Securing AI Agents with NVIDIA OpenShell and Harness
Harness, Tuesday, November 10th, 2026
Harness describes how its AI and API security tools complement NVIDIA's OpenShell reference architecture for AI agents.
Harness explains how its AI Security offering can complement NVIDIA's OpenShell Reference Architecture, which governs agent runtime access, with optional NVIDIA Sentry on BlueField for independent monitoring.
Harness discovers agents from runtime traffic and builds a context graph linking agents, models, MCP servers, tools and APIs.
AI red teaming simulates attacks on agents, API security testing assesses the APIs behind MCP servers and tools, and AI detection and response monitors agent interactions and applies protection against prompt injection, tool misuse, data exposure and API attacks.
The proposed deployment keeps the self-managed platform and its data lake within the customer environment.