AI Model Security Scan Results
Learn about Prisma AIRS AI Model Security scan results, including how to interpret
verdicts via SDK and web interface.
| Where Can I Use This? | What Do I Need? |
|---|
- Prisma AIRS (Model Security)
|
- Prisma AIRS Model Security License
|
Once a scan finishes, you can retrieve the results through either the SDK or the Strata
Cloud Manager web interface. The primary check involves determining whether your scanned
model is permitted or restricted.
Execute the following command to retrieve the scan result using the SDK:
scan_result.eval_outcome
To view the scan results in the Strata Cloud Manager, navigate to and locate your recently scanned model.
A red shield
indicates the model was blocked, while a green
shield
shows the model is allowed.
Select the Scan Request ID to open a detailed flyover with
comprehensive scan information:
This example shows a result of a scan that is blocked due to failure in compliance to the
rules: the use of non-approved file formats and detection of code execution in
unsafe_model.pkl that runs during model loading, indicating a
potential deserialization attack.
To view the complete JSON response that your SDK would receive, select
JSON:
