Content & Brand Voice

Content quality and brand voice enforcement for marketing, documentation, and user-facing writing.

Overview

Key Rules

Tone & Voice (Semantic)

• brand-voice-check — LLM validates professional, approachable tone

Quality (Static)

• no-placeholder-text — Blocks TODO, FIXME, Lorem ipsum
• no-broken-links — Detects empty links, example.com URLs
• no-double-spaces — Catches formatting issues
• heading-hierarchy — Warns on H4+ headings

Guidelines (Injected)

• writing-style — Lead with important info, avoid weasel words
• inclusive-language — allowlist/blocklist, they/them, accessibility

Quality Gates

quality_gates:
  content:
    max_reading_level: 10
    max_length: 5000
    min_length: 50
    tone_check: true

Knowledge: Publishing Checklist

Before publishing:

1. Spell check and grammar check passed
2. All links tested and working
3. All images have alt text
4. Reading level appropriate (Grade 8-10)
5. No placeholder text remaining
6. Reviewed by another person
7. SEO metadata complete

Sample Rules

- name: brand-voice-check
  category: style
  enforcement: semantic
  detect:
    type: semantic
    criteria: |
      Professional but approachable
      Active voice preferred
      Clear and direct — no unexplained jargon
    threshold: 0.8
  action: warn
 
- name: no-placeholder-text
  category: quality
  enforcement: static
  detect:
    type: pattern
    match: "\\b(TODO|FIXME|Lorem ipsum|XXX)\\b"
  action: block

Install

manifest install content-brand-voice

View full manifest on GitHub →