Index of every page on this site: llms.txt
All skills
Run it yourself

Interface copy that works — voice/tone, the PACE framework, alert anatomy, naming features, empty states, error messages, and the small-word edits that measurably improve UX. Distilled from Apple's writing sessions (2017–2026). Use when writing or reviewing any user-facing text — labels, alerts, onboarding, notifications, empty states, paywalls — or naming a feature.

Use this Skill: https://skilld.dev/gh/rshankras/claude-code-apple-skills/ux-writing

Nothing lands on disk. Nothing to clean up.

Fork this Skill

Edit a local copy. It keeps the author and licence.

SKILL.md

≈97 tokens for metadata: the name and description. ≈1.2k when used: this file.

Description uses 4.8% of example budget

Before choosing a Skill, your Agent reads its name and description. All available Skills share that space.

  • A shorter description leaves more room for other Skills. This entry exceeds our 1% size suggestion.

In our Claude Code example, all Skill names and descriptions share 8,000 characters. This Skill uses ≈382 characters, or 4.8%.

The 1% threshold is a size suggestion. Longer descriptions can still fit.

Your model, settings, and other Skills decide how much text your Agent can read.

Example settings and source

The example uses a 200k-token context and default Claude Code settings. The count includes the name, description, separators, and when_to_use when present. Codex also counts local file paths.

Skit's source and limits: Codex 0.160.1, Claude Code 2.1.292.

UX Writing

Interface text is interface design. This skill covers writing it: the frameworks Apple's writers use, the anatomy of the hardest components (alerts, errors, empty states, names), and the mechanical edits that improve any screen's copy in minutes.

When This Skill Activates

  • Writing or reviewing user-facing copy: labels, buttons, alerts, onboarding, notifications
  • Naming a new feature, tier, or setting
  • Error messages that frustrate, empty states that dead-end, permission prompts that get denied
  • Localization prep (copy that survives translation)

The PACE framework (Writing for Interfaces)

Run every screen's copy through four questions:

  • Purpose — what must this screen communicate? Headers and buttons should carry the screen alone; know what to leave out. Hierarchy in words mirrors hierarchy in layout.
  • Anticipation — what will the user ask or do next? Write for that ("What comes next?").
  • Context — where is the user right now (quiet home vs. busy airport, mid-task vs. idle)? That decides how much to say.
  • Empathy — write for a person having a problem, not for the system reporting one.

Voice vs. tone

  • Finding voice: describe the app as a person, cluster the personality adjectives, keep 2–3 as your voice attributes — constant across the app, while tone flexes per moment (more directness for high-stakes info, sparing warmth in celebrations).
  • The test that recurs in every Apple session: read it aloud. Does it sound like the best version of your app talking?
  • ❌ Jargon ("it can make people feel left out"); ❌ personality at the cost of usefulness.

The four small edits (biggest impact per minute)

  1. Cut filler: "easily," "quickly," "simply," "just" — ✅ "Enter your license plate to pay for parking" ❌ "Simply enter your license plate number to quickly pay for parking."
  2. Kill repetition — merge redundant sentences: ❌ "We're running late. Your driver won't make it on time. They'll be there in 10 minutes." ✅ "Delivery delayed 10 minutes."
  3. Lead with the why — benefit before action: ✅ "To get reservation updates, enter your phone number." Notifications too: ✅ "Keep your streak going by solving today's crossword."
  4. Keep a word list — three columns: approved term · alternatives to avoid · definition (e.g. "alias — not handle/username/title — the player's in-game name, shown to others"). Reuse identical button labels ("Next") everywhere; consistency builds trust.

Alert anatomy (Writing Great Alerts)

First: should this be an alert at all? Alerts interrupt by design — reserve them for things the user must know now.

  • Title: the main point, one sentence or less. Message: only if needed.
  • The scannability test: title + buttons alone must convey the situation.
  • ❌ Vague hedging ("You may need to…") — name the file, name the fix.

Naming features (Craft Clear Names, WWDC26)

Judge every candidate on three criteria: it belongs (fits the app and its location), it sets expectations (readers predict what they'll get), it works everywhere (languages, markets, platforms).

  • Process: Think / Feel / Do — what should the audience think, feel, and do? Group the answers into themes; name from the themes; test candidates in natural sentences ("just search for ___").
  • ✅ Industry-standard words in high-stakes domains: "Balance," not "Spending Power."
  • ✅ Compounds are fine when the parts self-explain ("AutoMix"); emotional names win where the moment is emotional ("Memories").
  • ❌ Implementation jargon — describe the outcome, not the mechanism.

Empty states, permissions, localization

  • Empty states: never dead-end — say what will appear and how, ideally with example input.
  • Permission prompts: explain value first, ask at the contextually right moment (mechanics in generators/permission-priming).
  • Localization: plain language over idiom/humor; leave layout room and accessibility text for every element.

Output Format

Copy review table:

Location | Current | Problem (PACE / filler / vague button / jargon) | Rewrite

ordered by frequency-of-exposure (navigation labels before deep-screen prose).

References

Source: SKILL.md on GitHub

No rule matched.

skilld matched fixed text patterns in SKILL.md and file names. Patterns miss obfuscated code.

skilld run checks every file with the same patterns. It asks for approval before it loads a Skill with a behavior marked Needs approval.

No alerts3mo3 checks · Risk SAFE
  • Gen Agent Trust Hub3mo

    The skill provides instructional guidelines for UX writing, voice and tone, and interface components based on industry design standards. No security risks were identified.

  • Socket3mo

    No alerts

  • Snyk3mo

    Risk: LOW · No issues

Signed by skilld at f958982. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.

Last checked against GitHub last week.

Activeupdated 4 months ago
What it can do
Reads files Edits files
last_verified
2026-07-16
review_by
2027-06-22
All 5 allowed tools
ReadWriteEditGlobGrep

README badge

README badge for rshankras/claude-code-apple-skills/ux-writing