agentleFS
hashiiiiihashiiiii/rules-for-ai/skills/hashiiiii-write-en/SKILL.md

hashiiiii-write-en

Skill10 starsChanged 3 months agoMITView on GitHub

Use when you write, edit, or review English prose, including documentation, messages, and articles.

The file

We split this skill into the rules it gives an agent

7 rules. About 0 are common, 4 are distinctive, and 3 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈4

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 3

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Write for the reader's task.

    line 8
  2. Very specific:

    Run node <skill-dir>/scripts/check.mjs <absolute-draft-path> before delivery.

    line 15
  3. Very specific:

    Check chat prose through a temporary file.

    line 23
  4. Very specific:

    If Node.js, npm, installation, or checking is unavailable, review manually and state which check could not run and why.

    line 24
  5. Distinctive:

    Do not invent actors, causes, sources, opinions, reactions, or personal experience.

    line 29
  6. Distinctive:

    Keep code, commands, identifiers, paths, URLs, quotations, and product names exact unless the task requires changes.

    line 30
  7. Distinctive:

    Keep errors inside quoted logs exact, including their punctuation.

    line 31

≈ “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

English writing

Write for the reader's task. Use the pinned Harper grammar checker to check the draft locally.

Workflow

  1. Identify the reader, purpose, tone, and requested English dialect from the task and surrounding text. Retain the source for comparison.
  2. Draft from the supplied facts. Lead with the point the reader needs. Give each paragraph one job and connect its sentences.
  3. Revise for natural English: check articles, agreement, tense, prepositions, word choice, and sentence flow. Prefer concrete verbs and ordinary words. Split sentences when this helps the reader; keep related conditions together.
  4. Run node <skill-dir>/scripts/check.mjs <absolute-draft-path> before delivery. Replace <skill-dir> with this skill's absolute directory. Markdown files use Harper's Markdown parser; other files use plain text. For another dialect, pass --dialect British, Australian, Canadian, or Indian before the path. The default is American English.
  5. Review each diagnostic in context. Apply a suggestion only when it improves the text and preserves the source's meaning. Quoted logs, identifiers, and deliberate wording can legitimately trigger findings.
  6. Compare the draft with the source, then read it as a whole. Check grammar, flow, repetition, missing conditions, and changes to certainty or timing. Rerun the checker after edits; deliver the exact checked text.

The checker requires Node.js 20 or later and npm 8.3 or later. On first use, run node <skill-dir>/scripts/check.mjs setup.
Setup uses npm to install the locked packages without lifecycle scripts in a user-owned cache, outside the project and skill directory.
Checks then run offline; the draft is never sent to an external service. See runtime dependencies for cache and update details.

Check chat prose through a temporary file. Remove only temporary files created for this task.
If Node.js, npm, installation, or checking is unavailable, review manually and state which check could not run and why.

Constraints

  • Preserve facts, quantities, conditions, exceptions, uncertainty, timing, scope, and requirement strength.
  • Do not invent actors, causes, sources, opinions, reactions, or personal experience. Request missing information or use the supplied facts.
  • Keep code, commands, identifiers, paths, URLs, quotations, and product names exact unless the task requires changes.
  • Keep errors inside quoted logs exact, including their punctuation. Put any added sentence punctuation outside the quotation.
  • Preserve the requested reader, tone, format, and length. Use the same term for the same meaning.
  • Preserve the writer's voice. Remove filler and repeated explanations without adding opinions or making every sentence the same length.

Output

For edits, change only passages that break these rules or the user's requirements.
Return the finished text unless the user requests an explanation. Keep drafts and diagnostic notes internal.
For reviews, report each problem passage, its reason, and a proposed correction. Rewrite the full text only when requested.
Harper reports suggestions, not a quality score or proof of factual accuracy. Never claim a check ran when it did not.

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.