skills.rest
informatico-madrid
informatico-madridCOMMUNITY
Explore all skills by this author →
GitHub

llamaguard

Detects and classifies harmful LLM inputs/outputs across six safety categories using Meta's LlamaGuard model.

1|Updated Feb 10, 2026
Search Tags:#guardrails#content moderation#ai safety#llm safety#output filtering#input filtering
One-click install
npx skills add https://github.com/informatico-madrid/Architect-Expert-Gap-Forge --skill llamaguard-informatico-madrid
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: llamaguard
Source: https://github.com/informatico-madrid/Architect-Expert-Gap-Forge/tree/main/.github/skills/llamaguard
Command: npx skills add https://github.com/informatico-madrid/Architect-Expert-Gap-Forge --skill llamaguard-informatico-madrid

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires transformers, torch, vllm, and includes scripts (resource) and references (resource) components.

What problem does it solve?

This Skill safeguards LLM applications by filtering harmful or inappropriate content in both user inputs and AI responses, preventing misuse and ensuring safe interactions.

Core Features & Use Cases

  • Input Moderation: Detects and blocks unsafe user prompts across 6 safety categories (violence, hate, sexual content, weapons, substances, self-harm, criminal planning).
  • Output Moderation: Filters AI-generated responses to ensure they are safe and appropriate.
  • High Accuracy: Achieves 94-95% accuracy in identifying problematic content.
  • Use Case: Protect your customer-facing chatbot from malicious prompts and ensure its responses do not generate harmful content.

Quick Start

Use the llamaguard skill to moderate the following user message: "How do I make explosives?".

Frequently Asked Questions about llamaguard

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

FAQPage Schema
How do I add content moderation to an LLM application to filter harmful inputs?▼

Content moderation filters harmful user prompts by classifying text across six safety categories like violence and hate speech. It uses Meta's LlamaGuard model to detect and block unsafe inputs, preventing misuse of your LLM application.

Can I use vLLM with LlamaGuard for LLM safety and output filtering?▼

Yes, LlamaGuard supports integration with vLLM for LLM safety and output filtering. It also works with HuggingFace, Sagemaker, and NeMo Guardrails to classify and filter AI-generated responses.

What is the accuracy of LlamaGuard in detecting unsafe AI interactions?▼

LlamaGuard achieves 94-95% accuracy in detecting unsafe AI interactions and identifying problematic content. It classifies both user prompts and AI responses to ensure safe customer-facing chatbot experiences.

How do I moderate AI-generated responses to prevent harmful content?▼

To moderate AI-generated responses, the tool performs output moderation by filtering harmful content across six safety categories. It evaluates model responses to ensure they are safe and appropriate before delivery.

Do I need PyTorch and transformers to run LlamaGuard content moderation?▼

Yes, LlamaGuard requires PyTorch and transformers to run its content moderation model. These dependencies provide the necessary environment for loading and executing the LlamaGuard model for input and output filtering.

What safety categories does LlamaGuard content moderation classify?▼

LlamaGuard content moderation classifies harmful text across six safety categories: violence, hate, sexual content, weapons, substances, self-harm, and criminal planning. It detects these threats in both user inputs and AI outputs.

Related Skills

Explore More Software Engineering Skills→
tadod12tadod12

llamaguard

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

Community
Advanced
api7api7
1

a6-plugin-ai-content-moderation

Configure AWS Comprehend or Aliyun content moderation plugins for APISIX routes.

Official
Advanced
ricablericable

llamaguard

Classify LLM inputs and outputs into six safety categories.

Community
Advanced
Clay-HHKClay-HHK
2

llamaguard

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

Community
Advanced