nemo-guardrails

Configure programmable runtime safety rails for LLM applications using Colang 2.0 DSL.

Updated Mar 18, 2026
One-click install
npx skills add https://github.com/tadod12/fraud-detection-research --skill nemo-guardrails-tadod12
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: nemo-guardrails
Source: https://github.com/tadod12/fraud-detection-research/tree/main/.agent/skills/07-safety-alignment/nemo-guardrails
Command: npx skills add https://github.com/tadod12/fraud-detection-research --skill nemo-guardrails-tadod12

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

NeMo Guardrails provides programmable runtime safety rails for LLM applications, helping teams detect and prevent unsafe prompts, injections, and outputs; it includes features for jailbreak detection, input/output validation, fact-checking, hallucination detection, PII filtering, and toxicity detection.

Core Features & Use Cases

  • Jailbreak detection
  • Input/output validation
  • Fact-checking (with retrieval)
  • Hallucination detection
  • PII filtering
  • Toxicity detection
  • Use cases: production LLM deployments needing strict safety rails

Quick Start

Configure NeMo Guardrails in your LLM application to enforce runtime safety rails and guard against unsafe prompts.

Frequently Asked Questions about nemo-guardrails

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

FAQPage Schema
How do I add runtime safety rails to my LLM application to prevent jailbreaks?

Runtime safety rails prevent jailbreaks in LLM applications by validating inputs and outputs programmatically. NeMo Guardrails enables this by detecting unsafe prompts and injections during inference to block malicious interactions.

What is the best way to implement PII filtering and toxicity detection for production LLMs?

PII filtering and toxicity detection for production LLMs are best implemented through programmable runtime guardrails. NeMo Guardrails applies these safety checks across deployment pipelines to intercept sensitive data and toxic content.

How does fact-checking with retrieval work to mitigate LLM hallucination?

Fact-checking with retrieval mitigates LLM hallucination by validating model outputs against retrieved external knowledge. NeMo Guardrails integrates this mechanism to detect ungrounded responses and prevent factual inaccuracies.

Can I configure custom input and output validation rules using Colang 2.0 DSL?

You can configure custom input and output validation rules using Colang 2.0 DSL. NeMo Guardrails supports this domain-specific language to define programmable safety rails tailored to specific LLM interaction requirements.

Does NeMo Guardrails support production-ready GPU deployment for LLM inference?

NeMo Guardrails supports production-ready GPU deployment for LLM inference. It applies runtime safety rails across deployment pipelines, ensuring low-latency input validation and jailbreak detection in high-throughput environments.