Extract brand voice from published content
brand-voice-extractorskillsetup L2★679
gooseworks-ai/goose-skills ↗Causal-lift measurements
brand-voice-management0pp vs no-skill baselinewith-skill 100% · baseline 100%
Measured by running the task with and without this artifact, K=5, graded by deterministic checks — no LLM judging.
What it does
Extract and document brand voice from published content
Best for
Content teams that want a reusable brand voice playbook extracted from real published examples, to apply consistently across new channels (X, email, outreach) without re-deriving style each time.
Inputs
- · Company name and content sample URLs (blog, landing pages, case studies, 10-20 pages)
- · Or: content inventory JSON from prior site-content-catalog run
Outputs
- · VOICE PROFILE markdown with documented tone, vocabulary, sentence structure, formatting patterns, persona, and audience
- · Tone spectrum positions (formality, emotion, authority, humor, directness)
- · Unique vocabulary and banned patterns
- · Sentence/paragraph length norms
Requires
- · WebFetch or firecrawl (content retrieval)
- · Optional: site-content-catalog output (inventory)
Preconditions
- · 10-20 content samples available (blog posts, landing pages, case studies)
- · Diverse topic mix to ensure consistency across subjects
Failure modes
- · Over-interpreting single anomalous post as brand voice
- · Voice analysis frozen at old content (brand evolves)
- · No separation of 'launch voice' from 'working docs voice' when both exist
Trust signals
- · 6 voice dimension categories (tone, vocabulary, sentence structure, formatting, persona, audience)
- · Formality spectrum provided (casual ↔ professional ↔ academic)
- · Power words extraction methodology
- · Banned patterns documented (what they never do)
- · Selection heuristic (8-10 blog posts, 2-3 landing pages, 2-3 case studies, mix of recent and older)