Zscaler launches WebMCP Security Controls for agentic web

Zscaler has introduced WebMCP Security Controls in its Zero Trust Browser, extending zero-trust principles to AI agents’ automated tool calls made via the emerging WebMCP standard.

The controls provide capability discovery, runtime monitoring, and policy enforcement for WebMCP tools, allowing organisations to inventory, audit, and block agent actions based on tool identity, arguments, and origin.

Available today, the feature aims to secure agentic browsing without sacrificing productivity, with policies that can allow, monitor, or block tool invocations in real time.