Guide · Agent skill
How to use the ljg-rank agent skill: decompose a field into its root rank
Updated · written with AI from the project's own documentation and checked against our catalogue data
ljg-rank is an agent skill, a folder with a SKILL.md file of instructions that a Claude agent loads when it needs a specific method. Give it a field such as startups, physics or economics, and it tries to find the few independent generators underneath that field instead of a long list of surface factors. This guide shows you how to install it, how to trigger it with natural requests, what the output looks like and where it is the wrong tool.
Root-Rank Decomposition Enginelijigang/ljg-skills · MIT Download the skill Quick start
- 1Download the ZIP. Get the ljg-rank ZIP from this page.
- 2Unzip into a skills folder. Extract it so SKILL.md sits in ~/.claude/skills/ or in a project's .claude/skills/ folder.
- 3Or upload in the Claude apps. Go to Settings, then Capabilities, then Skills, and upload the skill.
- 4Ask in natural Chinese. Use a trigger phrase such as 降秩 or 背后是什么 followed by a named domain.
- 5Read the essay and diagram. Check that the forces regenerate the phenomena and that none is redundant.
What the skill makes the agent do
The skill asks the agent to find the root rank of a domain: how many truly independent generators sit underneath it. A valid answer must be able to regenerate the phenomena one by one, and removing any single generator must leave some phenomenon unexplained.
The skill also applies two tests from David Deutsch's The Beginning of Infinity: reach, meaning the explanation predicts things beyond the original list that are actually observed, and hard to vary, meaning changing any detail of a good explanation makes its predictions fail.
- Input: a domain, for example 'startups' or 'economics'.
- Output: a Chinese-language prose essay that walks from scattered factors to a small set of core forces.
- Where useful, an ASCII diagram showing how the forces relate, such as a loop, a layered cross-section or a two-axis grid.
How to install ljg-rank
Download the ZIP for the skill from this page. Unzip it so that the folder containing SKILL.md ends up in one of two places: your personal skills folder at ~/.claude/skills/, or a project-specific .claude/skills/ folder inside the project you are working in.
If you use the Claude apps instead, open Settings, go to Capabilities, then Skills, and upload the skill there.
- Personal install: place the skill under ~/.claude/skills/.
- Project install: place the skill under .claude/skills/ in that project.
- Claude apps: Settings → Capabilities → Skills, then upload.
How to trigger the skill with natural requests
You do not need a special command. The skill's description lists trigger phrases, so ask in plain Chinese using words like these: 降秩, 找秩, 秩是什么, 这个领域靠什么撑着, 背后是什么. You can also ask for any domain to be decomposed into its irreducible generators.
Name the domain clearly. The more specific the field, the easier it is for the agent to find real generators rather than generic slogans.
- 降秩 创业
- 这个领域靠什么撑着？经济学
- 找秩：物理学的根骨架是什么
- 背后是什么？分析一下独立的生成器
Examples of what you get back
For a field like startups, the skill is meant to avoid stopping at surface items such as vision, execution, team and market. It keeps asking where each of those comes from until it reaches a layer that cannot be pushed further down, and then checks that the result can regenerate the original phenomena.
The output is prose, not a bullet list or a checklist. Expect the agent to lay out the structure, then add a diagram at the end whose shape follows the structure it found.
What it needs and what it runs
The skill lists no required tools or packages. It does not run scripts and does not use the network. The only file in the skill is SKILL.md, which contains the instructions the agent follows.
- Requires: nothing listed.
- Runs scripts: no.
- Uses network: no.
- Files: SKILL.md.
Safety notes and limits
Because the skill runs no scripts and makes no network calls, the main risk is interpretive rather than technical. The conclusions are essays built on the framework, and the framework asserts no empirical verification of them.
Do not rely on the output alone for decisions with real stakes. Check the claims against sources you trust.
- Treat the output as an interpretation, not a verified fact.
- The skill is tuned to Chinese-language writing and prose rules; English output is not what it is built for.
- It is not a source of practical instructions or quick facts.
What people use it for
- Explaining how a startup or organisation is held together by a few core forces, for an essay or internal memo drafted in Chinese.
- Building a structured mental model of a field such as physics or economics before reading deeper material.
- Writing a long-form Chinese essay about how a domain works, using the diagram as a structural backbone.
Frequently asked questions
How do I install the ljg-rank skill?
Download the ZIP from this page and unzip it so SKILL.md sits in ~/.claude/skills/ for personal use or in .claude/skills/ inside a project. In the Claude apps you can instead go to Settings, then Capabilities, then Skills, and upload it.
Does ljg-rank require any tools or packages?
The source lists no requirements. The skill runs no scripts and uses no network, and it consists of a single SKILL.md file.
Is ljg-rank free to use?
The skill is published under the MIT licence. The source does not mention pricing, so check the licence and your own Claude plan for any costs.
Can I use ljg-rank in English?
Not as intended. The skill is tuned to Chinese-language writing style and prose rules, so its output is in Chinese.
What should I use instead for quick facts or how-to steps?
Use this skill only for deep structural analysis. For factual lookups or practical instructions, the skill is a poor fit because its output is an essay, not instructions.
Why does the skill produce nothing useful when I ask for a short answer?
It is designed to produce a structured essay that walks from many factors to a few core forces. Trigger it with a clearly named domain and expect a long prose answer rather than a short summary.
New agent skills, repos and AI tools — every week
The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.
Our verdict
Use ljg-rank if you want a structured, Chinese-language essay that reduces a complex field to a few core forces, and you are comfortable reading interpretive prose and ASCII diagrams. Pick something else if you need English output, quick facts, practical how-to steps, or evidence-based conclusions for decisions with real stakes.
Download the skill