Palo Alto Networks · Capability

Palo Alto Networks AI Security

Unified AI security capability for scanning AI model inputs/outputs for threats and red-teaming AI applications for vulnerabilities across Prisma AIRS and AI Red Teaming APIs.

Run with Naftiko Palo Alto NetworksAI SecurityAI Red TeamingLLM SecurityPrompt Injection

What You Can Do

POST
Submit sync scan — Submit a synchronous scan of AI model inputs/outputs for threats.
/v1/ai-scans/sync
POST
Submit async scan — Submit an asynchronous scan of AI model inputs/outputs for threats.
/v1/ai-scans/async
GET
Get async scan results — Get the results of a previously submitted asynchronous scan.
/v1/ai-scans/async/{scan_id}/results
GET
List ai profiles — List all AI security profiles with pagination.
/v1/ai-profiles
GET
Get ai profile — Get a specific AI security profile by name.
/v1/ai-profiles/{profile_name}
POST
Create red team target — Create a new red team scan target.
/v1/red-team-targets
GET
List red team targets — List all red team scan targets.
/v1/red-team-targets
GET
Get red team target — Get a specific red team scan target by ID.
/v1/red-team-targets/{target_id}
DELETE
Delete red team target — Delete a specific red team scan target by ID.
/v1/red-team-targets/{target_id}
POST
Start red team scan — Start a new red team vulnerability scan against a target.
/v1/red-team-scans
GET
Get red team scan status — Get the status of a red team vulnerability scan.
/v1/red-team-scans/{scan_id}
GET
Get red team scan results — Get results of a red team vulnerability scan with optional filters.
/v1/red-team-scans/{scan_id}/results
GET
List attack categories — List all available attack categories for red team scans.
/v1/attack-categories

MCP Tools

submit-sync-scan

Submit a synchronous AI security scan of model inputs/outputs for threats like prompt injection, data leakage, and malicious content.

submit-async-scan

Submit an asynchronous AI security scan of model inputs/outputs for threats.

get-async-scan-results

Get results of a previously submitted asynchronous AI security scan.

read-only idempotent
list-ai-profiles

List all AI security profiles with pagination.

read-only idempotent
get-ai-profile

Get a specific AI security profile by name.

read-only idempotent
create-red-team-target

Create a new red team scan target for AI application vulnerability testing.

list-red-team-targets

List all red team scan targets with optional type filter.

read-only idempotent
get-red-team-target

Get a specific red team scan target by ID.

read-only idempotent
delete-red-team-target

Delete a specific red team scan target by ID.

idempotent
start-red-team-scan

Start a new red team vulnerability scan against a target AI application.

get-red-team-scan-status

Get the status of a red team vulnerability scan.

read-only idempotent
get-red-team-scan-results

Get results of a red team vulnerability scan with optional category and severity filters.

read-only idempotent
list-attack-categories

List all available attack categories for red team vulnerability scans.

read-only idempotent

Who This Is For

👤
AI Security Engineer
Secures AI applications with runtime scanning and vulnerability assessment.
👤
Red Team Operator
Conducts automated adversarial testing against AI systems and LLM applications.

APIs Used

prisma-airs prisma-airs-red-teaming