The current documentation for Prisma AIRS is missing the most recent update, which includes a new key-value pair that captures prompt detection details
We simply need to update the API schema here to reflect this new functionality within the response payload
{
"action": "block",
"category": "malicious",
"error": false,
"errors": [],
"profile_id": "59e0f670-7901-4ed7-aa38-1507788cb6cf",
"profile_name": "AI-Firewall-High-Security-Profile",
"prompt_detected": {
"agent": true,
"dlp": false,
"injection": true,
"malicious_code": false,
"topic_violation": false,
"toxic_content": true,
"url_cats": false
},
"prompt_detection_details": {
"toxic_content_details": {
"toxic_categories": [
"Indiscriminate Weapons",
"Violent Crimes"
]
}
},
"report_id": "Redd0e3c9-e7b3-4827-a200-c2eb8b618ddb",
"response_detected": {},
"scan_id": "edd0e3c9-e7b3-4827-a200-c2eb8b618ddb",
"session_id": "pan_be2fad45-6907-48bc-b616-d68cd8f8e556",
"source": "AI-Runtime-API",
"timeout": false,
"tool_detected": {},
"tr_id": "pan_be2fad45-6907-48bc-b616-d68cd8f8e556"
}
Documentation link
The current documentation for Prisma AIRS is missing the most recent update, which includes a new key-value pair that captures prompt detection details
Describe the problem
We simply need to update the API schema here to reflect this new functionality within the response payload
Suggested fix
Support for this example structure: