The Humanization skill automates ZeroGPT checks and iteratively rewrites essays to safely lower AI-detection scores below 70% while maintaining an academic tone.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install humanization
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 humanization 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 Humanization?
The Humanization skill is a robust browser-automation utility built for Openclaw Skills to help writers and academic professionals streamline essay compliance. Operating as an intelligent, loop-based optimization agent, the skill checks drafts against ZeroGPT and systematically refines text to lower the AI detection score below 70%. It bridges the gap between high-quality AI assistance and strict educational guidelines by preserving academic terminology, citations, and critical arguments.
Rather than applying simplistic word-spinning, this workflow applies structured linguistic techniques across multiple processing rounds. It represents a powerful integration within the ecosystem of Openclaw Skills, allowing developers and students to automate writing audits without relying on costly external paraphrasing APIs.
Humanization Use Cases
- When you need to verify if an essay will be flagged as AI-generated prior to submission.
- When aiming to lower the ZeroGPT AI-detection score of an academically sound draft.
- When preparing learning journals, reflection logs, or essays for academic peer reviews.
- When optimizing content flow and removing common structural or lexical patterns typical of default LLM output.
How Humanization Works
- Ingest Draft: The skill reads your writing directly from the ongoing conversation or from a workspace file using filesystem operations.
- Automated Verification: It launches a browser session to access ZeroGPT, clears past input, inserts the working draft, triggers detection, and extracts the numeric AI percentage.
- Loop-Based Refinement: If the AI score is equal to or higher than 70%, the skill executes up to five progressive rounds of revisions using custom optimization strategies.
- Dynamic Score Checks: After each optimization round, the updated draft is re-analyzed via ZeroGPT to measure incremental score drops.
- Completion & Exporting: Once the score drops safely below the 70% threshold, the final rewritten draft is generated, along with an audit trail showing the progress across all revision rounds.
Humanization Setup
To deploy and configure the Humanization skill within your workspace, ensure you have the required browser automation environment active.
# Activate and register the humanization skill within your directory
openclaw skills install humanization
# Ensure system-level browser instances (e.g., Playwright or Puppeteer) are operational
playwright install chromium
Configure permissions in your workspace configuration settings to allow filesystem access for reading local draft files and saving the final outputs.
Humanization Data Schema & Taxonomy
The Humanization skill processes raw text drafts in-memory or directly from your specified workspace directories. Metrics are systematically tracked and presented upon task completion.
Iterative Verification Metrics Schema
| Attribute Name | Data Type | Description |
|---|---|---|
Round |
Integer | The incremental step of the rewrite process (0 to 5) |
AI Score |
String | The percentage rating of AI writing extracted from ZeroGPT |
Status |
String | Evaluative feedback indicating if the score meets the requirement |
File Integration
- Input: Markdown (
.md) or Plain Text (.txt) documents loaded via filesystem reads. - Output: The optimized, humanized essay is displayed in the chat interface or optionally exported back to the directory as an output file.
Humanization Advanced Features
- Progressive Multi-Round Optimization: Utilizes layered rewriting tactics starting with sentence-level variation, transitioning through vocabulary naturalization, structural reordering, academic hedging, and voice refinement.
- Linguistic Tone Guardrails: Rejects informal vocabulary and emotional anecdotes, ensuring the document maintains a professional academic register and preserves original citations.
- DOM Injection Fallbacks: Integrates advanced Javascript execution hooks to simulate native input events if web platforms block traditional copy-paste inputs.
- Document Partitioning for Scale: Automatically flags long documents exceeding 3000 words, splitting them into logical segments to process and average across web execution boundaries.
SKILL.md
Loading
Related Openclaw Skills
Discount Hunter
An automated browser-based coupon hunter that finds and tests promo codes directly on checkout pages with complete financial safety.
Huo15 Odoo Integration
A zero-dependency Python toolset facilitating task tracking, timesheet aggregation, CRM administration, and calendar reminders on Odoo through XML-RPC.
Code Mirror
A powerful full-stack code generation and synchronization skill that mirrors APIs, data schemas, validation rules, and client logic between frontend and backend stacks.
Douyin Scraper
A versatile automation capability for Openclaw Skills that extracts trending Douyin videos, hot list metrics, and creative copy using browser replication or scripting.
FeliciaOS Community Audit
A comprehensive Web3 community health auditing and growth intelligence skill powered by the FeliciaOS engine.
FeliciaOS Confession Generator
Generates raw, anonymous, first-person crypto win/loss confession stories with real numbers and zero moralizing.
Featured*
