SECURITY ATLAS / EVALUATION BRIEF Zscaler AI Gateway Research reviewed: 2026-10-18 https://atlasofsecurity.com/landscape/vendors/zscaler-ai-gateway/ THE ATLAS TAKE Zscaler AI Gateway is worth evaluating when the requirement is governing agent communications across supported MCP and A2A paths. The current product page describes per-request authorization, an agent registry and context-aware policy, which is a different control purpose from a simple model API proxy. Atlas would make trusted identity binding and bypass resistance central to the trial. A gateway can govern what traverses it; agents retaining direct access to the same destination still need independent constraints. Broader discovery, red teaming and Agentic SOC functions should not be folded into this profile automatically. SCOPE Evaluate AI Gateway independently from Zscaler endpoint discovery, red teaming and Agentic SOC products. Verify protocol support and traffic routing. WORTH A LOOK WHEN Teams ready to route supported agent-to-tool or agent-to-agent communication through an explicit gateway. THINK TWICE WHEN The architecture cannot prevent privileged agents from reaching protected destinations by an alternate route. IMPLEMENTATION 1. Inventory every destination route Record agent credentials, protocol transports and direct alternatives to the intended gateway path. 2. Prove identity binding Use two test agents with different owners and permissions; verify that client-editable fields cannot select another identity. 3. Operate policy safely Define review, exception expiry, failure behavior and rollback for a mistakenly denied business workflow. OPERATING REALITY Platform owners must maintain routing, registry ownership and policy as agents and tools change. Security reviewers need to know which context is trusted and which comes from agent-controlled content. Operations should track missing identity, denied calls and gateway failure separately, with an owned procedure for legitimate access recovery. COMMERCIAL CONSIDERATIONS Request an AI Gateway-specific quote with supported protocols, deployment, registry and authorization scope. Itemize other Zscaler AI Security functions and routing infrastructure separately. - Gateway deployment and covered agent/interaction scope - Required identity, discovery and routing integrations - Separate testing modules and policy operations EVALUATION PLAN / PROPOSED TESTS, NOT COMPLETED RESULTS 1. Permit one test tool and deny another Acceptance evidence: Preserve bound identity, policy and target-side invocation evidence. Owner: __________ Pass criterion: __________ Observed result: __________ Evidence location: __________ 2. Attempt the same destination outside the gateway Acceptance evidence: Document the independent control that removes or limits the bypass. Owner: __________ Pass criterion: __________ Observed result: __________ Evidence location: __________ 3. Change an agent permission under review Acceptance evidence: Verify effective timing, audit and reversal for subsequent calls. Owner: __________ Pass criterion: __________ Observed result: __________ Evidence location: __________ QUESTIONS FOR THE VENDOR - Which MCP and A2A transports are supported? - How is agent identity bound to each request? - Can agents reach the same destination outside the gateway? EVIDENCE GAPS Public evidence limitation: Two accessible public product pages establish the gateway purpose and adjacent platform scope. A versioned technical deployment and protocol-support reference was not available in the reviewed sources. - Atlas has not independently tested authorization correctness or gateway performance. - Obtain current transport support, failure semantics and deployment/data-handling documentation. SOURCES Zscaler: AI Gateway control model and scope (Product information) https://www.zscaler.com/products-and-solutions/ai-gateway Zscaler AI Security portfolio boundaries (Product information) https://www.zscaler.com/products-and-solutions/ai-security Use synthetic fixtures in an authorized environment. Agree rollback and acceptance criteria before enforcement. Documentation research and Atlas analysis. Independent hands-on effectiveness, effort, and complete price remain unassessed.