Pangea AI Guard integration

Warning: The Pangea AI Guard integration is deprecated and will be removed in a future release. Users should migrate to the CrowdStrike AIDR integration.

The Pangea guardrail uses configurable detection policies (called recipes) from the AI Guard service to identify and mitigate risks in AI application traffic, including:

• Prompt injection attacks (with over 99% efficacy)
• 50+ types of PII and sensitive content, with support for custom patterns
• Toxicity, violence, self-harm, and other unwanted content
• Malicious links, IPs, and domains
• 100 spoken languages, with allowlist and denylist controls

All detections are logged in an audit trail for analysis, attribution, and incident response. You can also configure webhooks to trigger alerts for specific detection types.

The following environment variable is required to use the Pangea AI Guard integration:

• PANGEA_API_TOKEN: Pangea API token with access to the AI Guard service.

You can also optionally set:

• PANGEA_BASE_URL_TEMPLATE: Template for constructing the base URL for API requests. The {SERVICE_NAME} placeholder will be replaced with the service name slug. Defaults to https://ai-guard.aws.us.pangea.cloud for Pangea’s hosted (SaaS) deployment.

Setup

Colang v1:

1
# config.yml
2
3
rails:
4
  config:
5
    pangea:
6
      input:
7
        recipe: pangea_prompt_guard
8
      output:
9
        recipe: pangea_llm_response_guard
10
11
  input:
12
    flows:
13
      - pangea ai guard input
14
15
  output:
16
    flows:
17
      - pangea ai guard output

Colang v2:

1
# config.yml
2
3
colang_version: "2.x"
4
5
rails:
6
  config:
7
    pangea:
8
      input:
9
        recipe: pangea_prompt_guard
10
      output:
11
        recipe: pangea_llm_response_guard

# rails.co
import guardrails
import nemoguardrails.library.pangea
flow input rails $input_text
    pangea ai guard input
flow output rails $output_text
    pangea ai guard output

Next steps

• Explore example configurations for integrating Pangea AI Guard with your preferred Colang version:
  • Pangea AI Guard for NeMo Guardrails v1
  • Pangea AI Guard for NeMo Guardrails v2
  • Pangea AI Guard without LLM (guardrails only) – Use this setup to evaluate AI Guard’s detection and response capabilities independently.
• Adjust your detection policies to fit your application’s risk profile. See the AI Guard Recipes documentation for configuration details.
• Enable AI Guard webhooks to receive real-time alerts for detections in your NeMo Guardrails-powered application.
• Monitor and analyze detection activity in the AI Guard Activity Log for auditing and attribution.
• Learn more about AI Guard Deployment Options to understand how and where AI Guard can run to protect your AI applications.