What happened
Linx Security announced general availability of Linx Agentic Access Control on June 9, 2026 — an inline enforcement layer positioned between AI platforms and enterprise applications that intercepts every Model Context Protocol (MCP) tool call before execution. The product enables tool-level policy controls (granular to individual read/write/admin operations within each MCP server), real-time approve/block decisions, and full audit logging attributable to human, non-human, or autonomous agent identities. The company cites McKinsey's 2026 AI Trust Maturity Survey (65% of orgs cite security as the top barrier to scaling agentic AI) and May 2026 NSA guidance warning of uncontrolled MCP automated actions as validation of the problem space.
Why it matters
MCP traffic currently bypasses downstream application logs in most enterprise environments, leaving a forensic and governance blind spot that is actively being exploited in the wild. Linx's inline MCP gateway approach addresses this at the protocol layer — a tighter control point than agent-side guardrails or post-hoc log analysis. Paired with Linx's existing human and non-human identity governance, it is the most purpose-built MCP enforcement product announced to date.
Applicability
Organisations deploying enterprise AI agents connected to CRM, ERP, database, or API systems via MCP; priority evaluation for regulated sectors (finance, health) where auditability of automated actions is a compliance requirement.