Tenable opens its exposure intelligence to outside AI agents with MCP-based OPEN for AI
Tenable announced Tenable OPEN for AI on October 20, 2026, an expansion of its Open Partner Exchange Network that lets security teams connect the company's exposure intelligence to the AI models, agents and tools they already use, through standardized Model Context Protocol (MCP) connectors and direct APIs.
The program targets a real operational gap: AI security copilots often recommend remediation without knowing whether an asset is actually vulnerable, reachable, or privileged. By connecting the Tenable Hexa AI MCP Server with partner MCP servers, an AI assistant can query live vulnerability and exposure data alongside datasets from other vendors in a single conversation, without custom integration work.
Launch partners include BeyondTrust, Cohesity, Command Zero, Corelight, Infoblox, Recorded Future, Securonix, SentinelOne, Setyl, Snyk and Splunk. Validated launch workflows include cross-platform security triage with SentinelOne, network threat and blast-radius mapping with Corelight, code-to-runtime risk correlation with Snyk, and identity-enriched exposure triage with BeyondTrust.
Each workflow ships in two forms: a documented configuration teams can replicate through the Hexa AI MCP Server with their AI tool of choice, or a ready-to-use skill on the CyberAgents Exchange, Tenable's vendor-agnostic AI agent marketplace. Tenable chief product officer Eric Doerr framed the pitch as freedom from single-vendor stacks: validated AI use cases without forcing reliance on one proprietary platform.
For buyers, the practical read is that exposure management is becoming an MCP data source rather than a closed console. Using the workflows requires the relevant Tenable platform and the partner products involved, and the launch partners span endpoint, network, code, identity and SIEM - so most enterprise stacks can connect at least one integration on day one.
The boundary to keep in mind: these are query-and-triage workflows that put better context in front of analysts and agents. Automated remediation on production systems still belongs behind human approval, and the value of any integration depends on the quality of the exposure data feeding it.