From majestic-marketing
Performs forensic style analysis on writing samples, measuring sentence length, rhythm, punctuation DNA, vocabulary profile, tone, rhetorical devices, and outputs Style DNA report.
npx claudepluginhub majesticlabs-dev/majestic-marketplace --plugin majestic-marketingThis skill is limited to using the following tools:
Analyze: $ARGUMENTS
Reverse-engineers voice profiles from text samples by analyzing sentence structure, vocabulary, tone markers, and usage patterns like contractions and pronouns. Outputs YAML profiles for formality, confidence, warmth.
Analyzes user's Yuque documents to extract writing style profile—vocabulary habits, sentence patterns, tone—for consistent voice when generating new content. Useful for style analysis requests.
Analyzes writing structure via paragraph decomposition into openings, sentence shapes, and tropes; aggregates patterns to detect repetition during editorial reviews.
Share bugs, ideas, or general feedback.
Analyze: $ARGUMENTS
Audience: Writers and content creators who need a precise, measured voice fingerprint for use with style-writer and humanizer.
Goal: Extract quantitative style metrics from writing samples and produce a Style DNA report. Measure, don't guess.
Strip the file to prose only before measuring anything. Remove:
...)Save the cleaned prose to a temp file for all subsequent analysis.
Report: Total prose word count.
Extract individual sentences from the prose. Handle:
Measure and report:
| Metric | How |
|---|---|
| Total sentences | Count |
| Average sentence length | words / sentences |
| Median sentence length | Middle value of sorted lengths |
| Min / Max | Extremes |
| Short sentence ratio | % of sentences under 8 words |
| Sentence length distribution | 4 bands: 1-7, 8-15, 16-25, 26+ words with counts and percentages |
Describe the rhythm pattern: How do short and long sentences alternate? Do short sentences cluster? Are long sentences followed by punchy landings?
Count across the full prose text:
| Mark | Measure |
|---|---|
| Commas | Total and per 100 words |
| Semicolons | Total and per 100 words |
| Colons | Total and per 100 words |
| Em dashes (— and –) | Total |
| Exclamation marks | Total |
| Parenthetical asides | Total, per 100 words, list each one |
| Rhetorical questions | Total, per 100 words, list each one |
Identify the punctuation fingerprint: Which marks carry the structural load? What's absent? Note any unusual ratios (e.g., colon-to-semicolon ratio).
| Metric | How |
|---|---|
| Total prose paragraphs | Count (exclude pure list paragraphs) |
| Average paragraph length | Words |
| One-sentence paragraph ratio | % of paragraphs with 0-1 sentence-ending punctuation |
| Metric | How |
|---|---|
| Total words | Count alpha-only tokens |
| Unique words | Count distinct |
| Lexical diversity | unique / total |
| Long words (8+ chars) | Count and percentage |
| Common word ratio | % that are top-100 English function words |
| Top 15 long words | With frequency counts |
Describe the vocabulary character: Academic or accessible? Jargon-heavy or plain? Are key terms repeated deliberately (thematic anchoring) or is vocabulary varied?
Search for: probably, perhaps, maybe, might, could be, I think, kinda, sort of, likely, arguably, roughly, nearly, pretty, fairly, basically, genuinely, honestly, admittedly, I guess, I don't think
Report total, per 100 words, and each instance. Note whether hedges appear on claims, predictions, or data.
Search for and quantify each of these patterns. Not all will be present. Report what IS there and what IS NOT.
Search for:
Describe the macro-structure:
How does the piece begin? (Bold thesis? Scene setting? Credential stack? Anecdote?)
How are sections structured internally? What's the repeating unit?
How does the piece end? (Call to action? Honest concession? Restatement? Promise?)
Write the full Style DNA report as a markdown file saved next to the source file, named [source-filename]-style-dna.md.
The report MUST include:
If analyzing multiple files by the same author, add a Cross-Article Comparison section with a comparison table and analysis of which traits are stable fingerprints vs. sliding dials.
After presenting the report, ask the user:
Style DNA report ready. Want me to save it as
[source-filename]-style-dna.mdnext to the source file?
Wait for confirmation before writing. If the user says yes, save the full report. If analyzing multiple files, also offer to save the cross-article comparison as a separate file.
Produces Style DNA reports consumed by other skills:
[writing samples] --> style-forensics --> Style DNA report
|
+--> style-writer (write new content in that voice)
+--> humanizer (preserve style when stripping AI tells)
+--> copy-editor (respect voice during editing)
style-writer to write new content matching the voicebrand-voice for qualitative brand guidelines, copy-editor for grammar review