Guide · Agent skill
How to use the ljg-think drill-down root cause skill
Updated · written with AI from the project's own documentation and checked against our catalogue data
ljg-think is an agent skill, a folder of instructions that a Claude-based agent loads when your request matches it. It makes the agent drill vertically into one statement, phenomenon or question, asking "why is this so?" layer after layer until it reaches something that cannot be broken down further. This guide covers what it does, how to install it, how to trigger it, and what to watch for before you use it.
Drill-Down Root Cause Thinkerlijigang/ljg-skills · MIT Download the skill Quick start
- 1Download the ZIP. Download the skill ZIP from this page and unzip it so you have the ljg-think folder containing SKILL.md.
- 2Place the folder. Move the folder into ~/.claude/skills/ for all projects, or into .claude/skills/ inside one project.
- 3Or upload in the apps. In the Claude apps, open Settings, then Capabilities, then Skills, and upload the skill.
- 4Give a clear statement. Write one claim or question and use a trigger such as 想透, 本质是什么, or drill down.
- 5Read the layers. The agent returns a layered essay that moves from surface to basic principle.
- 6Find the saved file. Open the Markdown file the agent saved in ~/Documents/notes and reported to you.
What does the ljg-think skill make the agent do?
Instead of surveying a topic broadly, the agent stays on one line of questioning. Each layer answers "why is this so?" and moves from surface explanations toward a more basic framework, such as the structure of human nature, physical laws, or logic itself. The source describes the path as switching to a more fundamental explanatory frame at each step, for example from sociology to psychology to biology to physics.
The skill asks the agent to write the descent as a short, vivid piece rather than an analysis report. Each layer gets a short name, ends with an open question or contradiction that leads to the next layer, and the final layer is meant to land hard. The skill sets no fixed number of layers; shallow topics may bottom out in three, deeper ones in about seven.
- Stays vertical: asks why, not what else there is
- Skips background and citations of prior work
- Names each layer and ends it with a crack that leads down
- Saves the result as a Markdown file and reports the path
How do you install the ljg-think skill?
Download the ZIP for this skill from this page and unzip it. The skill is a folder containing SKILL.md. Put that folder in one of two places: your personal skills folder at ~/.claude/skills/, or a project-level .claude/skills/ folder inside a specific project if you want it only there.
You can also upload the skill in the Claude apps. Go to Settings, then Capabilities, then Skills, and upload it there.
- Personal: ~/.claude/skills/ (available across your projects)
- Project: .claude/skills/ inside the project folder
- Claude apps: Settings → Capabilities → Skills
How do you trigger ljg-think with a natural request?
The skill loads when your request matches its description. The source lists Chinese trigger phrases and English ones. Chinese triggers include 想透, 追本, 本质是什么, 为什么会这样, 深挖 and 钻到底. English triggers include "think deep" and "drill down". You can also trigger it without a keyword by giving the agent a statement and asking for depth analysis rather than a broad survey.
Give it one clear claim or question. The skill works best on a single idea, phenomenon or statement, because it peels one thread downward.
- "想透这件事：为什么拖延总是在最后一刻发生？"
- "Drill down: why do startups so often fail at distribution?"
- "本质是什么？人为什么需要被认可？"
- "这个观点我想钻到底：效率越高的团队越容易崩溃。"
What output should you expect?
The agent writes a Markdown file to ~/Documents/notes/ with a timestamped name that includes the topic, then reports the file path back to you. The file name pattern in the source is {timestamp}--追本-{topic}__think.md. The skill tells the agent not to use Org-mode syntax.
Expect a layered essay with named layers and a closing conclusion, not a list of sources or a balanced overview.
~/Documents/notes/{timestamp}--追本-{topic}__think.mdWhat does it need, and what are the safety notes?
The listing states no required tools or dependencies. The skill does not run scripts and does not use the network, according to the listing. Its only side effect is writing a file to ~/Documents/notes without asking you first. If you do not want files created on your computer, do not install it, or remove it after use.
The skill also tells the agent to run two date commands to get timestamps. Those commands are part of the skill's instructions, so review SKILL.md before installing if you want to know every command the agent will run.
- Requires: nothing listed
- Runs scripts: no
- Uses network: no
- Writes a file to ~/Documents/notes without asking
- Output is interpretive and not verified against sources
Tips for getting useful results
Start with a concrete statement rather than a topic word. "Remote work lowers productivity" gives the descent something to push on; "remote work" alone invites a survey, which this skill avoids.
If you need facts, check them separately. The skill does not instruct the agent to research or cite sources, so treat the layers as a reasoning exercise, not evidence.
- Phrase the input as a claim or a "why" question
- Use it for writing, study or thinking, not for lookups
- Read the saved file; the path is reported in the reply
What people use it for
- Working out why a habit or behaviour persists, down to human nature
- Tracing a business assumption to the underlying principle behind it
- Writing a dramatic, layered essay that goes from a claim to its foundation
- Studying why a phenomenon occurs, as a first-principles exercise for a class
Frequently asked questions
How do I install the ljg-think skill?
Download the ZIP from this page and unzip it. Place the ljg-think folder in ~/.claude/skills/ for personal use or in a project's .claude/skills/ folder. In the Claude apps you can instead upload it under Settings, then Capabilities, then Skills.
Is ljg-think free to use?
The skill is listed under the MIT licence. The listing does not mention any price, so check the repository at lijigang/ljg-skills for licence terms.
What does ljg-think require?
The listing states no requirements, does not run scripts and does not use the network. It does write a Markdown file to ~/Documents/notes.
Why did the skill not give me a broad overview?
The skill is built for depth only. It deliberately avoids breadth and background, so it follows one line of why-questions downward. For an overview, ask a normal question without the drill-down trigger words.
Can I use ljg-think for research with sources?
Not as a source of verified evidence. The skill does not tell the agent to research or cite sources, and its output is interpretive. Verify any factual claims yourself.
Where does the output file go, and can I change it?
The skill writes to ~/Documents/notes using a timestamped name that includes the topic, and it reports the path to you. The source does not describe a setting to change that location, so edit SKILL.md if you want a different folder.
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-think if you want a vivid, layered essay that pushes one idea down to its basic principles, for writing, study or personal thinking. Pick something else if you need quick facts, broad background, or sourced evidence for a decision, or if you do not want the agent saving files to your computer.
Download the skill