| name | humanizer-pro |
| description | Transform AI-generated text into authentic human writing. Detects and removes 24 AI patterns, replaces 500+ AI vocabulary terms, analyzes statistical signals (burstiness, vocabulary diversity), and injects personality through strategic misspellings, parenthetical asides, tangents, and random thoughts. Use when humanizing content, checking for AI tells, removing robotic patterns, adding natural voice, making text sound less polished, or when asked to write like a specific person. Works with social posts, articles, emails, marketing copy, documentation—any text that needs to sound genuinely human. |
Humanizer Pro
Transform AI-generated text into authentic human writing by removing robotic patterns and injecting natural personality.
What This Skill Does
Core Capabilities:
- Detects 24 AI writing patterns (significance inflation, promotional language, filler phrases, etc.)
- Identifies 500+ AI vocabulary terms across 3 severity tiers
- Analyzes statistical signals (burstiness, type-token ratio, sentence uniformity)
- Removes chatbot artifacts and sycophantic tone
- NEW: Injects personality through parenthetical asides, strategic typos, tangents, and random thoughts
Based on: Wikipedia's "Signs of AI writing", Copyleaks research, and real-world pattern analysis.
Quick Start
Basic Humanization
When asked to humanize text:
- Scan for patterns → Check all 24 patterns in
references/patterns.md
- Check vocabulary → Flag Tier 1/2/3 terms from
references/vocabulary.md
- Analyze statistics → Compute burstiness, TTR, sentence variance
- Rewrite → Remove AI-isms, add personality
- Verify → Read aloud, check if it sounds natural
Adding Personality
Use the new personality injection features from references/personality-injection.md:
- Parenthetical asides → (honestly, this part gets me every time)
- Strategic misspellings → Natural typos that don't hurt credibility
- Tangents → "wait, that reminds me..." moments
- Random thoughts → Going off script with authentic reactions
When to Use Each Component
Always Check (Core Patterns)
- Tier 1 vocabulary → DEAD GIVEAWAYS, ban completely (delve, tapestry, vibrant, seamless, etc.)
- Filler phrases → "in order to" → "to", "due to the fact that" → "because"
- Chatbot artifacts → "Great question!", "I hope this helps!", "Let me know if..."
- Generic conclusions → "The future looks bright", "Exciting times lie ahead"
Check When Relevant
- Promotional language → If writing about places, products, services
- Significance inflation → If discussing history, events, milestones
- Vague attributions → If making claims without sources
- Em dash overuse → If text has many — dashes — everywhere
Advanced Analysis
For comprehensive humanization or when scoring text:
- Read
references/patterns.md → All 24 patterns with examples
- Read
references/vocabulary.md → Complete AI term database
- Read
references/statistical-signals.md → Burstiness, TTR, readability formulas
- Read
references/personality-injection.md → How to add human touches
Core Principles
Write Like a Human, Not a Press Release
- Use "is" and "has" freely — "serves as" is pretentious
- One qualifier per claim — don't stack hedges
- Name your sources or drop the claim
- End with something specific, not vague optimism
Add Actual Personality
- Have opinions → React to facts, don't just report them
- Vary rhythm → Short. Then longer ones that meander a bit.
- Acknowledge complexity → "I genuinely don't know how to feel about this"
- Let mess in → Perfect structure feels algorithmic
- Use contractions → "don't", "won't", "it's" (natural speech)
- Sentence fragments → When it makes sense. Like this.
Cut the Fat
Remove these automatically:
- "In order to" → "to"
- "Due to the fact that" → "because"
- "At this point in time" → "now"
- "It is important to note that" → (just say it)
- "In the event that" → "if"
The 24 Patterns (Quick Reference)
For full details with examples, see references/patterns.md.
| # | Pattern | What to Watch For |
|---|
| 1 | Significance inflation | "marking a pivotal moment in the evolution of..." |
| 2 | Notability name-dropping | Listing media outlets without specific claims |
| 3 | Superficial -ing analyses | "...showcasing... reflecting... highlighting..." |
| 4 | Promotional language | "nestled", "breathtaking", "stunning", "renowned" |
| 5 | Vague attributions | "Experts believe", "Studies show", "Industry reports" |
| 6 | Formulaic challenges | "Despite challenges... continues to thrive" |
| 7 | AI vocabulary (500+ words) | "delve", "tapestry", "landscape", "showcase" |
| 8 | Copula avoidance | "serves as", "boasts" instead of "is", "has" |
| 9 | Negative parallelisms | "It's not just X, it's Y" |
| 10 | Rule of three | "innovation, inspiration, and insights" |
| 11 | Synonym cycling | "protagonist... main character... central figure..." |
| 12 | False ranges | "from the Big Bang to dark matter" |
| 13 | Em dash overuse | Too many — dashes — everywhere |
| 14 | Boldface overuse | Mechanical emphasis everywhere |
| 15 | Inline-header lists | "- Topic: Topic is discussed here" |
| 16 | Title Case headings | Every Main Word Capitalized |
| 17 | Emoji overuse | 🚀💡✅ decorating professional text |
| 18 | Curly quotes | "smart quotes" instead of "straight quotes" |
| 19 | Chatbot artifacts | "I hope this helps!", "Let me know if..." |
| 20 | Cutoff disclaimers | "As of my last training...", "While details are limited..." |
| 21 | Sycophantic tone | "Great question!", "You're absolutely right!" |
| 22 | Filler phrases | "In order to", "Due to the fact that" |
Vocabulary Tiers
Tier 1 (Dead Giveaways - NEVER USE)
delve, tapestry, vibrant, crucial, comprehensive, meticulous, embark, robust, seamless, groundbreaking, leverage, synergy, transformative, paramount, multifaceted, myriad, cornerstone, reimagine, empower, catalyst, invaluable, bustling, nestled, realm
Tier 2 (Suspicious - Use Sparingly)
furthermore, moreover, paradigm, holistic, utilize, facilitate, nuanced, illuminate, encompasses, catalyze, proactive, ubiquitous, quintessential
Tier 3 (Context-Dependent - Watch Density)
landscape (abstract), journey (metaphorical), ecosystem, framework, roadmap, touchpoint, pain point, streamline, optimize, scalable
See references/vocabulary.md for the complete 500+ term database.
Statistical Signals
Check these when doing comprehensive analysis:
| Signal | Human | AI | Why |
|---|
| Burstiness | 0.5-1.0 | 0.1-0.3 | Humans write in bursts; AI is metronomic |
| Type-token ratio | 0.5-0.7 | 0.3-0.5 | AI reuses vocabulary |
| Sentence length CoV | High | Low | AI makes same-length sentences |
| Trigram repetition | <0.05 | >0.10 | AI reuses 3-word phrases |
See references/statistical-signals.md for formulas and implementation.
Personality Injection (NEW)
When asked to add human touches or make text sound less perfect:
1. Parenthetical Asides
Add authentic reactions in parentheses:
- (honestly, this part gets me every time)
- (which, let's be real, is kind of absurd)
- (I still don't fully understand why)
Rule: Use for genuine reactions, not filler. 1-3 per 500 words max.
2. Strategic Misspellings
Natural typos that don't hurt credibility:
- Acceptable: "teh" → "the", "recieve" → "receive", "definately" → "definitely"
- Avoid: Professional contexts, names, technical terms, key claims
Rule: 1-2 per 1000 words in casual contexts only.
3. Tangents and Going Off Script
Let thoughts wander naturally:
- "Speaking of which..."
- "That reminds me..."
- "Okay, slight tangent..."
- "Wait, this connects to..."
Rule: Allow 1-2 per longer piece (1000+ words).
4. Random Thoughts
Inject authentic human processing:
- "Honestly didn't think this would work but..."
- "I keep coming back to..."
- "There's something unsettling about..."
- "I genuinely don't know how to feel about this"
Rule: Use when genuine uncertainty or strong reaction fits.
See references/personality-injection.md for comprehensive examples and context guidance.
Process
- Read the input text carefully
- Identify all pattern instances (use the 24-pattern checklist)
- Flag vocabulary issues (Tier 1 = remove, Tier 2 = reduce, Tier 3 = watch density)
- Check statistics (if comprehensive analysis requested)
- Rewrite problematic sections
- Add personality (if requested or if text is sterile)
- Verify naturally (read aloud, check rhythm)
- Present humanized version with brief change summary
Example Transformation
Before (AI-sounding)
Great question! Here is an overview of sustainable energy. Sustainable energy serves as an enduring testament to humanity's commitment to environmental stewardship, marking a pivotal moment in the evolution of global energy policy. In today's rapidly evolving landscape, these groundbreaking technologies are reshaping how nations approach energy production, underscoring their vital role in combating climate change. The future looks bright. I hope this helps!
After (Humanized)
Solar panel costs dropped 90% between 2010 and 2023 (according to IRENA data). That single fact explains why adoption took off—it stopped being an ideological choice and became an economic one.
Germany gets 46% of its electricity from renewables now. The transition is happening, but it's messy and uneven, and the storage problem is still mostly unsolved.
Changes Made
- Removed "Great question!" and "I hope this helps!" (chatbot artifacts)
- Removed "serves as an enduring testament" (significance inflation)
- Removed "marking a pivotal moment" (AI vocabulary)
- Removed "rapidly evolving landscape" (AI vocabulary)
- Removed "groundbreaking", "underscoring", "vital" (AI vocabulary)
- Removed "The future looks bright" (generic conclusion)
- Added specific data points and sources
- Added personal observation (parenthetical)
- Added acknowledgment of complexity
- Varied sentence rhythm
Using the CLI Tool
The scripts/humanize.js tool provides command-line access:
node scripts/humanize.js score "Your text here"
node scripts/humanize.js analyze -f draft.md
node scripts/humanize.js suggest article.txt
node scripts/humanize.js fix --autofix -f article.txt
Always-On Mode
To make this skill's rules your default writing style (not just when explicitly asked to humanize):
- Ban Tier 1 vocabulary completely from all writing
- Kill filler phrases automatically
- No sycophancy, chatbot artifacts, or generic conclusions
- Vary sentence length, have opinions, use concrete specifics
- If you wouldn't say it in conversation, don't write it
These rules can be added to SOUL.md or agent personality config for permanent application.
Tips for Effective Humanization
- Start with Tier 1 vocabulary → Easiest wins, most obvious AI tells
- Check for em dashes → Quick scan for — overuse
- Read the last paragraph → Generic conclusions are common
- Listen to rhythm → Read aloud, hear if it sounds robotic
- Don't over-polish → Some imperfection is human
- Match the context → Formal documents need different treatment than social posts
- Preserve meaning → Remove patterns without changing core message
Troubleshooting
Text still sounds robotic after fixing patterns?
→ You removed AI tells but didn't add personality. Review references/personality-injection.md.
Humanization made text too casual?
→ Personality injection should match context. Use fewer asides/typos in formal writing.
Text is TOO perfect now?
→ Add strategic imperfection: vary sentence length, include a tangent, acknowledge uncertainty.
Not sure if a word is AI vocabulary?
→ Check references/vocabulary.md. When in doubt, ask "Would I say this in conversation?"
