OpenPrompts
← Back to catalog
NVIDIAGuardrailsSafety & Moderation

Prompt Security (NeMo Guardrail)

INPUT RAILS flow protect prompt """Check if the prompt is valid according to Prompt Security.""" $result = await ProtectTextAction(user_prompt=$user_m

# INPUT RAILS

flow protect prompt
  """Check if the prompt is valid according to Prompt Security."""
  $result = await ProtectTextAction(user_prompt=$user_message)
  if $result["is_blocked"]
    if $system.config.enable_rails_exceptions
      send PromptSecurityRailException(message="Prompt not allowed. The prompt was blocked by the 'protect prompt' flow.")
    else
      bot inform answer unknown
    abort
  else if $result["is_modified"]
    $user_message = $result["modified_text"]


# OUTPUT RAILS

flow protect response
  """Check if the response is valid according to Prompt Security."""
  $result = await ProtectTextAction(bot_response=$bot_message)
  if $result["is_blocked"]
    if $system.config.enable_rails_exceptions
      send PromptSecurityRailException(message="Response not allowed. The response was blocked by the 'protect response' flow.")
    else
      bot inform answer unknown
    abort
  else if $result["is_modified"]
    $bot_message = $result["modified_text"]
Automated safety scan: no suspicious patterns found.

Heuristic text scan aligned to the OWASP Agentic Skills Top 10. How we scan

Provider
NVIDIA
Origin
Official
Type
Guardrails
License
Apache-2.0
Language
English
Added
2025-10-06
#guardrail#nemo#rails#colang