Not in the top 1,000 by GitHub stars

ChangWenC/understanding-ladder

Karpathy's understanding ladder as agent skills: auto-pick text, diagram, interactive HTML, or video storyboard for any explanation. Nothing cut. English + 中文.

Claude Code Engineering

History

Review weekly changes to SKILL.md before updating. Dates mark the weeks in which we observed each version; checks apply to the recorded commits.

View weekly changes

`skills/plain-chinese/SKILL.md`

Week of Oct 26, 2026 · Line counts unavailable

Before Not checked yet → After Instructions only

First observation → 27b0957818f5

No earlier version was observed for comparison.

Compare on GitHub

`skills/understanding-ladder/SKILL.md`

Week of Oct 26, 2026 · Line counts unavailable

Before Not checked yet → After Instructions only

First observation → 27b0957818f5

No earlier version was observed for comparison.

Compare on GitHub

View on GitHub Add to my skills

Verified commit 27b0957818f57a9c1d0ecdde1f50c2e3752f972c

Version-pinned installation · Claude Code

git clone --filter=blob:none --no-checkout https://github.com/ChangWenC/understanding-ladder.git .hotskills/changwenc-understanding-ladder && git -C .hotskills/changwenc-understanding-ladder checkout --detach 27b0957818f57a9c1d0ecdde1f50c2e3752f972c
# Copy each selected skill directory from .hotskills/changwenc-understanding-ladder into .claude/skills/<skill-name>/

121GitHub stars

7stars in 7 days

2skills in this repository

0commits in 30 days

Heat

3.6

Star growth, last 7 daysNot available
Star growth, last 24 hoursNot available
Rank movement, last 7 daysNot available
Commits, last 30 daysNot available
Accounts asking about it through MCP, last 7 daysNot available
Hot votes this weekNot available

Heat never changes the order by stars, and sponsorship never touches it. How heat works

Stars over time

30-day change not yet available: not enough history.

Individual skills in this repository

plain-chinese skills/plain-chinese
understanding-ladder skills/understanding-ladder

How it works

Choose a skill

plain-chinese

When your agent uses it

用简明中文写解释、步骤和总结：一句一事、一词一义、主动句、不铺垫，内容一点不删。写中文的技术解释、操作步骤、学习笔记、报告，或者解释 AI/agent 的产出时使用。用户说「说人话」「写清楚点」「太绕了」「看不懂」「简明中文」「用 STE 的写法」「改得好懂一点」时也使用。受 ASD-STE100 启发，不是它的中文版。

What it loads into context

Frontmatter

119 tokens

Always available to the agent

SKILL.md

1,499 tokens

When the skill is activated

Reference documents

889 tokens

Loaded on request

Token counts are estimates. Scripts run separately; they are not counted as reference context.

The playbook

什么时候用
规则
强度
怎么改写
输出格式
交付前自检
限制

Files

../../.claude-plugin/marketplace.jsonasset557 bytes
../../.claude-plugin/plugin.jsonasset655 bytes
SKILL.mdcore5,493 bytes
references/examples.mddoc3,351 bytes

Read SKILL.md


# 简明中文（plain-chinese）

这个 skill 规定「怎么写」：用一套显式规则写中文，让读者读一遍就懂。

背景：英文有 ASD-STE100（简化技术英语）。在提示里写一句「in ASD-STE100」，模型就会调出整套写法。中文没有这样一个名字。所以本 skill 把规则写成清单。

核心原则只有一条：**简单指语言简单，不指内容变少。** 改写后可以比原文长。

## 什么时候用

- 写中文的技术解释、概念说明、操作步骤。
- 写学习笔记、复习材料、报告、总结。
- 解释 AI 或 agent 的产出：它做了什么、为什么、结果在哪。
- 用户要求改写一段难懂的中文。

View the complete source on GitHub

understanding-ladder

When your agent uses it

Pick the output format that makes an explanation easiest to understand, following Karpathy's understanding ladder — (1) plain controlled text, (2) diagram, (3) interactive HTML page, (4) 3Blue1Brown-style explainer video. Classify the content (definition/steps, process/causality, parameter→outcome, unfolding derivation), recommend the lowest rung that fully answers, then produce it. Works in English and Chinese. Use when the user says "explain this clearly", "help me understand", "draw a diagram", "make it interactive", "make an explainer video", "what's the best way to explain X", "understanding ladder", "Karpathy ladder", or 「讲清楚」「帮我理解」「画个图」「做个交互页」「做个讲解视频」「用哪种方式讲最好懂」「理解阶梯」; also when explaining a non-trivial concept, mechanism, algorithm, or an AI agent's output.

What it loads into context

Frontmatter

213 tokens

Always available to the agent

SKILL.md

1,828 tokens

When the skill is activated

Reference documents

2,367 tokens

Loaded on request

Token counts are estimates. Scripts run separately; they are not counted as reference context.

The playbook

The four rungs
Step 1: classify the content, pick a rung
Step 2: produce the rung
Demo mode: same content, every rung
Always append "What to check"
Do not

Files

SKILL.mdcore7,268 bytes
references/level-2-diagram.mddoc2,703 bytes
references/level-3-html.mddoc2,709 bytes
references/level-4-video.mddoc4,099 bytes

Read SKILL.md


# Understanding ladder

This skill decides **which format** to explain in. Writing quality inside every format is handled by level 1 rules (below).

Source: Andrej Karpathy, 2026-10-02 <https://x.com/karpathy/status/2105819303471976479>. LLM output is nearly free; human understanding is the bottleneck. So ask for formats that are easier to understand. He ranked four, each "even better" than the last — and each more expensive.

**Language:** reply in the user's language. These instructions are in English; the output is not.

## The four rungs

| Rung | Format | Why it is easier | Cost |
|---|---|---|---|
| 1 | Plain controlled text | One idea per sentence, one meaning per term; read once, no misreading | lowest |
| 2 | Diagram | Text is linear; a diagram lays the structure out at once | low |

View the complete source on GitHub

Skills included

plain-chinese
understanding-ladder

Repository facts

License: MIT
Created: Oct 4, 2026
Last pushed: Oct 6, 2026
Forks: 4
Open issues: 0
Project activity: Active

Before you install

Instructions only

Automated signals, not a guarantee

Checked Nov 4, 2026

No broad tool permissions
No dangerous shell commands
19 network referencesin 4 files: understanding-ladder/references/level-4-video.md:40 .claude-plugin/plugin.json:5 .claude-plugin/marketplace.json:3 and 1 more
No hidden or obfuscated text
No instruction overrides
No executable scripts
1 reference to secretsin 1 file: plain-chinese/references/examples.md:23
No requests to vote in Hot

Show all 20 signals by file

`.claude-plugin/marketplace.json`

Network accessline 3

"owner": {     "name": "ChangWenC",     "url": "https://github.com/ChangWenC"

Network accessline 4

"name": "ChangWenC",     "url": "https://github.com/ChangWenC"   },

Network accessline 5
```
"url": "https://github.com/ChangWenC"
```

`.claude-plugin/plugin.json`

Network accessline 5

"author": {     "name": "ChangWenC",     "url": "https://github.com/ChangWenC"

Network accessline 6

"name": "ChangWenC",     "url": "https://github.com/ChangWenC"   },

Network accessline 7
```
"url": "https://github.com/ChangWenC"
```

Network accessline 8

},   "homepage": "https://github.com/ChangWenC/understanding-ladder",   "repository": "https://github.com/ChangWenC/understanding-ladder",

Network accessline 9

"homepage": "https://github.com/ChangWenC/understanding-ladder",

Network accessline 10

"repository": "https://github.com/ChangWenC/understanding-ladder",

`skills/plain-chinese/references/examples.md`

Access to secretsline 23
```
[Credential-related source omitted]
```

`skills/understanding-ladder/SKILL.md`

Network accessline 10

Source: Andrej Karpathy, 2026-10-02 <https://x.com/karpathy/status/2105819303471976479>. LLM output is nearly free; human understanding is the bottleneck. So ask for formats that are easier to understand. He ranked four, each "even better"

Network accessline 84

- A storyboard: per shot — visual, narration (rung-1 rules), duration. - Tool links, e.g. [showtime](https://github.com/FavioVazquez/showtime) (open source, early; v0.4.0+ reads the storyboard table directly), [Manim](https://www.manim.comm

Network accessline 85

- Tool links, e.g. [showtime](https://github.com/FavioVazquez/showtime) (open source, early; v0.4.0+ reads the storyboard table directly), [Manim](https://www.manim.community/), local TTS [Kokoro](https://github.com/thewh1teagle/kokoro-onnx

`skills/understanding-ladder/references/level-4-video.md`

Network accessline 40

| Need | Tool | Notes | |---|---|---| | Let the coding agent produce the whole video | [showtime](https://github.com/FavioVazquez/showtime) | Open source, renders locally, no API keys. Early by the author's own description. **v0.4.0+ reads

Network accessline 41

|---|---|---| | Let the coding agent produce the whole video | [showtime](https://github.com/FavioVazquez/showtime) | Open source, renders locally, no API keys. Early by the author's own description. **v0.4.0+ reads this skill's storyboard

Network accessline 42

| Let the coding agent produce the whole video | [showtime](https://github.com/FavioVazquez/showtime) | Open source, renders locally, no API keys. Early by the author's own description. **v0.4.0+ reads this skill's storyboard table directly

Network accessline 43

| Hand-written math animation | [Manim Community](https://www.manim.community/) | Open-source library for 3Blue1Brown-style animation in Python |

Network accessline 44

| Free local voice | [Kokoro](https://github.com/hexgrad/kokoro) (ONNX: [kokoro-onnx](https://github.com/thewh1teagle/kokoro-onnx)) | Local TTS, offline. Listen to the voice for your language first |

Network accessline 45

| Paid high-quality voice | [ElevenLabs](https://elevenlabs.io/) | Needs an API key |

Network accessline 46

| Mux video, audio, captions | [ffmpeg](https://ffmpeg.org/) | |

How the signals fit together

Facts from a static read of the files at the pinned commit. They describe what the code does, not why.

Some files could not be read; the facts below still hold.

Signal confidence: 11 facts

Access to secrets in prose of documentationplain-chinese/references/examples.md:23Low confidence
Network access in prose of the SKILL.mdunderstanding-ladder/SKILL.md:10Low confidence
Network access in prose of the SKILL.mdunderstanding-ladder/SKILL.md:84Low confidence
Network access in prose of the SKILL.mdunderstanding-ladder/SKILL.md:85Low confidence
Network access in prose of documentationunderstanding-ladder/references/level-4-video.md:40Low confidence
Network access in prose of documentationunderstanding-ladder/references/level-4-video.md:41Low confidence
Network access in prose of documentationunderstanding-ladder/references/level-4-video.md:42Low confidence
Network access in prose of documentationunderstanding-ladder/references/level-4-video.md:43Low confidence
Network access in prose of documentationunderstanding-ladder/references/level-4-video.md:44Low confidence
Network access in prose of documentationunderstanding-ladder/references/level-4-video.md:45Low confidence
Network access in prose of documentationunderstanding-ladder/references/level-4-video.md:46Low confidence

Works with

Agent compatibility, from the files in this repository
Agent	Status	Evidence	Install to
Claude Code	Native	Manifest .claude-plugin/marketplace.json	`.claude/skills/<skill-name>/`
Codex	Compatible	Standard SKILL.md skills/plain-chinese/SKILL.md	`.agents/skills/<skill-name>/`
Cursor	Compatible	Standard SKILL.md skills/plain-chinese/SKILL.md	`.cursor/skills/<skill-name>/`
Gemini CLI	Compatible	Standard SKILL.md skills/plain-chinese/SKILL.md	`.gemini/skills/<skill-name>/`
GitHub Copilot	Compatible	Standard SKILL.md skills/plain-chinese/SKILL.md	`.github/skills/<skill-name>/`
OpenCode	Compatible	Standard SKILL.md skills/plain-chinese/SKILL.md	`.opencode/skills/<skill-name>/`

Native The repository ships a manifest, folder or frontmatter field for this agent.
Compatible A standard SKILL.md that this agent can read.
Unknown No evidence either way.

Derived automatically from paths, manifests and frontmatter. Not tested on each agent.

ChangWenC/understanding-ladder

History

skills/plain-chinese/SKILL.md

skills/understanding-ladder/SKILL.md

Share this page

Heat

Stars over time

Individual skills in this repository

How it works

plain-chinese

Frontmatter

SKILL.md

Reference documents

The playbook

Files

understanding-ladder

Frontmatter

SKILL.md

Reference documents

The playbook

Files

Skills included

Repository facts

Before you install

.claude-plugin/marketplace.json

.claude-plugin/plugin.json

skills/plain-chinese/references/examples.md

skills/understanding-ladder/SKILL.md

skills/understanding-ladder/references/level-4-video.md

How the signals fit together

Works with

You might also use

Astro-Han/karpathy-llm-wiki

coreyhaines31/marketingskills

karanb192/itr-wala

ComposioHQ/awesome-claude-skills

wondelai/skills

alexgreensh/anidoodle

`skills/plain-chinese/SKILL.md`

`skills/understanding-ladder/SKILL.md`

`.claude-plugin/marketplace.json`

`.claude-plugin/plugin.json`

`skills/plain-chinese/references/examples.md`

`skills/understanding-ladder/SKILL.md`

`skills/understanding-ladder/references/level-4-video.md`