AI & Enterprise
Microsoft releases AI agent control standard for developers
Microsoft has released an open-source standard called the Agent Control Spec to manage what AI agents are allowed or prohibited from doing. The specification helps developers and security and compliance teams define agent behavior rules in policy files. Policies can cover permitted and banned actions, points requiring human approval, and what records to keep for later review. The rules are applied at multiple checkpoints throughout an agent’s workflow and support plugins for major agent frameworks.
