Straiker arrived at RSAC 2026 with two products aimed squarely at the growing security gap in enterprise AI deployments: Discover AI, a new agent inventory and risk detection tool, and an expanded version of Defend AI built to handle the specific behaviors of coding agents, productivity agents, and custom-built agent platforms.

The premise behind both products is direct: “Agents are operating across enterprise systems with broad access, growing autonomy, and zero security oversight.” Straiker is building for that reality.

Discover AI focuses on finding what most organizations cannot currently see: the AI agents already running across their environments. The product does agent discovery and inventory across coding tools and agent frameworks, runs MCP vulnerability detection against a database of 12,000-plus MCP vulnerabilities, identifies security misconfigurations and hygiene issues, and classifies prompts to flag risky agent behavior before it escalates.

Defend AI, which launched earlier, has been expanded with new capabilities and is now trained on millions of real-world agent traces. Straiker reports sub-300ms latency and 98%-plus detection accuracy. New integrations allow deployment in monitoring mode via API with Amazon Bedrock AgentCore, Azure AI Foundry, and Microsoft Copilot Studio, and an inline gateway mode supports active blocking when monitoring is not enough.

The MCP and tool-chain risk detection built into Defend AI addresses one of the faster-moving attack surfaces in enterprise security right now. As organizations wire up AI agents to production systems through tool integrations and MCP servers, the risk profile of those connections has outpaced most security team awareness.

Straiker’s approach, combining discovery coverage with a high-accuracy enforcement layer, positions it as an early infrastructure play in what is becoming a crowded but still largely unsolved market.