A sophisticated writing editor for Openclaw Skills that identifies AI writing signatures and rewrites text to sound naturally human.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install ai-humanizer-2-1-0
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 ai-humanizer-2-1-0 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?
The Humanizer skill is a specialized tool for Openclaw Skills designed to bridge the gap between algorithmic output and natural human prose. By leveraging 24 distinct pattern detectors and statistical analysis, it identifies the hallmarks of large language models—such as predictable sentence rhythms and specific 'AI vocabulary'—and replaces them with specific, personality-driven writing. It is an essential utility for anyone looking to ensure their AI-generated content resonates with real readers rather than sounding like a sterile press release.
This skill goes beyond simple word swapping; it analyzes the structural DNA of the text, including burstiness and type-token ratios, to ensure the output mimics the natural variability of human thought. Whether you are refining technical documentation or creative blog posts, this addition to your Openclaw Skills library provides the linguistic nuance required to make writing sound like it was written by a person, not a model.
Humanizer Use Cases
- Rewriting AI-generated drafts to sound more natural, specific, and authentic.
- Scoring existing articles and essays to identify the probability of AI detection.
- Stripping away chatbot artifacts like 'I hope this helps' and sycophantic greetings.
- Enhancing content readability by diversifying sentence lengths and vocabulary usage.
- Implementing a 'Human-First' persona for agents using Openclaw Skills through Always-On mode.
How Humanizer Works
- The skill scans the input text for 24 specific patterns, including significance inflation, promotional language, and formulaic challenges.
- It performs a deep dive into vocabulary, checking against a database of 500+ terms frequently overused by AI models.
- Statistical indicators are calculated to measure text burstiness, sentence length variation, and trigram repetition.
- Problems are categorized into content, language, style, communication, or filler groups.
- The tool generates rewrite suggestions or applies auto-fixes to create a final version that preserves the original meaning while adding human-like personality.
- A final score from 0-100 is assigned to represent the text's human-like qualities.
Humanizer Setup
To integrate Humanizer with your Openclaw Skills environment, ensure you have the necessary dependencies and use the following CLI commands:
# Score text to determine AI-likelihood
echo "Your text here" | node src/cli.js score
# Generate a comprehensive analysis report
node src/cli.js analyze -f draft.md
# Apply humanization suggestions with auto-fixes
node src/cli.js humanize --autofix -f article.txt
# Output statistical data in JSON format
node src/cli.js analyze --json < input.txt
Humanizer Data Schema & Taxonomy
The Humanizer skill organizes its analysis based on four primary statistical signals and a multi-tier vocabulary taxonomy:
| Signal | Human Benchmark | AI Benchmark | Utility |
|---|---|---|---|
| Burstiness | 0.5 - 1.0 | 0.1 - 0.3 | Measures sentence rhythm |
| Type-Token Ratio | 0.5 - 0.7 | 0.3 - 0.5 | Measures vocabulary diversity |
| Sentence Length Var | High CoV | Low CoV | Identifies uniform structures |
| Trigram Repetition | < 0.05 | > 0.10 | Detects repetitive phrasing |
It also monitors three tiers of vocabulary, ranging from Tier 1 'Dead Giveaways' (e.g., delve, tapestry, robust) to Tier 2 'Suspicious Density' terms (e.g., furthermore, holistic).
Humanizer Advanced Features
- Always-On Mode: Copy-paste templates for Openclaw Skills SOUL.md to ensure agents consistently write with human personality.
- Statistical Profiling: Real-time calculation of readability and stylometric indicators to provide deep insights into writing style.
- Markdown Reporting: Generates full documentation of writing issues directly into GitHub-flavored markdown for easy review.
- Multi-Agent Support: Can be triggered by other tools within the Openclaw Skills ecosystem to audit and refine outgoing communications.
- Custom Auto-fixes: Automated removal of common filler phrases and 'chatbot artifacts' to streamline the editing process.
SKILL.md
Loading
Related Openclaw Skills
Academic Researcher
A specialized AI research assistant designed to streamline literature reviews, analyze scholarly papers, and automate academic writing workflows.
social-writer
A specialized skill for turning technical documentation and dev logs into opinionated, high-engagement social media content.
Social Reader
A comprehensive social media scraping and automation skill supporting real-time interaction and scheduled batch monitoring.
Safe Backup
A cross-platform utility for securely backing up OpenClaw state directories and workspaces while automatically excluding sensitive credentials.
Gowok
Gowok is a comprehensive Golang library designed to eliminate repetitive boilerplate when building project starters and frameworks.
ngamux
A lightweight and efficient HTTP router for Go designed for building high-performance web services with clean route organization.
Featured*
