Brand voice — voice guidelines, tone-by-channel adaptation, and before/after rewrites that keep identity consistent across pages, emails, and posts. Use when user says "brand voice", "tone of voice", "style guide".
Installs into .claude/skills of the current project.
Are you the author of Copy Brand Voice?
Add the live security badge to your README. It updates with every re-scan.
[](https://www.skillsdirectory.com/skills/meshcode-ai-copy-brand-voice)
---
name: copy-brand-voice
description: Brand voice — voice guidelines, tone-by-channel adaptation, and before/after rewrites that keep identity consistent across pages, emails, and posts. Use when user says "brand voice", "tone of voice", "style guide".
license: MIT
metadata:
source: "anthropics/skills brand-guidelines (MIT)"
category: copy
---
# Brand Voice — Guidelines and Application
**The voice is always the same thing; the tone is what varies by situation.** Confuse the two and every team ends up writing in its own style.
## Setting the 4 Voice Axes
Pick one pole per axis and state the forbidden spectrum explicitly.
- Formal ↔ Friendly (courtly register vs casual "—which pole)
- Restrained ↔ Expressive / Expert ↔ Plain / Confident ↔ Humble
Attach **two example pairs per axis (supports / does not support)**. Rules alone confuse both models and people.
## Guideline Document Structure
1. One-sentence voice declaration ("We speak like X, we never speak like Y")
2. Per-axis criteria + example sentences
3. Vocabulary table: 10 recommended / 10 forbidden words (forbidden list must include competitor and industry clichés)
4. Grammar rules: tense, number notation, capitalization policy
5. Variation by channel: Product UI = minimum / Email = medium / Social = maximum friendliness (±1 step only)
6. Six-item pre-publish self-proofing checklist
## Tone Response (by event)
Incident notice = calm and accountable / Good news = upbeat / Price increase = honest. Sentences written while emotions run high are off-voice — check them against the axes before sending.
## Output
One-sentence voice declaration + a 4-axis setup table (with examples) + 10 recommended/forbidden word pairs + 5 before/after rewrites of existing copy (axis rationale annotated).