Prisma Airs Ai Red Teaming Api Vulnerability Finding Structure

{
  "$schema": "https://json-structure.org/meta/core/v0/#",
  "$id": "https://raw.githubusercontent.com/api-evangelist/palo-alto-networks/refs/heads/main/json-structure/prisma-airs-ai-red-teaming-api-vulnerability-finding-structure.json",
  "name": "VulnerabilityFinding",
  "description": "VulnerabilityFinding schema from Palo Alto Networks Prisma AIRS AI Red Teaming API",
  "type": "object",
  "properties": {
    "finding_id": {
      "type": "string",
      "description": "Unique identifier of the finding."
    },
    "category_id": {
      "type": "string",
      "description": "Attack category identifier."
    },
    "category_name": {
      "type": "string",
      "description": "Human-readable attack category name."
    },
    "severity": {
      "type": "string",
      "description": "Severity level of the vulnerability.",
      "enum": [
        "informational",
        "low",
        "medium",
        "high",
        "critical"
      ]
    },
    "title": {
      "type": "string",
      "description": "Brief title describing the vulnerability."
    },
    "description": {
      "type": "string",
      "description": "Detailed description of the vulnerability and its impact."
    },
    "attack_prompt": {
      "type": "string",
      "description": "The adversarial prompt that revealed the vulnerability."
    },
    "model_response": {
      "type": "string",
      "description": "The AI model's response to the attack prompt, demonstrating the vulnerability. May be truncated."
    },
    "remediation": {
      "type": "string",
      "description": "Recommended steps to mitigate the vulnerability."
    }
  }
}