agentleFS
TrueImmortal82TrueImmortal82/open-code-skills/skills/english-request/SKILL.md

english-request

Skill0 starsChanged 41 days agoMITView on GitHub

Do the model's own work in English while keeping the reply in the user's own language, so a long transcript costs fewer tokens. Use when the user's message is not in English — Russian/Cyrillic, Uzbek, or any other language — especially for long or multi-step requests; when the user asks to work in English to save tokens; or when a conversation is growing long in a non-English language. Restate the request in English for yourself, run all reasoning, planning, notes and tool arguments in English, and write only the final reply in the user's language (e.g. `ru`, `uz`). If the message is already `en`, do nothing. Never translate or alter code, paths, identifiers, commands, or quoted text; do not echo the translation into the reply.

The file

We split this skill into the rules it gives an agent

6 rules. About 0 are common, 6 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈6

    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:

    Do your own work in English.

    line 8
  2. Distinctive:

    Internal English: Run planning, reasoning, todo items, code comments, commit messages, and tool-call arguments in English.

    line 25
  3. Distinctive:

    Change nothing verbatim: Leave code, identifiers, file paths, commands, URLs, error strings, and quoted user text exactly as given.

    line 29
  4. Distinctive:

    Do not translate the user's documents or files — that is a different task and would change their content.

    line 35
  5. Distinctive:

    Do not switch the reply to English "for consistency".

    line 37
  6. Distinctive:

    Do not fire when the message is already English.

    line 38

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 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

English request

Do your own work in English; keep the reply in the user's language. The point is token cost: the model's generated tokens — reasoning, plans, notes, tool arguments, drafts — dominate a long transcript, and the same ideas are roughly 2–4× cheaper in English than in Cyrillic or another non-Latin script.

Why the saving is not in the request

Translating the request does not shrink it: the user's original message stays in context and costs its tokens either way. The lever is your generated tokens. So restate the request in English for yourself and generate everything internal in English. Printing the translation back would add tokens for no gain, which is the opposite of the goal.

Rules

  1. Restate, do not echo. Understand the request by restating it in English in your working notes. Never print the translation or announce this skill.
  2. Internal English. Run planning, reasoning, todo items, code comments, commit messages, and tool-call arguments in English.
  3. Reply in the user's language. The reply stays in whatever language the user wrote in. This skill changes your working language, not theirs.
  4. Change nothing verbatim. Leave code, identifiers, file paths, commands, URLs, error strings, and quoted user text exactly as given. Rewrite only your own prose.

Do not

  • Do not translate the user's documents or files — that is a different task and would change their content.
  • Do not switch the reply to English "for consistency".
  • Do not fire when the message is already English.

Checklist

  • Worked from an English restatement I did not print.
  • Plan, notes, tool arguments, and comments are in English.
  • Reply is in the user's language and reads naturally, not as a translation.
  • Code, paths, identifiers, commands, and quotes are unchanged.

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.