mcpbeat

Writing User Facing Copy

posthog/writing-user-facing-copy

> Use ALWAYS before writing or editing user-facing copy, and whenever a code change adds or changes a string a user will see. Also carries feature-naming rules that are easy to get wrong, most importantly how to talk about the Wizard (name it "Wizard" once, then call it "the setup agent" / "the agent"). Not for internal code comments, commit messages, or variable names.

1k tokens
context cost
the whole folder, loaded on every use
1
files
instructions only
0
copies elsewhere
how many repositories repackaged it
690
stars on the repo
on the repository, not the skill itself

Install

one command, takes just this skill from the repository
npx skills add https://github.com/PostHog/posthog --skill writing-user-facing-copy

The instruction itself

9 sections, as written by the author

Writing user-facing copy

This is the operational gate for anything a person reads in the product or around it.

Run it before writing or editing user-facing text, and whenever a code change introduces or changes a visible string.

Applies to: UI labels, buttons, form fields, tooltips, empty states, error and success messages, toasts and notifications, onboarding flows, emails, docs, and support replies.

Does not apply to: code comments, commit messages, log lines, variable/function names, or other developer-only text.

Voice

Write the way a person would. Neutral and humane.

  • Sentence case. Capitalize only the first word and proper nouns. "Save as view", not "Save As View". "Product analytics", not "Product Analytics".
  • Be direct and friendly. Say what happened and what to do next.
  • Plain language, no jargon. Use the label the user sees, not the internal name. surveyPopupDelaySeconds becomes "Delay the survey popup".
  • Don't editorialize. State what is, not how exciting it is. Cut "powerful", "seamless", "effortless", "simply", "just", "easily", "supercharge", "unlock".
  • No sales-y or edgy copy. No marketing hooks, no clever one-liners, no hype.
  • American English spelling. "color", "analyze", "canceled".

Specific rules

No em-dashes

Do not use em-dashes (—) anywhere in user-facing copy.

Do not substitute an en-dash either.

Prefer rewriting the sentence so the dash isn't needed. If a connector is unavoidable, use a hyphen with spaces, a comma, a colon, or split into two sentences.

  • ❌ "Save this view — you can reuse it later."
  • ✅ "Save this view. You can reuse it later."
  • ✅ "Save this view to reuse it later."

Also avoid the sentence _shapes_ that lean on that dash, because they read as machine-written:

  • ❌ "This isn't just a filter, it's a saved view." (the "not just X, but Y" construction)
  • ❌ "It's fast, it's simple, it's yours." (rule-of-three padding)
  • ❌ Hedging preambles like "It's worth noting that…", "Keep in mind that…".

Write the plain version instead.

No punchy or clever phrasing

This is the failure mode that slips through even when you know the rule, so check for it specifically. It applies to every string, including headings, section titles, and button labels, not just body text.

The tells:

  • Antithesis ("X, not Y" / "X, they don't Y") used for effect. ❌ "Errors guide, they don't dead-end." → ✅ "Errors point to a next step."
  • A heading written to sound weighty. ❌ "The rules that are easy to break" → ✅ "Common mistakes".
  • A clever or dramatic closing line. ❌ "Never leave the user staring at a failure." → ✅ "Say what happened and what to do next."
  • Filler that sounds like a principle but adds nothing. ❌ "This is a tool people use to get work done." → cut it.
  • Emphasis the sentence doesn't need. ❌ "This will completely transform how you work." → ✅ "This changes how X works."

After drafting, re-read and cut anything that's there to sound good rather than to inform. If a line would feel out of place said plainly out loud to a colleague, rewrite it.

Errors and empty states point to a next step

Say what happened and give the next action. Don't stop at reporting that something failed.

  • ❌ "Something went wrong."
  • ✅ "Couldn't load your insights. Refresh the page, and if it keeps happening contact support."

How to talk about features

Use the names users see, and stay consistent across every surface.

Wizard

The setup tool is named Wizard. It confuses users as a description, because "wizard" reads like an old-style step-by-step form, not an AI. Users already have a mental model for AI agents.

  • Use "Wizard" only as the feature's proper name (the thing you're pointing at).
  • To explain what it does, always call it "the setup agent" or "the agent".

Examples:

  • ✅ "Wizard sets up PostHog for you. The setup agent installs the SDK and wires up your first events."
  • ✅ "Ask the agent to add error tracking."
  • ❌ "Use the wizard to walk through setup." (using "wizard" as a description)
  • ❌ "The wizard will guide you through each step." (reinforces the wrong mental model)

For any other feature, use its product-facing name exactly as it appears in the UI, and describe it in the terms users already understand.

When unsure

If you can't tell whether copy reads well, or whether a term is the right user-facing name, ask a human before shipping it.

How to use it

Copy the folder

Take posthog/writing-user-facing-copy from the repository into ~/.claude/skills for personal use, or into .claude/skills inside a project.

Check the name does not clash

The agent identifies a skill by the name field in its header. Two skills with the same name cannot sit side by side — one of them will be ignored.