Skip to main content

AgentSuite-Blue FAQ

What is AgentSuite?​

AgentSuite is the first AI-native, end-to-end security and compliance platform for agentic AI, providing automated red-teaming, real-time guardrails, MCP scanning, agent observability, and centralized governance across the agent lifecycle. AgentSuite also includes Agent ForgingGround with Built-In Red-Teaming Agents to continuously evaluate and stress-test AI agents across realistic enterprise environments.

What makes AgentSuite different?​

Traditional security tools leave organizations blind to how agents actually operate end-to-end. AgentSuite provides the industry's first comprehensive security platform purpose-built for autonomous AI, enabling enterprises to test, monitor, and control agents throughout their entire lifecycle.

What are the core features of AgentSuite?​

  • Agent ForgingGround is an enterprise testing ground that simulates complex single-agent and multi-agent workflows across 50+ production-grade environments such as Databricks, Gmail, Google Docs, PayPal, ServiceNow, and Atlassian. Within those environments, Built-In Red-Teaming Agents perform risk assessments and simulate multiple major AI attacks for agents and multi-agent systems using 1,000+ proprietary red-teaming algorithms. These attacks target injection points such as prompt injection, tool injection, environment manipulation, and skill injection to surface misconfigurations, vulnerabilities, and zero days.

  • MCPGuard performs continuous security analysis of MCP servers, analyzing tool descriptions to detect prompt injections and scanning source code for vulnerabilities such as data leakage and arbitrary code execution risks.

  • ActionGuard provides real-time guardrails for entire agent trajectories, including input queries, output responses, and tool calls, in a stateful way, evaluating every tool call before execution and alerting your team about unsafe, insecure, or out-of-policy actions before they happen.

  • Observability & Access Control delivers full observability into agent structure, configurations, and real-time behavior with visibility into conversations, actions, tool calls, and token traces, plus granular role-based access control.

  • Unified Agent Gateway is positioned between agents and all MCP tools, automatically scanning every MCP with MCPGuard and runtime actions via ActionGuard, with full agent observability and access control, creating a single enforcement point via one-click integration.

How does AgentSuite protect agents at runtime?​

ActionGuard provides real-time, stateful guardrails for every agent action. It evaluates every tool call before execution and alerts teams about unsafe or out-of-policy behavior before it happens.

Does ActionGuard impact agent performance?​

ActionGuard operates with super-low latency, offering real-time protection without impacting agent performance.

How does AgentSuite ensure consistent enforcement across agents and tools?​

The Unified Agent Gateway creates a single enforcement point between agents and all MCP tools. It automatically scans every MCP with MCPGuard and evaluates all runtime tool calls via ActionGuard.

What observability does AgentSuite provide into agent behavior and decisions?​

AgentSuite provides full observability into agent structure and real-time behavior, including visibility into conversations, actions, tool calls, token traces, and visual representations of agent decision trajectories.

What access control and audit capabilities are available for agents?​

AgentSuite includes granular role based access control, control over tool and data resource access by user group, and a centralized audit trail of access control activity.