“guardrails and AI safety”

Meta's 7-8B specialized moderation model for LLM input/output filtering.

This skill lists commands that are generally safe, and commands that are very NOT safe and should be avoided.