agentsclimarketplace
Sign in

Brand voice analyzer

Skill bg-szy/TOP-SKILLS/skills/claude-skills/brand-voice-analyzer

Analyzes a company's content to extract and codify their brand voice into a comprehensive style guide. Reads website copy, blog posts, emails, and social media to identify tone, vocabulary patterns, sentence structure, personality traits, and word preferences. Generates a brand-voice-guide.md and reviews new content against it.From its SKILL.md

Install
npx -y skills add bg-szy/TOP-SKILLS --skill brand-voice-analyzer

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

2 things to look at

  • no licenseNo license file was found in the repository. Code published without one is not open source by default, so using it at work is a question for whoever answers licensing questions where you are.
  • 5 stars5 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

3.8 KB, 724 tokens by cl100k_base, as published. Nobody here has run it

Brand Voice Analyzer

Ingest a company's existing content, reverse-engineer the voice behind it, produce a definitive brand voice guide any writer can follow, and review new content against an established guide.

Contents

  • references/voice-dimensions.md -- the seven-axis tone spectrum, vocabulary, sentence, personality, audience, and formatting dimensions to extract
  • references/guide-template.md -- required sections, quality standards, and pre-delivery checklist for brand-voice-guide.md
  • references/review-mode.md -- process and output format for scoring content against a guide
  • references/metrics.md -- quantitative and qualitative analysis techniques and citation standards
  • references/edge-cases.md -- handling for thin corpora, multiple voices, regulated industries, rebrands, plus interaction patterns

Workflow: Create a New Voice Guide

  1. Ask which content sources are available (URLs, files, directories, pasted text).
  2. Ask about any known brand attributes, values, or existing guidelines.
  3. Collect the corpus. Pull URLs with WebFetch, locate files with Glob/Read, find public content with WebSearch. Gather website copy, 5-10 blog posts, email campaigns, social posts, sales collateral, and support content. Document each source and its word count.
  4. Confirm corpus size. Require at least 3,000 words across 2+ content types. If smaller, flag the limitation and mark affected findings as lower confidence.
  5. Run the full voice extraction across every dimension in references/voice-dimensions.md, citing source examples per the standards in references/metrics.md.
  6. Present a summary of findings for validation before generating the guide.
  7. Generate the complete brand-voice-guide.md following references/guide-template.md. Verify it against that checklist before delivery.
  8. Offer to review a sample piece of content against the new guide.

Workflow: Review Content

  1. Locate or ask for the brand-voice-guide.md.
  2. Read the content to be reviewed.
  3. Run the review process and produce the scored output defined in references/review-mode.md.
  4. Present findings with actionable, guide-referenced fixes.

Workflow: Update an Existing Guide

  1. Read the current brand-voice-guide.md.
  2. Ingest the new content sources and run extraction on the new material.
  3. Compare findings against the existing guide.
  4. Propose specific updates with rationale.
  5. Apply approved changes with Edit.

Output Standards

  • Primary output: brand-voice-guide.md. For multiple brands: brand-voice-guide-[company-name].md. Deliver review output inline unless a file is requested.
  • Use markdown headers (H1-H4), tables for comparative data, blockquotes for source quotes, code blocks only for literal reproduced text. Bold key terms on first use. No emojis.

Operating Principles

  • Never fabricate examples. Every quoted example must come from actual source content.
  • Never assert voice attributes without evidence. If the data does not support a conclusion, say so.
  • Present findings as observations, not prescriptions, until the user validates them.
  • Distinguish voice (consistent: who we are) from tone (shifts by context: how we adapt) in every guide.
  • When reviewing, be specific: cite the exact word, phrase, or structure and the guide section it touches.
  • Treat the guide as a living document and defer to the user's judgment on subjective calls.

What ships with it

Read from the repository

Just SKILL.md. No reference files, no scripts.

Gives 0 of the 12 instructions most marketing audience skills give in 724 tokens

Counted across 690 of the 894 authors here whose files we hold, read 2026-08-07

  • Apply Poppins font to headingsin 41 of 690, across 6 files
  • Apply Lora font to body textin 41 of 690, across 6 files
  • Use Arial fallback for headingsin 39 of 690, across 4 files
  • Use Georgia fallback for body textin 39 of 690, across 4 files
  • Maintain text hierarchy and formattingin 39 of 690, across 4 files
  • Use accent colors for non-text shapesin 38 of 690, across 3 files
  • Use RGB values for precise color matchingin 38 of 690, across 3 files
  • Use brand colors for primary text and backgroundsin 36 of 690, across 1 file
  • Read product marketing context file before asking questions, starting, or auditingin 35 of 690, across 23 files
  • Use active voice instead of passive voicein 26 of 690, across 10 files
  • Implement or generate appropriate JSON-LD structured datain 24 of 690, across 17 files
  • Prioritize clarity over clevernessin 22 of 690, across 8 files

Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once. Length counted with cl100k_base; the agent that loads this file may tokenize it differently.

Keep looking

Skills are one crate of 325,949. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.