Brand Voice Profiler

The second foundation skill in the SCALE AI skills library. Brand DNA Builder captures high-level voice signals (tone profile, signature language). Brand Voice Profiler goes deeper — it extracts actual sentence patterns from real brand copy, builds a do/don't word list, and calibrates voice rules that every downstream copy skill (Hook Writer, Ad Script Writer, Email Newsletter Writer, etc.) reads before writing.

Critical Rules

• ALWAYS read ./brand/brand-dna.md first. If it doesn't exist, tell the user to run Brand DNA Builder first — do NOT proceed without brand DNA.
• ALWAYS scrape 2-3 real copy samples from the brand's website (product pages, about page, blog, landing pages) before asking the user to rate anything.
• ALWAYS show the user at least 3 example rewrites of the same sentence in different voice directions and ask them to pick the closest match. Calibration is the whole point.
• NEVER invent voice rules that aren't grounded in real copy samples or explicit user confirmation.
• Save the final file to ./brand/brand-voice.md.

Process

Step 1: Read brand DNA

Read ./brand/brand-dna.md from the project root. If it doesn't exist, stop and tell the user:

"I need a brand foundation file before I can build a voice profile. Run the brand-dna-builder skill first — it takes about 2 minutes and creates the file I need to read. Once you have brand/brand-dna.md, come back and I'll build the detailed voice rules on top of it."