Maps an industry onto visual terrain where you can see value flows like rivers, bottlenecks as narrow passes, and profit capture points as treasure piles. Built on Hawkins' reference frame theory from A Thousand Brains. The core insight: where flow gets constrained (ASML's chokepoint, regulatory gates) often isn't where money accumulates (platform rent, brand premiums), and making that gap visible is the point. Comes with three base rates to calibrate your priors and three unresolved questions at the frontier. Outputs a rendered PNG card with an embedded map in either Makoto Yoshida storybook style or dark cyberpunk pixel art. Useful when you're trying to understand power structure in an industry you're entering or where someone claims disruption is coming.

skills.sh installs
1.2K
GitHub stars
7.7K
last updated
Aug 21, 2026

Install

$ npx -y skills add lijigang/ljg-skills --skill ljg-map --agent claude-code

Installs into .claude/skills of the current project.

Files

ljg-map · 形式化思考地图

用《学习观》关于概念、判别模型和联结模型的理解,看到文稿背后的知识结构:哪些词有明确的所指,现实对象凭什么归入它们,概念之间的关系在什么条件下成立。把看到的结构画出来,形式化负责减少歧义,白话和实例负责让关系可理解、可使用。

Workflow Routing

请求执行
从文稿生成地图,或修订已有地图阅读 Workflows/Map.md
需要校准公式密度、节点解释和来源边界按需阅读 Examples.md;学习表达方式,不套用题材和节点

Gotchas

  • 知识观是洞察视角,图形由材料决定。 具象、概念与应用的对应帮助检查知识是否接通,不构成固定的三个图层、三个栏目或必经步骤。分类交叉就画交叉关系,反馈就画循环;材料确有层级时才画层级。
  • 节点来自当前文稿。 不把任何文章都画成「经验—抽象—概念—记忆—学习」;只有文章本身讨论这些概念时,它们才成为内容节点。判别、验证可以写在边或节点解释里,无须一律占框。
  • 地图不是换成公式的长文。 成品先让读者看见承重关系,再解释图里的节点。不能只给概念清单、长表格或一串没有地图的公式。
  • 形式化不以符号数量计。 优先「中文名称 + 短公式」。定义符号后再使用,不另造一套图论元语言来描述本来很简单的文章。
  • 分类与预测分开。 「有圆圈」可以定义一类;「这一类会加分」是另一条需要依据的关系。满足分类条件不自动证明预测成立,也不能推出反面类别有反面结果。
  • 箭头有语义。 定义、归类、计算、相关、因果、类比、推断和反馈不能混作一种关系;用短边标签和必要条件区分。猜想仍标猜想。
  • 图的完整不能靠编造。 文稿缺少判别标准、实例或验证办法,就在对应位置标缺口。补充的教学例子、应用设想与可行性判断要与原文陈述分清。

成品约定

默认直接在对话中交付 Markdown、Mermaid 和可读公式。用户要求保存、特定格式或修改已有产物时,按其指定交付;本技能不默认写笔记、覆盖原文或发布。

主图显出这篇文稿特有的知识关系,不预设层数、形状或统一应用流程。主要关系保持易读,次要分支可以放到有连接位置的补图或标记表。每个图中节点均能在解释中找到相同编号或名称,以及定义、公式、语义和具体对应。

Examples

Sponsors

Details

Repository
github.com/lijigang/ljg-skills
Category
Data Science & ML
First seen
Jul 14, 2026

More from lijigang/ljg-skills

All 25
1qodex-ai avatarlegal-document-analyzerqodex-ai/ai-agent-skillsBuild agents for legal document analysis, contract review, and compliance checking. Handles document parsing, risk identification, and legal research. Use when 1.2Kskills.sh installs
2ruvnet avatarswarm-advancedruvnet/rufloAdvanced swarm orchestration patterns for research, development, testing, and complex distributed workflows1.2Kskills.sh installs
3duneanalytics avatarduneduneanalytics/skillsDune CLI for querying blockchain and on-chain data via DuneSQL, searching decoded contract tables, managing saved queries, managing visualizations, managing das1.2Kskills.sh installs
4brave avatarbxbrave/brave-search-skillsUSE FOR web search, research, RAG, grounding, browse, find, lookups, fact-checking, documentation, agentic AI. All-in-one, optimized for AI agents. Pre-extracte1.2Kskills.sh installs
5himself65 avatarfunda-datahimself65/finance-skillsQuery Funda AI financial data via two surfaces: the MCP server at https://funda.ai/api/mcp for analyst-grade research synthesis (DCF, comps, earnings previews/r1.2Kskills.sh installs
6kostja94 avatarai-traffic-trackingkostja94/marketing-skillsWhen the user wants to track AI search traffic in GA4 or GSC. Also use when the user mentions "AI traffic," "ChatGPT referral," "Perplexity traffic," "AI Overvi1.2Kskills.sh installs