A sophisticated text processing skill that identifies and removes AI-generated signatures to transform robotic LLM output into natural human writing.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install humanizer-2
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install humanizer-2 using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is Humanizer AI Text Refinement?
Humanizer is a professional-grade writing editor designed to strip away the tell-tale signs of AI-generated content. By analyzing text against 28 distinct linguistic patterns and a database of over 560 overused AI terms, it identifies artifacts like significance inflation, superficial analyses, and formulaic structures. This component of Openclaw Skills ensures that content retains its technical accuracy while adopting the nuance and personality expected from a human author.
At its core, the skill evaluates the stylometric signature of text, looking for the metronomic rhythm typical of large language models. By adjusting sentence length variation and removing sychophantic chatbot filler, Humanizer produces text that is indistinguishable from professional human copy. It is an indispensable tool for anyone looking to refine AI drafts into high-quality, engaging material.
Humanizer AI Text Refinement Use Cases
- Scoring marketing drafts to determine their AI-detection probability before publication.
- Rewriting technical documentation to remove repetitive Tier 1 AI vocabulary like "delve" and "tapestry."
- Optimizing automated content pipelines by integrating humanization rules into Openclaw Skills workflows.
- Analyzing academic or professional writing for burstiness and sentence length uniformity.
- Removing sycophantic tones and generic chatbot artifacts from customer-facing communications.
How Humanizer AI Text Refinement Works
- The input text is scanned against 28 specific pattern detectors to identify content inflation and promotional language.
- A statistical analysis is performed to calculate burstiness and the type-token ratio, identifying low-variance patterns common in LLMs.
- The text is cross-referenced with a three-tier vocabulary library containing over 500 dead giveaway terms and suspicious phrases.
- Problematic segments are flagged for rewriting, focusing on replacing copula avoidance with natural "is/has" structures.
- The skill applies human-centric writing principles, such as adding personality and varying sentence rhythm, to the final output.
- A comprehensive report or humanized version of the text is generated, often including a score from 0-100 reflecting its AI-like qualities.
Humanizer AI Text Refinement Setup
To integrate this tool within your Openclaw Skills environment, follow these installation and configuration steps:
# Install the package and dependencies
npm install humanizer-cli
# Verify the installation by scoring a sample text
echo "In the realm of digital innovation..." | node src/cli.js score
# Generate a full markdown report for a specific file
node src/cli.js report draft.txt > report.md
Humanizer AI Text Refinement Data Schema & Taxonomy
The skill processes text and returns structured data regarding its linguistic composition and AI-signature density.
| Field | Type | Description |
|---|---|---|
| AI Score | Integer (0-100) | Probability that the text was generated by an LLM. |
| Burstiness | Float (0.0-1.0) | Measurement of sentence length variation; higher is more human. |
| TTR | Float (0.0-1.0) | Type-Token Ratio; measures vocabulary diversity. |
| Patterns Found | Array | List of specific AI artifacts detected (e.g., Significance Inflation). |
| Suggestions | Markdown List | Priority-grouped recommendations for manual or automatic fixes. |
Humanizer AI Text Refinement Advanced Features
- Auto-fix mode which leverages CLI flags to automatically replace Tier 1 vocabulary with human alternatives.
- Always-on personality templates designed for Openclaw Skills to ensure agents never adopt robotic writing styles.
- Programmatic JSON output for integration into CI/CD pipelines to audit content quality automatically.
- Statistical-only mode for researchers needing raw stylometric data without text modification.
SKILL.md
Loading
Related Openclaw Skills
Soaring Weather
A professional-grade thermal and gliding weather forecast tool using an 11-parameter scoring engine to predict soaring conditions.
Remember Me
A sophisticated memory layer that tracks user preferences, goals, and boundaries to provide highly personalized AI interactions over time.
Omarchy
Omarchy provides a comprehensive suite of system wrappers and operating guardrails to ensure AI agents manage local environments using safe, native workflows.
Glance
Glance is an AI-extensible personal dashboard that allows developers to create custom widgets using natural language, where AI agents handle the underlying data collection and caching.
Context Pruner
Advanced context management for DeepSeek models featuring auto-compaction and dynamic relevance-based optimization.
OpenMM Exchange Setup
An interactive configuration guide for setting up secure API credentials and environment variables for the OpenMM trading framework.
Featured*
