Armory
Source
Browse
Skills

nemo-guardrails

NVIDIA's runtime safety framework for LLM applications. Features jailbreak detection, input/output validation, fact-checking, hallucination detection, PII filtering, toxicity detection. Uses Colang 2.0 DSL for programmable rails. Production-ready, runs on T4 GPU.

Score
Unranked
Evidence
No signals yet
Last commit
Not known
Listed

Install

armory install nemo-guardrails --cli claude

writes the skill into.claude/skills/nemo-guardrails/SKILL.mdListed as compatible

Configuration
# fetches the source and writes it to:
.claude/skills/nemo-guardrails/SKILL.md

Needs the armory CLI · not on npm yet, build it from cli/ in the repository

What it is

NVIDIA's runtime safety framework for LLM applications. Features jailbreak detection, input/output validation, fact-checking, hallucination detection, PII filtering, toxicity detection. Uses Colang 2.0 DSL for programmable rails. Production-ready, runs on T4 GPU.

When to use it

NVIDIA's runtime safety framework for LLM applications. Features jailbreak detection, input/output validation, fact-checking, hallucination detection, PII filtering, toxicity detection. Uses Colang 2.0 DSL for programmable rails. Production-ready, runs on T4 GPU.

How to install / invoke

# Copy the skill into your .claude/skills/ directory
curl -sL https://raw.githubusercontent.com/davila7/claude-code-templates/main/cli-tool/components/skills/ai-research/safety-alignment-nemo-guardrails/SKILL.md -o .claude/skills/safety-alignment-nemo-guardrails/SKILL.md

Notes

Extracted from davila7/claude-code-templates, safety-alignment-nemo-guardrails category.