A specialized writing editor that identifies and strips away 24 common AI writing patterns to make text sound authentically human.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install irisgo-humanizer
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 irisgo-humanizer 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?
Humanizer is a professional-grade writing and editing tool designed to eliminate the robotic markers often found in Large Language Model (LLM) outputs. This skill is built upon extensive research from the Wikipedia AI Cleanup project, targeting 24 specific patterns that distinguish AI prose from human writing. By integrating Humanizer into your suite of Openclaw Skills, you can transform sterile, soulless content into engaging, opinionated, and rhythmic text that captures a genuine human voice. It goes beyond simple editing by injecting personality and soul into the writing process.
The skill addresses deep-seated issues like overused AI vocabulary, sycophantic tones, and formulaic structures that often flag content as machine-generated. Whether you are polishing a technical blog post or refining a Wikipedia entry, this addition to your Openclaw Skills provides a systematic way to ensure your writing resonates with readers rather than sounding like a statistical algorithm.
Humanizer Use Cases
- Refining AI-generated blog posts to sound more authentic and less promotional.
- Removing repetitive AI vocabulary like delve, testament, and pivotal from technical documentation.
- Converting sterile, neutral reports into opinionated pieces with a clear first-person perspective.
- Standardizing formatting by removing robotic bolding, em dash overuse, and curly quotes.
- Cleaning up chatbot artifacts such as I hope this helps or knowledge-cutoff disclaimers from final drafts.
How Humanizer Works
- The skill scans the provided text using tools like Grep and Read to identify 24 specific AI-writing patterns documented by researchers.
- It flags content markers such as undue emphasis on significance and vague attributions.
- The agent identifies stylistic and grammatical issues like copula avoidance and the overuse of the rule of three.
- Using its rewriting engine, it replaces AI-isms with natural alternatives, such as using simple is/are verbs instead of serves as.
- It varies sentence length and rhythm to mimic human breathing and thought patterns.
- Finally, it injects personality by acknowledging complexity or using first-person perspectives where appropriate.
Humanizer Setup
To add this capability to your environment, use the following commands to install the Humanizer repository and configure your agent:
# Clone the Humanizer skill repository
git clone https://github.com/IrisGoLab/humanizer-skill
# Install the skill to your Openclaw configuration
openclaw install humanizer
Ensure your agent has permissions for the Read, Write, Edit, Grep, and Glob tools to allow full text-processing capabilities.
Humanizer Data Schema & Taxonomy
The Humanizer skill organizes its refinement process based on a taxonomy of AI writing markers. It evaluates text across these data points:
| Category | Patterns Monitored |
|---|---|
| Content Patterns | Significance inflation, notability claims, -ing endings, promotional language |
| Grammar Patterns | AI vocabulary (delve, tapestry), copula avoidance, negative parallelisms |
| Style Patterns | Em dash overuse, mechanical boldface, inline-header lists, curly quotes |
| Communication | Chatbot artifacts, knowledge-cutoff disclaimers, sycophantic tone |
| Formatting | Title case in headings, emoji usage in bullet points |
Humanizer Advanced Features
- Multi-agent compatibility for collaborative humanization across large document sets.
- Personality Injection Engine to shift tone from neutral reporting to active, opinionated perspective.
- Automatic source verification to replace vague Expert claims with specific, citation-ready attributions.
- Integrated pattern detection for 24 distinct categories based on WikiProject AI Cleanup research.
- Formatting normalization that enforces professional publishing standards for Openclaw Skills users.
SKILL.md
Loading
Related Openclaw Skills
Contract Review
An AI-powered legal analysis tool that reviews contracts for risks, extracts key terms, and suggests redlines based on market standards.
Extract Pic Text
A specialized CLI tool for batch extracting specific text segments from image filenames based on customizable delimiters and positions.
clawctl
A centralized coordination layer for managing tasks, messaging, and activity feeds across autonomous agent fleets.
Stock Analysis Agent Skill
A high-performance AI agent skill for technical stock analysis, trend prediction, and automated market reporting.
Omni Research
A powerful multi-source research tool that queries Perplexity, Grok, and Gemini via your browser sessions without requiring individual API keys.
PCClaw - Windows-Native AI Agent Skills
A powerful suite of Windows-native capabilities designed to bring the Openclaw Skills ecosystem to the PC.
Featured*
