Przejdź do treści
Biblioteka skilli
Projektowanie i UX · Pisanie

write-ui-microcopy

/write-ui-microcopy

Pisze teksty w interfejsie: etykiety przycisków, komunikaty o błędach, potwierdzenia i podpowiedzi, w głosie produktu.

Repozytorium
truecast ↗
Licencja
MIT
Ostatnia zmiana repozytorium
17.10.2026

Ocena i bezpieczeństwo

90/ 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.
Treść25 / 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
0
Przestarzałość

Brak sygnałów. Repozytorium zmienione 17.10.2026, bez odwołań do wycofanych modeli i API.

Licencja

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

Co robi i kiedy

Pisze teksty w interfejsie: etykiety przycisków, komunikaty o błędach, potwierdzenia i podpowiedzi, w głosie produktu.

Oryginalny opis (EN)

Use when writing any words inside the UI — button verbs, labels, empty/error/success messages, tooltips, confirmations, onboarding. Copy is interface; make it specific, in the product's voice, and helpful at the moment of friction. Not generic CTAs.

Jak działa

  1. The rules
  2. The voice discipline
  3. The AI register is not a voice (the microcopy scrub)
  4. The discipline
Sekcje z nagłówków SKILL.md, w oryginale · 59 linii, 715 słów
Podgląd SKILL.md
---
name: write-ui-microcopy
description: Use when writing any words inside the UI — button verbs, labels, empty/error/success messages, tooltips, confirmations, onboarding. Copy is interface; make it specific, in the product's voice, and helpful at the moment of friction. Not generic CTAs.
---
# Write UI microcopy — words are interface

The words *inside* the UI are part of the design, not an afterthought for "content" — a label, a button
verb, an error line steers behavior more than most pixels. Microcopy is yours (the product's *external*
positioning and marketing copy is the marketer's — consult them when the line carries brand weight, but
the in-product words are interface and they're yours).

## The rules
1. **Verbs the user actually wants.** Not "Submit," "Get Started," "Learn More" — the specific action:
   "Connect repo," "Re-run build," "Talk it through →." The button says what happens when you press it.
2. **Labels in the user's language** (match-to-real-world, heuristic #2) — not internal jargon or DB
   field names. The label is what the user would call the thing.
3. **Errors: name what's wrong + the way out** (heuristic #9). Not "Error 500" / "Something went wrong" —
   "We couldn't reach the server. Retry, or check your connection." Plain language, honest, no dead end.
4. **Empty states say something specific** (`design-the-states`) — what this is for and the next step, in
   voice; never "Nothing here yet 🎉."
5. **Confirmations for the destructive** — name the consequence ("Delete 3 items? This can't be undone"),
   give the escape. Match the friction to the stakes.
6. **Set expectations** — especially for slow or AI work: "This usually takes a few seconds"
   (`design-the-ai-interaction`).

## The voice discipline
- **Consistent voice** across the product — one personality (direct, honest, calm — match the product's).
  Decide the voice once and **write it down**: a mini voice spec in this project's `instance/` — five
  words we use, five we never use, one reference error message, one reference empty state. "Off-voice"
  then cites a spec line, not a feeling; and the spec is derived from *this* product's users
  (`research-the-user`), not a house default.
- **The user's vocabulary, not yours** — labels and messages use the words this audience uses for these
  things (an on-call SRE and a first-run consumer don't get the same register, density, or patience).
- **Concise** — every word earns its place; users scan, they don't read. Front-load the meaning.
- **Honest, not cute** — humor and emoji rarely survive the user who's frustrated or confused at that
  exact moment; respect the user's intent over a quip.
- **Clear over clever** — if a clever line risks ambiguity, the plain line wins. Clarity is the job.

## The AI register is not a voice (the microcopy scrub)
Unedited model output has a recognizable in-product register, and users have learned to read it as
"nobody wrote this." Scrub for it wherever copy was generated (clusters, not single hits — era: 2026):

- **Exclamation cheer.** "You're all set! 🎉" "Great choice!" — the model celebrates; a product speaks.
  State what happened and what's next.
- **Apology theater.** "Oops! Something went wrong on our end. Don't worry —" is three tells in one line:
  the mascot "Oops," the vague failure, the instruction to feel. Name what broke + the way out (rule 3).
- **Hedge padding.** "It looks like you may not have any projects yet" — the user either has projects or
  doesn't; the softener is the model's, not the product's. "No projects yet. Create one."
- **Chatbot pleasantries in chrome.** "Sure!", "Let's get you set up!", "Feel free to…" in buttons,
  tooltips, onboarding — conversational filler where an interface should be an instrument.
- **Explaining feelings instead of the state.** Copy that narrates emotion ("We know this is
  frustrating…") instead of giving the fact and the action. Empathy in UI is a working way out, fast.
- **The cold read at the worst moment.** Read every error/empty/confirmation as the user who just lost
  work at 2am. The register that survives *that* read ships; the writer of the line doesn't get to be
  its judge.

## The discipline
- Microcopy is most load-bearing exactly where the user is stuck (error, empty, confirmation) — write
  those with the most care, not the least.
- Test labels for ambiguity in a `usability-test` — if two users read a label two ways, it's a bug.
- Pull in the **marketer** when in-UI copy bleeds into positioning/brand voice; pull in the
  **product-manager** when the words imply a promise about what the product does.

Wersje

W bazie nie ma innych skilli o nazwie write-ui-microcopy.

Migawka bazy z 13.11.2026 · wersja paczki 4d6b082b32cb (skrót SHA-256 całej paczki)

Pobrania w ostatnich 30 dniach: 1 · ranking

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

Podobne skille

  • 95ux-microcopy
    Pisze i poprawia mikrotekst interfejsu: etykiety przycisków, komunikaty błędów i pustych stanów, podpowiedzi, etykiety formularzy, powiadomienia i potwierdzenia, tak by były jasne, zwięzłe i spójne z głosem marki.
  • 90ux-copy
    Pisze albo ocenia mikrokopię: przyciski, komunikaty o błędach, puste stany, potwierdzenia i podpowiedzi w interfejsie.
  • 100ux-writer
    Pisze mikrocopy i teksty w interfejsie: komunikaty o błędach, tooltipy, etykiety przycisków, stany puste, onboarding i okna potwierdzeń.
  • 100open-ds-write
    Pisze teksty interfejsu w stylu schematu Open DS: etykiety, komunikaty o błędach, potwierdzenia, puste stany i podpowiedzi.
  • 100microcopy
    Pisze mikrotekst interfejsu: komunikaty błędów, puste stany, podpowiedzi, etykiety przycisków, strony 404 i 500 oraz onboarding, prostym językiem i bez obwiniania użytkownika.
  • 100ux-writing
    Pisze skuteczny mikrocopy, komunikaty o błędach, etykiety przycisków i teksty interfejsu według zasad UX writingu, także dla stanów pustych i konwersacyjnego UI.

Uwagi i zgłoszenia

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