Munger Decision Free
v1.2.4Charlie Munger's mental model decision assistant. Analyzes your decision scenario, recommends the most relevant thinking models, and guides you through struc...
Security Scan
OpenClaw
Benign
high confidencePurpose & Capability
Name, description, SKILL.md, and included code/data (detector, recommender, dialogue, reporter, model data and reference docs) align: this is a local decision-assistant implemented in Node/TypeScript and no capabilities appear unrelated to that purpose.
Instruction Scope
SKILL.md limits runtime behavior to scenario detection, model recommendation, guided Q&A, and report generation. Example commands and API usage target the packaged code; there are no instructions to read unrelated system files, access credentials, or post data to unexpected external endpoints (other than an optional upgrade link in README).
Install Mechanism
The skill is instruction-only (no install spec) which is lower-risk. However the bundle includes full source, package.json, and a large package-lock — if you run npm install locally you will pull many dev packages from the npm registry. Nothing in the manifest points to downloads from untrusted URLs, but running npm install would be the main friction point if you choose to execute the code.
Credentials
The skill declares no required environment variables, no credentials, and no special config paths. SKILL.md mentions using an LLM as a fallback in detection logic, but this appears to rely on the agent/platform LLM (no keys requested).
Persistence & Privilege
The skill does not request always:true and does not declare persistent system-wide config or cross-skill modifications. It is user-invocable and may be autonomously invoked per platform defaults (normal behavior).
Assessment
This package appears to be a coherent, local Node/TypeScript decision-assistant. Before installing or executing any code: 1) inspect src/index.ts and any network/HTTP calls (none were obvious in provided excerpts) to confirm no unexpected endpoints; 2) if you run npm install, do so in a sandbox/container because the package-lock will pull many dev packages; 3) note the README contains a Gumroad upgrade link (for paid full version) — that is a marketing link, not required for the free skill; 4) if you only want to use the skill via the platform, no credentials are required. If you want extra assurance, ask for a quick scan of src/index.ts and any files that perform network I/O so I can point out any outgoing connections or data‑export behavior.Like a lobster shell, security has layers — review code before you run it.
Runtime requirements
🧠 Clawdis
Binsnode
latest
芒格决策助手
将查理·芒格的核心思维模型转化为可执行的决策工具。
当前版本: 免费版(12 个核心模型)
完整版: 83 个模型 + 83 个详细文档(即将推出)
🎯 核心功能
- 场景识别: 自动识别决策场景(投资、产品、人员、战略)
- 模型推荐: 根据场景推荐 3-5 个相关思维模型
- 引导式分析: 通过结构化问题引导用户思考
- 决策报告: 生成 Markdown 格式的分析报告
📦 免费版包含的 12 个核心模型
核心决策模型(8 个)
- 第一性原理 - 回归事物本质
- 机会成本 - 评估替代方案
- 复利效应 - 长期价值积累
- 能力圈 - 了解自己的边界
- 逆向思维 - 反过来想,避免失败
- 护城河 - 竞争优势分析
- 安全边际 - 容错空间评估
- 多元思维模型 - 跨学科交叉验证
心理学模型(4 个)
- 确认偏误 - 避免选择性证据
- 锚定效应 - 警惕第一印象
- 损失厌恶 - 理性评估风险
- 社会认同 - 独立思考,避免从众
🚀 使用方法
命令行
# 开始决策分析
/munger analyze 是否应该投资中宠股份
# 查看所有模型
/munger models
# 查看历史记录
/munger history
代码调用
import { assistant } from './src/index';
// 开始分析
const response = await assistant.startAnalysis('session-123', '是否应该投资中宠股份');
console.log(response);
// 处理回答
const next = await assistant.handleAnswer('session-123', '7分,我对行业有一定了解');
console.log(next);
🏗️ 架构设计
模块划分
src/
├── index.ts # 主入口
├── detector.ts # 场景识别器
├── recommender.ts # 模型推荐引擎
├── dialogue.ts # 对话管理器
├── reporter.ts # 报告生成器
└── types.ts # 类型定义
data/
└── munger-knowledge.js # 12 个核心模型库
核心流程
用户输入
↓
场景识别(detector)
↓
模型推荐(recommender)
↓
多轮对话(dialogue)
↓
报告生成(reporter)
🧪 测试
npm test
🎓 芒格模型应用
第一性原理
核心问题: 用户真正需要什么?
- 不是"学习芒格模型"
- 而是"做出更好的决策"
设计推导:
- 最小化学习成本 → 自动场景识别
- 结构化思考 → 引导式问题
- 可执行建议 → 决策报告
逆向思维
什么会导致失败?
- 场景识别不准 → 关键词 + 正则 + LLM 兜底
- 问题太学术 → 白话文 + 实际案例
- 报告太长 → 控制在 1 页内
能力圈
我们擅长:
- ✅ Node.js + TypeScript
- ✅ 状态机设计
- ✅ Markdown 生成
🎁 完整版(即将推出)
完整版将包含:
83 个思维模型
- 核心模型(10 个): 多元思维、机会成本、复利效应等
- 心理学模型(29 个): 确认偏误、锚定效应、损失厌恶等
- 经济学模型(15 个): 供需关系、边际效用、规模经济等
- 数学模型(12 个): 概率论、贝叶斯、幂律分布等
- 物理学模型(8 个): 临界质量、杠杆原理、惯性等
- 生物学模型(9 个): 进化论、适者生存、生态位等
83 个详细文档
- 每个模型配有独立 Markdown 文档
- 包含定义、案例、应用场景、常见误区
- 总计超过 50,000 字的深度内容
高级功能
- 智能评分算法(基于 LLM 分析)
- 历史记录持久化
- 自定义模型库
- 批量决策分析
敬请期待!
📄 许可证
MIT
开发者: edu-dev
版本: v1.2.0-free
日期: 2026-03-29
Comments
Loading comments...
