Setup & Installation
Or with the ClawHub CLI, for registry-managed skill folders outside a full OpenClaw workspace:
What This Skill Does
Profanity detection library for JavaScript, TypeScript, and Python that handles evasion techniques like leetspeak, Unicode homoglyphs, and obfuscated text. Supports 24 languages with optional ML-powered toxicity scoring via TensorFlow.js. Includes context-aware analysis for domains like medical, gaming, and technical content.
Unlike simple word-list filters, it catches evasion attempts like leetspeak and Unicode lookalikes that users commonly exploit to bypass moderation.
When to use it
- Filtering user comments on a forum before they go live
- Blocking profane usernames during account registration
- Moderating chat messages in a multiplayer game
- Validating form submissions in a web app
- Scanning batch-uploaded content for policy violations
Example Workflow
Here's how your AI assistant might use this skill in practice.
User asks: Check if a chat message contains profanity before sending
- 1Import Filter from glin-profanity
- 2Initialize filter with detectLeetspeak and normalizeUnicode enabled
- 3Call filter.checkProfanity() on the message text
- 4Inspect result.containsProfanity and result.profaneWords
- 5Block or replace the message using result.processedText if flagged
Message is either sent as-is or replaced with a censored version, and the user is warned if violations are found