The Agent Skills Directory

[SAFE]: The skill is a comprehensive guide for using the NeMo Guardrails library for AI safety.- [PROMPT_INJECTION]: The content contains phrases associated with prompt injection (e.g., 'Ignore all previous instructions', 'DAN'); however, these are used exclusively as data examples within code blocks to demonstrate how the framework detects and blocks such attacks. They do not represent an attempt to override the agent's behavior.- [EXTERNAL_DOWNLOADS]: The skill references the standard nemoguardrails Python package and provides links to official NVIDIA GitHub repositories and documentation. These are trusted sources for the technology described.- [DYNAMIC_EXECUTION]: The skill demonstrates the use of RailsConfig.from_content to define safety flows using Colang DSL. This is the intended functional use of the documented library and does not involve executing untrusted or hidden code.

nemo-guardrails