Microsoft today announced their newest open-source (MIT-licensed) software project.. the Agent Governance Toolkit. Microsoft is trying their hand at coming up with runtime security governance for autonomous AI agents.
Microsoft proclaims their new open-source project is the first toolkit that addresses all ten agentic AI risks identified last year by the OWASP. These risks include goal hijacking, tool musiuse, identity abuse, supply chain risks, code execution, memory poisoning, insecure communications, cascading failures, human-agent trust exploitation, and rogue agents. This new Agent Governance Toolkit is available for Python, Rust, TypeScript, Go, and .NET.
Microsoft’s Agent Governance Toolkit provides Agent OS as a policy engine that intercepts every agent’s action before execution, Agent Mesh for securing agent-to-agent communication, Agent Runtime for dynamic execution rings, Agent SRE for various safeguards, Agent Compliance for automated governance verification with compliance grading, Agent Marketplace for plug-in life-cycle management, and Agent Lightning for reinforcement learning training governance.
Microsoft says this new toolkit is “open source by design” under the MIT license.
For those trusting Microsoft for handling AI agent security, you can learn more about the Agent Governance Toolkit via the Microsoft Open-Source Blog. The code is hosted on GitHub.
