In summary
- Netskope has launched Skylight Agent Action Control, which classifies every AI agent action and applies policy-based controls to block high-risk actions before they execute.
- The launch coincides with Netskope renaming its AI security portfolio from Netskope One AI Security to Netskope Skylight, adding Agent Action Control to the family.
- The capability classifies agent actions into nine intent-based categories and lets security teams set risk-based block, allow or alert policies by agent type, running on traffic the platform already inspects.
Netskope has launched Netskope Skylight Agent Action Control, a capability that classifies every action an AI agent attempts and applies policy-based controls to block high-risk actions before they execute.
The launch coincides with Netskope renaming its AI security portfolio. The suite formerly known as Netskope One AI Security is now branded Netskope Skylight, covering Netskope Skylight AI Command Center, AI Guardrails, AI Gateway, Agentic Broker and AI Red Teaming, with Agent Action Control added to the family.
The new capability classifies every agent action into one of nine intent-based categories prior to execution: access control changes, configuration changes, credential and secret manipulation, data destruction, infrastructure provisioning, potential data exfiltration, potential external communication, remote code execution, and source code changes.
Security teams can then set risk-based policy profiles - block, allow or alert - across low, medium, high or critical risk levels, with profiles attached to specific agent types. A coding assistant and a chat application can operate under different rules.
Every action is logged, and policy alerts can be filtered by cost exposure, source code changes, infrastructure updates or external communication.
According to Netskope, the capability runs on network traffic the platform already inspects, requiring no new console or additional agent deployment.
"AI agents tend to act first and explain later, and most security teams only learn what happened after it is done," said John Martin, chief product officer at Netskope.
"Agent Action Control puts a decision in front of every action an agent takes, so a team can say yes to agentic AI without saying yes to the one action that could cost them a production system."




