](https://github.com/cnfjlhj/paperreview) | 论文评审 |
| [skills-governance](https://github.com/cnfjlhj/skills-governance) | Skills 治理 |
| [session-recovery-codex](https://github.com/cnfjlhj/session-recovery-codex) | 会话恢复 |
| [collaborating-with-codex](https://github.com/cnfjlhj/collaborating-with-codex) | Codex 协作 |
| [completion-learn](https://github.com/cnfjlhj/completion-learn) | 任务完成后的三轴复盘：self → collaboration → tool |
| [xhs-note-creator](https://github.com/cnfjlhj/xhs-note-creator) | 小红书笔记创作 |
| [prompt-polisher](https://github.com/cnfjlhj/prompt-polisher) | 提示词润色 |
| [writing-anti-ai](https://github.com/cnfjlhj/writing-anti-ai) | 去 AI 味写作 |
| [xhs-longform-private-publisher](https://github.com/cnfjlhj/xhs-longform-private-publisher) | 小红书长文发布 |

## 反馈

- 随手留言 / 读后反馈：[Discussions](https://github.com/cnfjlhj/ai-collab-playbook/discussions/1)
- 勘误 / 结构建议：[提 Issue](https://github.com/cnfjlhj/ai-collab-playbook/issues/new/choose)
- 小红书转发版：[链接](https://www.xiaohongshu.com/discovery/item/69ab040f000000001a02d99e)

## 友链

- [linux.do](https://linux.do/)

---

<details>
<summary>Star History</summary>

<a href="https://www.star-history.com/?repos=cnfjlhj%2Fai-collab-playbook&type=date&legend=top-left">
  <picture>
    <source
      media="(prefers-color-scheme: dark)"
      srcset="https://api.star-history.com/image?repos=cnfjlhj/ai-collab-playbook&type=date&theme=dark&legend=top-left"
    />
    <source
      media="(prefers-color-scheme: light)"
      srcset="https://api.star-history.com/image?repos=cnfjlhj/ai-collab-playbook&type=date&legend=top-left"
    />
    <img
      alt="Star History Chart"
      src="https://api.star-history.com/image?repos=cnfjlhj/ai-collab-playbook&type=date&legend=top-left"
    />
  </picture>
</a>

</details>

Excerpt of 3,811 characters

Read on GitHub

Top contributors

No contributor data yet — it is fetched during the contributor sync pass.

Quality score

Would you bet a product on this? Bounded 0–100 and slow moving.

51/10051

Maintenance: 70%; How recently the default branch was pushed to
Releases: 25%; Recency and cadence of tagged releases
Community: 20%; Contributor breadth and bus factor
Issue backlog: 100%; Open issues measured against the size of the audience
Documentation: 81%; README depth, plus a homepage and topics
Licensing: 0%; How freely the code can be adopted

What to know before adopting

No tagged releases

Details

Created: 5 months ago
Last push: last month
Last release: none
Release tag: —
Releases in the last year: 0
Default branch: main
Size: 50k KB
Owner type: User
Owner location: —
Discovered via: phrase:agentic
First indexed: 7 days ago
Last synced: 6 days ago

Top AI Repos — open-source AI, indexed and scored

cnfjlhj/ai-collab-playbook

From the README

Top contributors

Quality score

What to know before adopting

Categories

Details

cnfjlhj/ai-collab-playbook

From the README

Top contributors

Quality score

What to know before adopting

Categories

Details

Top AI Repos — open-source AI, indexed and scored

Trend

From the README

Top contributors

Quality score

What to know before adopting

Categories

Details

Trend

From the README

Top contributors

Quality score

What to know before adopting

Categories

Details