ljg-is

Writes plain-language explanations connecting a concept's definition, mechanism, and actionable judgment.

7.3k|818|Updated Mar 8, 2026
One-click install
npx skills add https://github.com/lijigang/ljg-skills --skill ljg-is
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: ljg-is
Source: https://github.com/lijigang/ljg-skills/tree/main/skills/ljg-is
Command: npx skills add https://github.com/lijigang/ljg-skills --skill ljg-is

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve?

When users ask "what is X" or "how should I understand X", a bare definition or a list of uses rarely produces real understanding. This Skill turns a noun or concept into a usable explanation: it states what X is and how it differs from neighboring concepts, explains how X actually operates, derives what old judgment must be revised, and ends with concrete guidance for acting on that understanding.

Core Features & Use Cases

  • Definition-to-action pipeline: Connects four questions (what it is, how it works, what belief it changes, how to act next) into one continuous reasoning path instead of a checklist.
  • Org/Markdown note generation: Produces a timestamped note in ~/Context with structured metadata (definition, operation, recognition, guidance, basis, falsifier) following the ljg-is-v5 schema.
  • Built-in validation: Ships a Bun-based validator (Tools/ValidateNote.ts) that checks schema, headings, paragraph counts, vague guidance, rumor-based evidence, and legacy contract residue.
  • Use Case: Ask "what is an objective function and how do I judge when to trust an 'optimal' result" and receive a plain-Chinese essay that defines it against constraints and metrics, walks through how it ranks candidate solutions, corrects the misconception that optimization is objective, and tells you to inspect what the function rewards before trusting high scores.

Quick Start

Ask the assistant to use ljg-is to explain a concept you want to truly understand, for example: explain what recursion is and how to decide when to use it.

Frequently Asked Questions about ljg-is

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I get a deep explanation of a concept instead of just a definition?

Invoke ljg-is with the concept name. It produces an essay that defines the concept against its nearest neighbors, walks through how it operates with a concrete example, states which prior belief must change, and gives guidance for acting on the new understanding.

What kinds of concepts does ljg-is work best on?

It suits technical concepts like objective functions or recursion, plus institutions, platforms, roles, methods, and norms. It is not meant for pure fact lookup, formal proofs, step-by-step tutorials, or deep causal drilling, which belong to other skills.

What output format does ljg-is generate?

It writes a timestamped Org file (or Markdown variant) to ~/Context named with the pattern timestamp--理解-target__is.org. The note carries ljg-is-v5 schema metadata fields: definition, operation, recognition, guidance, basis, and falsifier.

How is the generated note validated?

Run bun Tools/ValidateNote.ts with the note path. The validator checks schema, filename and identifier consistency, heading quality, paragraph counts, vague guidance, rumor-based evidence, and residue from older contract versions, exiting nonzero on failure.

When should I not use ljg-is?

Skip it when you only need a formal definition, a formula proof, a translation, or an operational tutorial. For vertical causal drilling use ljg-think, and for recurring motif structures use ljg-structure, as noted in the skill's boundary rules.