Microsoft Releases RAMPART and Clarity for AI Agent Security Testing
Add RAMPART to your project and write tests for AI agent safety.
Add RAMPART to your project and write tests for agent safety.
Summary
Microsoft has released two open‑source tools, RAMPART and Clarity, to help developers test the security of AI agents. RAMPART is a Pytest‑native framework that allows developers to write and run safety and security tests for AI agents, covering a wide range of attack scenarios. Clarity complements RAMPART by providing visibility into agent behavior and enabling analysis of decision paths. Both tools are designed to integrate seamlessly with existing CI pipelines and support a variety of agent architectures. The release aims to standardize security testing for AI agents and reduce the risk of unsafe or malicious behavior. Developers can now write tests that assert agent compliance with safety guidelines and verify that the agent does not produce harmful outputs. The tools are available under an open‑source license and can be added to projects that build AI agents.
Key changes
- RAMPART is a Pytest‑native framework for AI agent testing
- Clarity provides visibility into agent behavior
- Both tools are open‑source
- RAMPART covers safety and security test cases
- Clarity supports analysis of decision paths
- Designed for CI pipeline integration
- Supports various agent architectures
- Released by Microsoft