Ux Writing logo

Ux Writing

Community
nahid-sparktales
ux-writing

Write or repair the words inside an interface — button labels, empty states, error messages, confirmation dialogs, form hints, success and loading text — so a reader knows what just happened and what to do next. Use when a screen's copy is being written or reviewed, when an error says something unhelpful, when a confirmation is vague about what it will do, or when users hesitate at a control. Not for marketing or landing-page copy, not for documentation, and not for deciding layout or visual design.

Overview

Publishernahid-sparktales
Repositoryagent-dispatcher
Skill nameux-writing
Stars
58
Forks
5
Bundled files
1
LicenseMIT
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • 1 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

    Published by nahid-sparktales on GitHub. Read the source before you install it.

Installation

Install the Ux Writing AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/nahid-sparktales/agent-dispatcher/tree/main/skills/design/ux-writing
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/nahid-sparktales/agent-dispatcher.git /tmp/agent-dispatcher
mkdir -p .claude/skills
cp -r /tmp/agent-dispatcher/skills/design/ux-writing .claude/skills/ux-writing
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Ux Writing in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Ux Writing on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Ux Writing is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

UX writing

Interface copy is read mid-task by someone who did not come to read. Every word is either doing work — naming an outcome, removing a doubt, offering the next move — or it is in the way.

When this fires

A screen's strings are being written, reviewed or translated; an error, empty state or confirmation is unclear; a control's label does not match what it does. It does not fire for prose the user chose to read: docs, marketing pages, release notes.

Procedure

  1. Collect the actual strings. Grep the component, route or feature for every user-visible string, including the ones not on screen right now — error branches, empty results, disabled reasons, toast text. Copy you never saw is copy you never fixed.
  2. Name the reader and the moment. What were they doing one action ago, what do they already know, and what do they need to decide next. Write to that person, not to a persona.
  3. Fix labels first. A label names the outcome in the reader's words, not the mechanism in yours: "Save changes", not "Submit"; "Delete project", not "Confirm". Avoid Yes/No button pairs — a button that repeats the verb can be answered without re-reading the question.
  4. Rewrite errors to be actionable. Say what happened, why when you actually know, and what the reader can do. Never blame the reader, never show a raw exception or bare status code as the whole message, and keep any support-quotable identifier alongside plain words rather than instead of them. Do not promise a retry, a saved draft or a notification that the code does not perform.
  5. Separate the three empty states. Nothing yet (say what will appear here and give one way to start), filtered to nothing (name the filter and offer to clear it), and failed to load (that is an error, not an empty state — do not let it read as "you have none").
  6. Make confirmations specific. State the object by name, state the consequence, and repeat the verb on the primary button. For anything destructive, say what is lost and whether it can be undone — if the code has no undo, the copy must not imply one. When the copy and the code disagree about reversibility, stop and raise it; do not paper over it with softer wording.
  7. Cut the words carrying no information. Padding ("please", "simply", "just"), fake cheer ("Oops!"), exclamation marks, and restating the screen's own title. Then check the result still sounds like the rest of the product — consistency beats each string being individually clever.
  8. Check every string against what the code does. A label that describes behaviour the function does not have is a bug report, not a copy edit. Report it as a bug.
  9. Check the string survives contact with reality. Longest plausible value, truncation, sentence-case consistency with neighbours, and an accessible name for every icon-only control. If the product is localized, note that changed strings need re-translation and that other languages run longer.

Checklist

  • Every user-visible string in the changed surface was listed, including unrendered branches
  • Buttons name outcomes; no ambiguous Yes/No pair
  • Each error says what happened and what to do, with no raw exception as the message
  • Empty, filtered-empty and failed-to-load are distinguishable
  • Destructive confirmations name the object and state reversibility truthfully
  • Copy matches actual code behaviour; mismatches reported, not smoothed over
  • Icon-only controls have accessible names
  • Translation and length impact noted where it applies

Failure handling

  • The right words depend on behaviour you cannot determine — read the handler. If it is still ambiguous, write the copy that is true of both branches and flag the question rather than inventing the more reassuring version.
  • Legal, consent, pricing, security or privacy wording — propose, do not replace. These are approved text somewhere; stop and ask before changing them.
  • The string is a translation key or appears in many places — check every use before editing. A label reworded for one screen can be wrong on three others.
  • The real problem is the flow, not the sentence — say so. Copy that has to apologize for an interaction is a design finding; writing around it hides the defect.

Evidence to report

The before and after of each changed string, with the file and the state it appears in; the mismatches you found between copy and code; the strings you deliberately left alone and why; and anything that needs approval or re-translation. "Improved the microcopy" with no strings quoted is not evidence.

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the Ux Writing AI skill do?

Write or repair the words inside an interface — button labels, empty states, error messages, confirmation dialogs, form hints, success and loading text — so a reader knows what just happened and what to do next. Use when a screen's copy is being written or reviewed, when an error says something unhelpful, when a confirmation is vague about what it will do, or when users hesitate at a control. Not for marketing or landing-page copy, not for documentation, and not for deciding layout or visual design.

Why use Ux Writing on TypingMind?

Because you install it once and use it with any model. Ux Writing is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Ux Writing in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/nahid-sparktales/agent-dispatcher/tree/main/skills/design/ux-writing. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Ux Writing?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Ux Writing?

As many as you like. As long as a model supports skills, you can use Ux Writing with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Ux Writing AI skill free?

Yes. It is published on GitHub by nahid-sparktales under the MIT license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇