nemo-guardrails

Implement NeMo Guardrails with Colang 2.0 for LLM safety and security.

Updated Jan 12, 2026
One-click install
npx skills add https://github.com/MesferAli/XCircle --skill nemo-guardrails-mesferali
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: nemo-guardrails
Source: https://github.com/MesferAli/XCircle/tree/main/.claude/skills/nemo-guardrails
Command: npx skills add https://github.com/MesferAli/XCircle --skill nemo-guardrails-mesferali

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This Skill provides a robust framework for adding programmable safety and security layers to Large Language Model (LLM) applications, preventing harmful outputs and ensuring reliable interactions.

Core Features & Use Cases

  • Runtime Safety: Implements guardrails to detect and prevent jailbreaks, toxic content, and PII leakage.
  • Input/Output Validation: Ensures both user inputs and LLM outputs adhere to defined policies.
  • Fact-Checking & Hallucination Detection: Integrates mechanisms to verify factual accuracy and mitigate hallucinations.
  • Use Case: Protect a customer-facing chatbot from generating inappropriate responses, leaking sensitive data, or providing factually incorrect information by implementing these guardrails.

Quick Start

Install the library and use the provided Python code to define and apply safety configurations to your LLM interactions.

Frequently Asked Questions about nemo-guardrails

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

FAQPage Schema
How do I prevent prompt injection and jailbreaks in my LLM application at runtime?

To prevent prompt injection and jailbreaks at runtime, you can implement programmable guardrails using Colang 2.0 DSL to validate user inputs and LLM outputs against defined security policies. This framework detects malicious prompts and blocks unsafe interactions dynamically.

What is the best way to filter PII and detect hallucinations in LLM outputs?

The best way to filter PII and detect hallucinations is by integrating runtime guardrails with tools like Presidio and LlamaGuard. This validates factual accuracy, mitigates hallucinations, and prevents sensitive data leakage from model outputs.

Can I use Colang 2.0 to define custom safety policies for customer-facing chatbots?

Yes, you can use Colang 2.0 DSL to define custom programmable rails for customer-facing chatbots. This allows you to enforce specific input and output validation rules, preventing toxic content and ensuring reliable interactions.

How do I implement input and output validation to stop toxic content generation?

You implement input and output validation to stop toxic content by applying runtime safety guardrails that inspect both user prompts and LLM responses. This ensures interactions adhere strictly to your defined toxicity and security policies.

Does NeMo Guardrails work with Presidio and LlamaGuard for LLM security?

Yes, NeMo Guardrails integrates directly with Presidio and LlamaGuard to enhance LLM security. This combination provides robust PII filtering, toxicity detection, and fact-checking capabilities to secure runtime model applications.