Brand voice
Skill euphoriawolf/ai-assisted-app-kit/skills/new-app/templates/project-kit/brand-voice
Write and review customer-facing copy for {{APP_NAME}} so it matches the brand voice in BRAND.md. Use before shipping any marketing page, UI string, email, error message, or docs copy for {{APP_NAME}}.From its SKILL.md
npx -y skills add euphoriawolf/ai-assisted-app-kit --skill brand-voiceAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
One thing to look at
- 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
SKILL.md
1.5 KB, 332 tokens by cl100k_base, as published. Nobody here has run it
{{APP_NAME}} — brand voice
Enforces BRAND.md (repo root), the source of truth for how {{APP_NAME}} sounds. When voice and
layout disagree, voice wins.
Before writing or editing ANY customer-facing copy
- Read
BRAND.md— especially §0 the two non-negotiable mechanics rules and §1 the one-line brief. - Draft in that voice.
- Run the §pre-publish checklist at the end of
BRAND.md.
The two non-negotiables (restated so they're never missed)
- {{NON_NEGOTIABLE_1}} (default: never use the em dash
—in any customer-facing copy — it reads as an AI-content marker; use a period, comma, colon, or restructure). - {{NON_NEGOTIABLE_2}} (the one product-specific rule that matters most for {{APP_NAME}}).
Quick reference
- CTAs: start with a verb, name the outcome.
- Errors: what happened, why, how to fix.
- Empty states: what this is, why it's empty, how to start.
- Lexicon: use canonical terms; avoid the banned words. See
BRAND.md§lexicon.
Reviewing copy
When asked to review, quote the offending line, name which rule it breaks, and offer a fix. Do not rewrite silently. Flag any em dash, any banned term, and any copy doing a job a visual/example should do instead.
What ships with it
Read from the repository
Just SKILL.md. No reference files, no scripts.