Guide · Agent skill
How to use the Megacave Classical Chinese Mode agent skill
Updated · written with AI from the project's own documentation and checked against our catalogue data
Megacave is an agent skill that makes an AI agent answer in Classical Chinese (文言文), using fewer characters while keeping code, commands, paths, error messages and numbers exactly as written. This guide explains what the skill does, how to install it, how to trigger it, and what to watch for. After reading it, you will be able to add the skill to your agent, switch the mode on and off, and judge whether it suits your work.
Megacave Classical Chinese ModeJuliusBrussee/caveman · Apache-2.0 Download the skill Quick start
- 1Download the ZIP. Get the megacave ZIP from the skill's page on this directory.
- 2Unzip the skill. Unzip it into ~/.claude/skills/ for all projects, or into .claude/skills/ inside one project.
- 3Or upload it in the Claude apps. In the Claude apps, open Settings, go to Capabilities, and upload the skill under Skills.
- 4Turn the mode on. Type /megacave or /caveman wenyan in your session.
- 5Check the status. Run /caveman status to see the current mode without changing it.
- 6Turn the mode off. Say "stop caveman" or "normal mode" to return to standard answers.
What the megacave skill makes the agent do
Once the skill is on, the agent replies in a compressed Classical Chinese register. Subjects are omitted where they can be recovered, the verb comes before the object, and particles such as 之, 乃, 為, 其, 則, 故 and 以 replace longer connective phrases. The result has fewer characters on screen.
The skill is a style change only. It does not add new knowledge or tools. Technical content is kept: code, commands, paths, API names and error strings stay in their original script, and negations and numbers are preserved. If a clause is ambiguous in Classical Chinese, the skill says the agent should switch to plain Chinese (白話) for that clause.
Install the megacave skill
Option one: download the ZIP from the skill's page on this directory and unzip it into your user skills folder, ~/.claude/skills/, or into a project's .claude/skills/ folder if you want the skill for one project only. The package contains a single file, SKILL.md.
Option two: in the Claude apps, open Settings, go to Capabilities, and upload the skill under Skills.
- Personal install: ~/.claude/skills/
- Project install: .claude/skills/ inside the project folder
- Claude apps: Settings → Capabilities → Skills
Trigger the skill with a request
The skill's description says it is invoked only with /megacave or /caveman wenyan. Natural requests such as 'please answer in caveman style' are not what the skill lists as its trigger, so use one of the two slash commands to turn it on.
Once on, it stays active for the whole session. It remains on until you say "stop caveman" or "normal mode". If you are unsure whether it is still active, run /caveman status. That command reports the current mode and changes nothing.
/megacave
/caveman wenyan
/caveman statusExamples of what the answers look like
The skill's own example shows a React question. The plain Chinese version says the component re-renders each time because a new object reference is created on every render, and suggests wrapping it with useMemo. The megacave version reads: 每繪新生對象參照，故重繪；以 `useMemo` 包之則免。
For code edits, the skill asks for only the changed lines plus context, not the whole file. Identifiers stay in their original form, so you can copy them directly into your project.
Safety notes: scripts and network
According to the skill's listing, megacave runs no scripts and uses no network access. It requires nothing beyond the agent itself, and the package contains only SKILL.md.
The skill also says that security warnings, irreversible actions, and anything saved outside the chat (such as code, commits, docs, issues or pull requests) switch the agent back to plain Chinese or the user's language, and that it confirms before irreversible steps.
Tips for getting the most out of it
Use the skill when you read Classical Chinese and want dense answers, or when you want to test whether an agent can keep a strict writing register. Keep a plain-language check in mind: if an answer is hard to follow, ask the agent to switch to plain Chinese or turn the mode off.
Do not expect savings. The skill states that token savings are small and noisy and should never be claimed.
- Say "normal mode" or "stop caveman" to return to standard answers.
- Ask for a clear explanation in plain Chinese when a clause is unclear.
- Check code and commands in the output before running them.
What people use it for
- Reading short, dense Chinese answers about code review or debugging while keeping commands and error messages exact.
- Experimenting with playful output styles for a demo or internal test.
- Checking whether an agent can hold a strict Classical Chinese register across a long session.
- Asking for quick, condensed updates on a project while keeping file paths and numbers unchanged.
Frequently asked questions
Is the megacave skill free?
The skill is licensed under Apache-2.0, as listed on this page. The source does not mention pricing, so check the license terms in the repository before using it commercially.
What do I need to run megacave?
The listing states that it requires nothing beyond the agent itself. It runs no scripts and uses no network access, and the package contains only SKILL.md.
How do I install the megacave skill?
Download the ZIP from this page and unzip it into ~/.claude/skills/ or a project's .claude/skills/ folder. In the Claude apps, you can instead upload it under Settings → Capabilities → Skills.
How do I turn megacave off?
Say "stop caveman" or "normal mode". The skill stays on for the whole session until one of these phrases is used.
Does megacave save tokens or money?
No. The skill says token savings are small and noisy and should never be claimed, so do not rely on it for cost reduction.
Can I read the answers if I do not know Classical Chinese?
Probably not easily. The skill is aimed at readers who read Classical Chinese, and it says unclear clauses should switch to plain Chinese, but the terse style may still be hard to follow.
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
Megacave suits readers of Classical Chinese who want short, dense answers, and developers who want to test a strict output style while keeping technical text exact. It is a poor fit for beginners, anyone who needs full explanations, or anyone expecting real token or cost savings.
Download the skill