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, strate
Install
npx skills add Renhuai123/world-model-method --skill world-model-methodgit clone --depth 1 https://github.com/Renhuai123/world-model-method.git /tmp/world-model-method
mkdir -p ~/.claude/skills
cp -r /tmp/world-model-method/. ~/.claude/skills/world-model-methodPersonal skills live in ~/.claude/skills. Use .claude/skills in a project folder to share a skill with your team.
From the SKILL.md
# 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：
Details
- Skill name
world-model-method- Repository
- Renhuai123/world-model-method
- Repo updated
- 4 June 2026
FAQ
What does the World Model Method 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, strate
How do I install the World Model Method skill?
skills CLI (Claude Code, Cursor, Codex and more): npx skills add Renhuai123/world-model-method --skill world-model-method. Skills go in ~/.claude/skills (personal) or .claude/skills (project), and Claude loads them automatically when a task matches the description.
Is World Model Method popular?
It lives in Renhuai123/world-model-method, which has 49 GitHub stars and 8 forks.
Related skills
Conference Watch
67Zsun79/ConferenceWatch
Watch and report on upcoming AI/ML conferences: submission deadlines (every stage), dates, location, special sessions / calls for papers, official links, and a 5-year acceptance-rate trend. Use when the user asks about future AI conference deadlines, wants to plan submissions, compare venues, or track a set of conferences by area/reputation/difficulty. Works in any agent that has a web-search / web-fetch capability (Claude Code, Codex, etc.).
Handoff
43ToolMonsters/handoff-skill
Generate a complete handoff document of the current conversation so another LLM (ChatGPT, a fresh Claude, Gemini...) can pick up exactly where this one left off. Use when the user types "handoff", says the session is almost full, mentions hitting the context/usage limit, or wants to continue this conversation in another AI.
Humanize Chinese
29swaylq/humanize-chinese
中文去 AI 腔一整套（装完即用，零 API Key —— 装了本 skill 的 agent 自己就是那个 LLM）。 六个流程：① 生成即去 AI —— 写中文时按本 skill 的写作指南落笔，初稿就不带 AI 腔； ② 复查改写 —— 产出后对照本 skill 的清单再过一遍，定点拆模板句式； ③ Python 词语替换 —— ./humanize replace 按文体自动路由词表（学术 120+ 条 / 通用 220+ 条），只做短语级、只在子句边界替换，改完自动核对数字/专名/段落； ④ Claude 腔转中文 —— 把 Claude 助手腔的中文改写成平实中文， 规则提炼自 gvzdv/claudish-to-english 并本地化； ⑤ 去水印 —— ./humanize watermark 扫清零宽字符/双向控制符/同形替身， 并对照 guillaumemeyer/watermarks-remover 说明统计水印的边界； ⑥ 纠错顺句 —— 修错词、病句、标点，拗口和用词不准处轻手顺一顺；改动范围有硬校验。 另有检测（0-100 分）、风格转换、学术降重等 CLI。 Use when user says: "去AI味", "降AIGC", "人性化文本", "humanize chinese", "AI检测", "AIGC降重", "去除AI痕迹", "文本
Cnki Aigc Rewrite
51qingshanliuci/cnki-aigc---skill
改写中文学术论文中被知网AIGC检测器标为"红色显著"的段落，目标是将AI率落出红色区间且不改变事实/数字/术语/引用。基于实战将20.6%降至10.1%（净降10.5个百分点）的工作流提炼。使用场景：用户提到"知网AIGC""降AI率""红色显著""降重"或贴出AIGC检测报告片段时触发。
Math Modeling Skill
55anticipate218/math-modeling-skill
数学建模竞赛全流程助手：审题选题、模型选型与求解、结果检验（误差/灵敏度/稳健性）、论文写作与提交自检， 覆盖国赛 CUMCM、研赛华为杯、美赛 MCM/ICM。当用户准备建模竞赛、写建模论文、问赛题思路、写摘要、 做灵敏度分析、选模型、检查论文格式或提交前自检时使用——即使没说"数学建模"三字也应使用。
Prompt Refine
52Apeironics/prompt-refine-skill
Silently restructures the user's natural-language prompt into the format the model CURRENTLY running this skill handles best, then answers. On activation it identifies which model family is executing it (Claude, GPT, Gemini, Llama, DeepSeek, Mistral, Qwen, Grok, Perplexity, Kimi, GLM, Command, Nova, or Phi) and loads that one model's official strategy — so the optimization always matches the model that actually runs it. Activate with /prompt-refine. Use when users want better answers without learning prompt engineering.
Browse the most popular Claude skills, all ai, llms & agents skills, or write your own with the skills generator.