Is there an API to evaluate and monitor prompt injection protection for Copilot Studio agents?

I want to assess whether agents are protected against prompt injection attacks.

I want to perform security drills using controlled test prompts and retrieve:

Agent name and ID.

Test scenario.

Prompt injection detection result.

Attack blocked or successful.

Guardrail triggered.

Test execution timestamp.

Security test pass/fail status.

Risk score, if available.

Does Microsoft provide an official API, security evaluation framework, or testing capability for automating prompt injection security drills?

Can these test results be retrieved through Copilot Studio, Azure AI Foundry, Microsoft Defender, or Application Insights?

Please share the recommended implementation approach and relevant APIs.