A sophisticated profanity detection and content moderation library that thwarts evasion attempts like leetspeak and Unicode homoglyphs.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install glin-profanity
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 glin-profanity 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 glin-profanity?
glin-profanity is a comprehensive content moderation library designed to safeguard digital platforms from toxic user-generated content. As a versatile addition to the Openclaw Skills ecosystem, it provides developers with the tools to detect not only standard profanity but also complex evasion techniques. It effectively identifies leetspeak patterns (like f4ck or sh1t) and Unicode homoglyphs (such as using Cyrillic characters that look identical to Latin ones) to ensure that moderation bypasses are minimized.
Supporting 24 languages and offering optional machine learning integration, glin-profanity is built for high-performance environments. It allows for deep customization, including domain-specific whitelisting for industries such as medicine or gaming. This ensures that technical or medical terminology is not incorrectly flagged while still maintaining a safe environment for users.
glin-profanity Use Cases
- Real-time chat and comment moderation to maintain community standards.
- Validating user-generated profile names and bios during registration.
- Sanitizing content in React-based applications using specialized hooks.
- Implementing automated moderation bots for social platforms using Openclaw Skills.
- Domain-specific filtering for gaming, medical, and educational applications.
How glin-profanity Works
- The text input is received and processed through a Unicode normalization layer to resolve homoglyph-based evasion.
- The engine applies leetspeak detection logic to identify obfuscated words based on the selected sensitivity level.
- The normalized text is compared against extensive dictionaries spanning 24 different languages.
- If enabled, the content is analyzed by an ML model to score it for toxicity and category-specific insults.
- Custom whitelists and domain-specific rules are applied to prevent false positives in technical or medical contexts.
- The system returns a structured result object containing the profanity flag, detected words, and a censored version of the input.
glin-profanity Setup
Integrating this library into your project is straightforward. Use the following commands to install the necessary packages for your environment:
# For JavaScript or TypeScript environments
npm install glin-profanity
# For Python environments
pip install glin-profanity
Once installed, you can initialize the filter within your Openclaw Skills workflow by configuring your preferred languages and detection levels.
glin-profanity Data Schema & Taxonomy
The library produces detailed analysis objects that can be easily integrated into larger data pipelines. The schema for a standard check result includes:
| Property | Type | Description |
|---|---|---|
| containsProfanity | boolean | True if profane content is detected. |
| profaneWords | string[] | Array of the specific words identified as profane. |
| processedText | string | The input text with profane segments replaced by a mask (e.g., ***). |
| confidence | number | The probability score provided by the ML toxicity model. |
Configuration objects allow you to define languages, leetspeakLevel, customWords, and ignoreWords to tailor the behavior to your specific needs.
glin-profanity Advanced Features
- Multi-language support for 24 global languages including Arabic, Chinese, Russian, and Hindi.
- Context-aware analysis that adjusts sensitivity based on the domain (e.g., medical vs. gaming).
- Optional ML-powered toxicity detection utilizing TensorFlow.js for deep intent analysis.
- High-performance batch processing capabilities for analyzing multiple strings simultaneously.
- Built-in LRU caching to minimize redundant processing in high-traffic Openclaw Skills deployments.
SKILL.md
Loading
Related Openclaw Skills
Glin Profanity MCP Server
A Model Context Protocol server providing AI assistants with robust tools for profanity detection, text censoring, and content safety analysis.
Echo Repeater
A minimal utility skill designed to echo user input back for testing skill routing and message integrity.
RALSTP Consultant
An AI-driven strategic planning consultant that applies Recursive Agents and Landmarks Strategic-Tactical Planning (RALSTP) to decompose complex workflows and identify agent dependencies.
Quality Gateways (Generic Web + API Applications)
A language-agnostic framework for enforcing six universal quality gateways across backend APIs and web frontends using LLM-friendly operational standards.
OpenClaw Profanity Plugin
A sophisticated content moderation plugin for OpenClaw and Moltbot AI agents featuring leetspeak detection and multi-language support.
clawplot
ClawPlot is a specialized art service that allows AI agents to order physical, archival ink pen-plotted artwork from SVG designs.
Featured*
