World Model Method logo

World Model Method

Community
Renhuai123
world-model-method

A general-purpose working mode that makes the assistant produce output the "world-model way" — define an explicit goal and cost, build an abstract model of the situation, predict the consequences of several candidate approaches, search/optimize for the best full plan, and only then render the concrete output — instead of generating it token-by-token off the top. This operationalizes Yann LeCun's objective-driven / world-model / planning ideas (JEPA, "A Path Towards Autonomous Machine Intelligence") as a reusable prompt-level method for ANY non-trivial task: important writing, decisions, strategy, problem-solving, system design, planning. Trigger when the user says "用世界模型工作法 / 按 LeCun 那套来做 / 别顺嘴生成,给我规划出来的 / 想清楚再答 / objective-driven / plan it, don't just write it", or when a task is complex enough that deliberation and comparing alternatives clearly beats a first-draft answer.

Overview

PublisherRenhuai123
Repositoryworld-model-method
Skill nameworld-model-method
Stars
50
Forks
8
Bundled files
1
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • 1 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

    Published by Renhuai123 on GitHub. Read the source before you install it.

Installation

Install the World Model Method AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/Renhuai123/world-model-method/tree/main
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/Renhuai123/world-model-method.git \
  .claude/skills/world-model-method
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable World Model Method in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads World Model Method on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so World Model Method is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

世界模型工作法 (World-Model Method)

默认的大模型是"一个 token 接一个 token 顺嘴生成"。这个 skill 让 AI 换一种产出方式——像 LeCun 设想的自主智能那样:先理解情境、定目标,在抽象层预测后果、搜索规划出整条路径,最后才渲染成具体产出。

一句话区别:产出是**"规划出来的",不是"生成出来的"**。


什么时候开这个模式

  • ✅ 用:复杂、后果重要、有多条可行方案、值得想清楚的活——写重要文案/方案、做决策、定策略、解难题、设计系统、写关键代码。
  • ❌ 不用:闲聊、查事实、简单改写、赶时间的小活。这是 System-2,慢而稳,杀鸡不必用牛刀。

核心循环(把 LeCun 的自主智能架构翻成 6 步工作法)

LeCun 在《A Path Towards Autonomous Machine Intelligence》(2022) 里设想的智能体有几个模块:感知 → 世界模型 → 成本 → 行动者 → 配置器。把它翻译成你能直接用的步骤:

  1. 定目标 + 成本函数(Objective & Cost) 先问:"成功长什么样?怎么算好/坏?硬约束是什么?" 写下可衡量的目标和要最小化的代价。没有目标就没有方向感——这正是 LeCun 说大模型缺的东西。

  2. 感知 + 建世界模型(Perceive & Model) 在抽象层把情境拆成:关键要素、它们的关系、动态怎么变、有哪些约束。理解这件事的"物理规则"。这一步不产出任何成品。

  3. 预测后果(Predict) 生成 2–4 条候选路径/方案,对每一条在脑内推演:"它会导向什么结果?哪里会崩?代价多大?" 这是世界模型的核心动作——预测行为的后果。

  4. 搜索 + 规划(Search & Plan) 不要拿第一个顺手的想法就跑。 在方案空间里按成本函数比选,挑出整条最优路径。复杂任务分层:先定高层骨架,再细化每一步(对应 H-JEPA 的分层规划)。

  5. 渲染(Render) 到这一步才把选中的抽象方案展开成具体产出(文章 / 方案 / 代码 / 决策)。

  6. 校验(Verify) 对照第 1 步的目标和约束自检。不达标?回到第 3 或第 4 步重来,而不是硬把次品交出去。


让它区别于普通"想一步步来"的 5 条原则

  1. 目标驱动:永远先有显式的目标 + 成本,再动手。
  2. 表征先于细节:先在概念层规划,别一上来抠字句——这是"预测表征、不预测像素"的工作版。
  3. 比选 > 续写:至少预测并比较 2 条路径,而不是顺着第一个想法写到底。
  4. 渲染放最后:成品 = "被选中方案的渲染",且和规划过程显式分开。
  5. 硬约束当护栏:不可逾越的约束在第 1 步就钉死,当架构级护栏,而不是事后补丁。

默认输出格式

先给一段"规划",再给成品,让"规划出来的"看得见:

【目标 & 成本】成功标准 / 要最小化什么 / 硬约束
【世界模型】关键要素 · 关系 · 动态 · 约束(抽象、精炼,不堆细节)
【候选路径 & 后果预测】
  方案 A → 预测后果 / 代价
  方案 B → 预测后果 / 代价
  (方案 C …)
【选择】选哪条 + 为什么(对照成本函数)
──────── 以下是渲染出的成品 ────────
<具体产出>
【自检】对照目标/约束过没过;不过则指出并修正

赶时间、或用户只要成品时,可把规划部分压成几行——但六步不能跳,区别就在这。


一个例子

见 examples/example-content-strategy.md:同一个"策划一条短视频选题"的任务,对比"默认顺嘴生成" vs "世界模型工作法",看产出差在哪。


诚实说明

  • 这是把 LeCun 的目标驱动 / 世界模型 / 规划思路翻译成提示词层面的工作法——改的是"怎么用 AI",不是真的换了 JEPA 神经网络架构。别把它当成"我现在跑的是世界模型"。
  • 它拿深思和比选换速度:好处是少踩坑、方案更优;代价是慢、啰嗦。值不值,看任务大小。
  • 它不保证答案正确,只保证想得更结构化——复杂任务里这通常就够把次品挡在门外了。
  • 参考:Yann LeCun, A Path Towards Autonomous Machine Intelligence (2022);Meta V-JEPA 2 (2025)。

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the World Model Method AI skill do?

A general-purpose working mode that makes the assistant produce output the "world-model way" — define an explicit goal and cost, build an abstract model of the situation, predict the consequences of several candidate approaches, search/optimize for the best full plan, and only then render the concrete output — instead of generating it token-by-token off the top. This operationalizes Yann LeCun's objective-driven / world-model / planning ideas (JEPA, "A Path Towards Autonomous Machine Intelligence") as a reusable prompt-level method for ANY non-trivial task: important writing, decisions, str...

Why use World Model Method on TypingMind?

Because you install it once and use it with any model. World Model Method is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install World Model Method in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/Renhuai123/world-model-method/tree/main. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use World Model Method?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with World Model Method?

As many as you like. As long as a model supports skills, you can use World Model Method with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the World Model Method AI skill free?

It is published on GitHub by Renhuai123. Check the repository for licensing terms. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇