agentleFS
JakeschincariolJakeschincariol/unhinged-claude-skills/chinese-grandpa/SKILL.md

chinese-grandpa

Skill10 starsChanged 5 months agoMITView on GitHub

Compresses long, repeated sub-agent instructions by translating them into Mandarin (more meaning per character) to slash the token bill, then dispatches the compressed brief to a fleet of sub-agents while preserving original intent; invoke via /chinese-grandpa when you are about to fan out the same heavy instructions to many sub-agents and the cost is climbing.

The file

We split this skill into the rules it gives an agent

3 rules. About 0 are common, 3 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈3

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Treat the user's full English prompt/brief as the canonical source-of-truth.

    line 15
  2. Distinctive:

    Do a back-translation sanity check: confirm the Mandarin still encodes every critical requirement before sending it anywhere.

    line 18
  3. Distinctive:

    If anything critical is ambiguous after compression, fall back to English for that clause rather than risk meaning loss.

    line 21

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Chinese Grandpa

Mandarin carries more meaning per character than English, so the same instruction costs fewer tokens. This skill keeps an English source-of-truth, ships a Mandarin-compressed brief to your sub-agent fleet, and pockets the difference. Wise, frugal, and faintly disappointed in your token spend.

When to use

  • You are dispatching the same long instruction set to many sub-agents.
  • The token bill on a fan-out job is climbing and the instructions are repetitive.

Behavior

  1. Treat the user's full English prompt/brief as the canonical source-of-truth. Never discard it; keep it in the working context.
  2. Identify the portion that will be dispatched repeatedly to sub-agents (the reusable instruction block), separate from one-off, agent-specific parameters.
  3. Translate that reusable block into clear, modern Mandarin, optimizing for density while keeping every constraint, acceptance criterion, and edge case intact.
  4. Do a back-translation sanity check: confirm the Mandarin still encodes every critical requirement before sending it anywhere.
  5. Dispatch sub-agents with the Mandarin brief plus the English-language agent-specific parameters (file paths, IDs, targets stay in their original form — do not translate identifiers).
  6. Instruct each sub-agent to reason in whatever language it prefers but to RETURN results in English.
  7. If anything critical is ambiguous after compression, fall back to English for that clause rather than risk meaning loss.

Output

A Mandarin-compressed dispatch brief, a one-line note of the estimated token savings, and the preserved English source-of-truth for reference.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.