skills.rest
Clay-HHK
Clay-HHKCOMMUNITY
Explore all skills by this author →
GitHub

llamaguard

Filter and classify user and model content across six safety categories.

2|Updated Apr 12, 2026
Search Tags:#deployment#safety#moderation#guardrails#ai-safety#llama#content-filtering
One-click install
npx skills add https://github.com/Clay-HHK/claude-config --skill llamaguard-clay-hhk
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: llamaguard
Source: https://github.com/Clay-HHK/claude-config/tree/main/skills/AI-research-SKILLs/07-safety-alignment/llamaguard
Command: npx skills add https://github.com/Clay-HHK/claude-config --skill llamaguard-clay-hhk

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This skill provides automated, multi-category moderation for LLM conversations, reducing unsafe prompts and outputs across six safety categories (violence/hate, sexual content, weapons, substances, self-harm, criminal planning) with high accuracy and easy integration with vLLM, HuggingFace, Sagemaker, and NeMo Guardrails.

Core Features & Use Cases

  • Input and output moderation to block or flag unsafe content in user prompts and model responses.
  • Deployable with vLLM, HuggingFace, Sagemaker, and NeMo Guardrails for production environments.
  • Suitable for customer support bots, enterprise copilots, and educational assistants that require strict content safety controls.

Quick Start

Install dependencies, login to HuggingFace, load the LlamaGuard-7b model, and begin moderating conversations.

Frequently Asked Questions about llamaguard

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I moderate user prompts and model responses in LLM deployments?▼

Content moderation for LLM deployments involves filtering and classifying user prompts and model responses to block unsafe content. This skill automates multi-category moderation across six safety areas including violence, sexual content, and self-harm.

Can I integrate LlamaGuard content filtering with vLLM and HuggingFace?▼

Yes, content filtering integrates directly with vLLM, HuggingFace, Sagemaker, and NeMo Guardrails for production environments. You load the LlamaGuard-7b model to establish moderation workflows and API endpoints across these supported platforms.

What safety categories does real-time content moderation cover for LLMs?▼

Real-time content moderation for LLMs covers six safety categories: violence and hate, sexual content, weapons, substances, self-harm, and criminal planning. The system automatically classifies prompts and responses to block or flag content in these areas.

How do I set up guardrails for an enterprise copilot or customer support bot?▼

Setting up guardrails for enterprise copilots or customer support bots involves loading the LlamaGuard-7b model and applying input and output moderation. This blocks or flags unsafe content in user prompts and model responses for strict safety controls.

Does NeMo Guardrails support automated multi-category moderation workflows?▼

Yes, NeMo Guardrails supports automated multi-category moderation workflows. You load the LlamaGuard-7b model to establish end-to-end moderation API endpoints and deployment configurations that filter conversations across six safety categories.

Related Skills

Explore More Software Engineering Skills→
tadod12tadod12

llamaguard

Classify and filter unsafe user and model content with LlamaGuard-7b.

Community
Advanced
ricablericable

llamaguard

Classify LLM inputs and outputs into six safety categories.

Community
Advanced
arsityarsity

llamaguard

Classify user-generated content into six safety categories for LLM moderation.

Community
Intermediate
ovachieverovachiever
52

llamaguard

Classify user input and system outputs across six safety categories.

Community
Advanced