MS unveils ‘Rampart/Clarity’, an open source tool to strengthen AI agent security

Microsoft (MS) has released two new tools as open source that support the application of safety verification from the early stages of AI agent development. The tools, named ‘Rampart’ and ‘Clarity’, were released last week as part of Microsoft’s strategy to extend agentic AI’s safety engineering to the actual operational level. Ram Shankar Shiva Kumar, founder of MS AI Red Team, said on his official blog, “AI safety is no longer a periodic inspection, but must be established as a continuous engineering system.