agentleFS
zznabilzznabil/lean-agent-skill-collection/skills/writing/SKILL.md

writing

Skill1 starsChanged 3 months agoMITView on GitHub

Draft or edit instructions, UI text, errors, help, emails, documentation, reports, and other prose for clarity, accessibility, evidence, and task fit while preserving meaning and the requested voice.

51 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use addskill, apply-firmware and 10 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • ryanportfolioSkill

    addskill

    Create, import, update, or install repo or personal skills: runtime ownership, resources, discovery checks.

    Also in 6 projects that use this

  • ryanportfolioSkill

    apply-firmware

    Bring current folder (plain folder, existing repo, old or partial firmware copy) up to full Harness Firmware in place. Mirror external repo → adopt-repo.

    Also in 6 projects that use this

  • +10 more from ryanportfolio

The file

Show its rules and text

We split this skill into the rules it gives an agent

14 rules. 0 are common, 14 are distinctive, and 0 are very specific to this project.

  • 0

    Common

    Most of the 45 edited versions we read keep these.

  • 14

    Distinctive

    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:

    Language — ASD-STE100-inspired: Use short, active sentences, direct verbs and concrete words.

    line 15
  2. Distinctive:

    Terminology — ISO 704-inspired: Use one preferred term per concept.

    line 16
  3. Distinctive:

    Use TL;DR only when it helps a long response, and place it first.

    line 20
  4. Distinctive:

    Make meaning explicit: Use familiar, literal words.

    line 26
  5. Distinctive:

    Make action manageable: Use short paragraphs, useful headings and lists.

    line 27
  6. Distinctive:

    Put prerequisites, warnings, costs and irreversible effects before the action.

    line 28
  7. Distinctive:

    Reduce memory demands: Keep decision information nearby.

    line 29
  8. Distinctive:

    Keep essential steps visible.

    line 30
  9. Distinctive:

    Verify material claims, names, dates, numbers, calculations, quotations, citations, and links.

    line 36
  10. Distinctive:

    Check that the reader can find the point and, where relevant, identify the next action, expected result, and recovery.

    line 38
  11. Distinctive:

    Check consistent terms and exact labels.

    line 39
  12. Distinctive:

    Treat readability scores and Vale checks as clues, not proof of comprehension or accessibility.

    line 40
  13. Distinctive:

    Read USER-INFORMATION.md for substantial manuals, onboarding, embedded help, forms, UI text, warnings, errors, or recovery guidance.

    line 46
  14. Distinctive:

    Read INSTRUCTION-EDITING.md for skills, agent policies, prompts, workflows, checklists, handoffs, or consequential procedures.

    line 48

“Common” is measured: kept by at least 6 in 10 of the 45 edited versions of this file we could read. “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

Writing

Choose draft or edit. Help the reader find, understand, and use the information.

For user-facing text, default to familiar language that does not assume technical knowledge. Respect requested expert detail and the reader's stated needs. Do not infer ability from a diagnosis or impose a universal reading age. Use respectful language for the intended audience.

Communication kernel

Number ordered actions; group long lists without hiding required items. After interruption or task change, restore relevant completed/current/pending state, not repetitive status each turn. Preserve language, exact tokens, formats, permissions and evidence. Never invent estimates or diagnose readers. Show verified outcomes and factual errors; give a concrete next action only if work remains. Apply this kernel to the skill, companions, examples, generated text and surrounding reply. For agent instructions, the executing agent is the reader. Trusted root AGENTS.md governs when loaded; otherwise use this standalone fallback. Never claim root activation without evidence.

  1. Language — ASD-STE100-inspired. Use short, active sentences, direct verbs and concrete words. Keep one main idea or action per sentence when practical.
  2. Terminology — ISO 704-inspired. Use one preferred term per concept. Distinguish related concepts and explain necessary unfamiliar terms near first use. Preserve exact labels, commands, identifiers and quotations; explain them alongside.
  3. Structure — Diátaxis. Separate tutorial, how-to, reference and explanation when useful; do not force categories into short answers. Lead with the answer or action. Use only necessary detail. Omit routine Summary headings and closing recaps. Use TL;DR only when it helps a long response, and place it first. These influences do not establish formal conformance; other frameworks remain task-selected.

Draft or edit

  1. Identify the need. Establish the reader, task, result, evidence, medium and requested voice. Inspect supplied context before asking.
  2. Expose the point. Lead with the supported answer or action when the requested form allows. Remove filler, praise, request restatement and repetition.
  3. Make meaning explicit. Use familiar, literal words. Resolve unclear pronouns, abbreviations and assumptions. Name actors and actions; illustrate difficult ideas when helpful.
  4. Make action manageable. Use short paragraphs, useful headings and lists. Number ordered steps; give each one main action or tightly coupled group. Put prerequisites, warnings, costs and irreversible effects before the action. Include needed expected results and safe recovery.
  5. Reduce memory demands. Keep decision information nearby. Repeat necessary conditions rather than require recall. Keep essential steps visible; place optional explanation afterwards without omitting requested expert depth.
  6. Preserve the contract. Retain facts, actors, negation, numbers, conditions, exceptions, permissions, requirement strength, uncertainty, evidence limits, quotations and exact machine-sensitive text. Preserve format and voice without weakening factual or safety-critical meaning. Do not expand scope.

Check before delivery

  • Verify material claims, names, dates, numbers, calculations, quotations, citations, and links.
  • Compare the source and revision in both directions. Check what disappeared and what was added.
  • Check that the reader can find the point and, where relevant, identify the next action, expected result, and recovery.
  • Check consistent terms and exact labels. Do not turn an unknown outcome into a failure or success.
  • Treat readability scores and Vale checks as clues, not proof of comprehension or accessibility.

Deliver the finished text. Flag only material unresolved claims, decisions, or verification limits. Do not add Summary or TL;DR to the drafted text unless requested.

Supporting references

Read USER-INFORMATION.md for substantial manuals, onboarding, embedded help, forms, UI text, warnings, errors, or recovery guidance. Use its audience-specific design and evaluation checks. Easy-to-Read is a specialised adaptation that requires intended-reader involvement; do not claim verification or certification from this skill.

Read INSTRUCTION-EDITING.md for skills, agent policies, prompts, workflows, checklists, handoffs, or consequential procedures. Preserve requirement strength, source links, resources, exceptions, and completion evidence.

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.