Przejdź do treści
Biblioteka skilli
Dokumenty i biuro · Pisanie

plain-writing

/plain-writing

Pisze i przepisuje prozę, podsumowania i listy tak, żeby czytelnik bez kontekstu zrozumiał je za pierwszym razem.

Repozytorium
dotfiles ↗
Licencja
MIT
Ostatnia zmiana repozytorium
13.11.2026

Ocena i bezpieczeństwo

100/ 100Klasa: dobry
Format25 / 25
Frontmatter z nazwą i opisem, poprawna nazwa, opis do 1024 znaków, nazwa zgodna z folderem.
Odkrywalność25 / 25
Opis mówi, kiedy użyć skilla, ma sensowną długość i jest w 3. osobie.
Treść35 / 35
Co najmniej 150 słów i do 500 linii, nagłówki, listy kroków, przykłady albo kod.
Higiena15 / 15
Bez zaślepek typu TODO, bez ścieżek Windows, z licencją.
Ryzykobrak sygnałów w skanie
Pliki w paczce
1
Skrypty
0
Komentarze HTML
0
Adresy URL w SKILL.md
3
Przestarzałość

Brak sygnałów. Repozytorium zmienione 13.11.2026, bez odwołań do wycofanych modeli i API.

Licencja

MIT, otwarta: można używać i zmieniać, z zachowaniem informacji o autorze.

Co robi i kiedy

Pisze i przepisuje prozę, podsumowania i listy tak, żeby czytelnik bez kontekstu zrozumiał je za pierwszym razem.

Oryginalny opis (EN)

Writes and rewrites prose, summaries, and lists so a reader without context understands them in one pass. Use for /plain-writing or deslopify.

Jak działa

  1. Orwell's Rules for Writing
  2. Structure
  3. Sentences
  4. Lists and feeds
  5. Response Style
  6. Formatting
  7. Deslopify
Sekcje z nagłówków SKILL.md, w oryginale · 70 linii, 795 słów
Podgląd SKILL.md
---
name: plain-writing
description: Writes and rewrites prose, summaries, and lists so a reader without context understands them in one pass. Use for /plain-writing or deslopify.
disable-model-invocation: true
---

# Plain writing

Write so that a smart reader who lacks your context understands the text in one pass. For docs, READMEs, and RFCs, also pick a structure from [technical-writing](https://github.com/backnotprop/pstack/blob/main/skills/technical-writing/SKILL.md). After drafting, reread the text as that reader and fix whatever they would stumble on.

## Orwell's Rules for Writing

From George Orwell, "Politics and the English Language" (1946):

1. Never use a metaphor, simile, or other figure of speech which you are used to seeing in print.
2. Never use a long word where a short one will do.
3. If it is possible to cut a word out, always cut it out.
4. Never use the passive where you can use the active.
5. Never use a foreign phrase, a scientific word, or a jargon word if you can think of an everyday English equivalent.
6. Break any of these rules sooner than say anything outright barbarous.

## Structure

- Lead with the point. In chat, restate the problem in a sentence before saying what changed, and skip the history.
  - Before: The exporter now waits on the reset header, and the tests pass.
  - After: The log exporter reread whole hours, so retries wrote duplicate events. It now waits for the reset header, and the six acceptance tests pass.
- Open each paragraph with its main point, then support it with a fact or example.
- Walk through sequences in order, one step per sentence or numbered item.
- Prefer prose for explanations. Keep a list to about four items, nesting extra items under groups.

## Sentences

- Use plain everyday words and established terms, and define a technical term the reader may not know. Call each thing by one name, using the real file, command, or symbol name.
- Use the concrete name instead of a coined term, metaphor, or slogan: "base" for "substrate", "the API" for "the API surface", "a limit that only decreases" for "a ratchet". [unslop](https://github.com/backnotprop/pstack/blob/main/skills/unslop/SKILL.md) lists more of these patterns.
- Name the actor in active voice: "the loader parses the file".
- State the mechanism or number.
  - Before: This change matters and makes ranking much faster.
  - After: The new ranking model cut p95 latency from 900 ms to 220 ms.
- Write complete sentences with articles and verbs, and spell out arrows and abbreviations.
- Join closely related clauses with "and", "because", or "so", and give unrelated ideas their own sentences.
  - Before: The gate runs on every merge. It blocks regressions. Nobody bypasses it.
  - After: The gate runs on every merge and blocks any change that fails a regression case.
- Say directly what a thing is, and end the sentence when the fact ends instead of adding an "-ing" phrase such as "ensuring reliability".
- Repeat the noun when "this", "that", or "it" could point at more than one thing.

## Lists and feeds

Readers scan the first two words of each item and assume that every difference between items means something, as [Making AI text readable at scale](https://fsdatalab.github.io/blog/agent-text-at-scale) explains with examples.

- Start each item with its verb or subject, in the tense readers expect: past for done, continuous for active, future for planned. Drop a subject that every item shares.
  - Before: `KV persistence`, `Login bug fixed by agent`
  - After: `Adding KV persistence`, `Fixing login bug`
- Put the words that tell items apart first: `Shipped: order #1001`, not `Your order #1001 has shipped`.
- Give every item the same structure, vocabulary, and kind of detail, with verbs whose tone matches the purpose, such as "Users complain" for problems.

## Response Style

- Answer in plain language, like one person talking to another.
- Keep answers short and concise: the direct answer first, in a few sentences.
- Skip background, caveats, and long lists unless the user asks for more detail.
- If something important doesn't fit, offer to expand in one line instead of including it.
- Do not omit critical information, but prefer brevity.
- Use the real name for things: the code identifier, the user's own words, or a concise description or appositive. Don't make up labels like "the switch-back note".
- Avoid nonce terms, and avoid cataphors with unclear antecedents or referents.
- Don't say "it", "that", and similar words unless the referent is obvious. Prefer precision.

## Formatting

- Use sentence case headings, straight quotes, and commas, periods, or "to" in place of em dashes, en dashes, and middle dots.
- Use colons only to introduce a list or as a short label such as "Summary:", and bold only for terms the reader must find.

## Deslopify

When the user asks to deslopify, rewrite the previous response, or the given text, for a sharp reader with no project context. Return only the rewrite.

Wersje

RepozytoriumWynikLicencja
Ta strona · davidgasquez/dotfiles100MIT
Agenta-AI/agenta100MIT
atomicbot-ai/atomic-agent100MIT
mobbin/skills100MIT
lxsmnsyc/agent-skills100MIT
KirillSachkov/agent-clarity100MIT
ciro-maciel/ccm-stack90MIT
shaunmarsden/say-it-plainly90MIT

Migawka bazy z 13.11.2026 · wersja paczki 20ac3fbfc135 (skrót SHA-256 całej paczki)

Pobrania w ostatnich 30 dniach: 0 · ranking

Zgłoś problem z tym skillem: szkodliwa treść, błędny opis albo prośba autora o usunięcie.

Podobne skille

  • 92say-it-plain
    Wyjaśnia dowolny temat tak, by czytelnik zrozumiał go za pierwszym razem: najpierw odpowiedź, krótkimi i prostymi zdaniami.
  • 100write-plain
    Pisze lub sprawdza tekst prostym, krótkim, międzynarodowym angielskim, by czytelnik zrozumiał go za pierwszym razem.
  • 90clear-writing
    Pomaga pisać tak, żeby czytelnik zrozumiał od pierwszego przeczytania: proste słowa, konkretne rzeczowniki i jedna myśl w zdaniu.
  • 100explain
    Wyjaśnia pojęcie techniczne albo przepisuje zawiły tekst tak, by czytelnik zrozumiał go przy pierwszym czytaniu, bez utraty dokładności.
  • 81language_humanizer
    Przepisuje lub ocenia dokument tak, by czytelnik zrozumiał go za pierwszym razem, bez gubienia warunków i liczb.
  • 100plain-language-rewriting
    Przepisuje gęste, urzędowe lub pełne żargonu teksty, żeby czytelnik znalazł, zrozumiał i działał od pierwszego czytania, bez utraty znaczenia.

Uwagi i zgłoszenia

Wiadomość trafi do Bartosza Sokolińskiego. Adres e-mail jest potrzebny tylko do odpowiedzi. Szczegóły w polityce prywatności.