Tenable opens CyberAgents Exchange for AI security tooling
Tenable has launched the CyberAgents Exchange, a free open-source registry purpose-built for cybersecurity AI agents, skills, Model Context Protocol (MCP) servers and multi-agent playbooks. The NASDAQ-listed exposure management vendor announced the platform at Black Hat USA 2026 in Las Vegas, positioning it as a vendor-agnostic alternative to general-purpose AI exchanges that security teams have found ill-suited to operational use cases.
SentinelOne and Recorded Future have joined as founding members, each committing to contribute AI components, integration frameworks and playbooks to the community. More than 50 components are already listed under open-source licences, ranging from Tenable's own Navi vulnerability management agent to a Recorded Future skill that maps APT group techniques from the MITRE ATT&CK framework against live Tenable findings to surface actively exploitable attack paths.
The platform
The Exchange provides code-level provenance for every listed component, including authorship, creation date and peer-review status. Tenable says this transparency is designed to address a specific concern in the agentic AI market: that autonomous security tooling deployed without auditability creates new blind spots rather than closing existing ones.
Jamie Zajac, Chief Product Officer at Recorded Future, articulated the dependency clearly: "The promise of AI agents in security depends entirely on whether they can be trusted. Intelligence doesn't just make agents smarter. It makes them traceable, auditable and repeatable."
Alongside the Exchange, Tenable is running a multi-day build event at Black Hat, branded SWARM, sponsored by AWS and supported by Anthropic. Practitioners are competing to build open-source agents, skill files or MCP servers, with winners announced on 6 August. The event is a direct community-building exercise: it generates contributed components while simultaneously training practitioners in a discipline Tenable describes as a career asset.
Market context
The agentic AI wave is arriving in cybersecurity at a moment when security operations teams are under sustained pressure on headcount and alert volume. The market for AI-native security tooling is crowded: hyperscalers embed security copilots into their platforms, while pure-play vendors including CrowdStrike, Palo Alto Networks and a cohort of well-funded startups are building proprietary agent frameworks. The problem Tenable is addressing, that bespoke agent development is being duplicated across hundreds of security teams independently, is genuine and widely acknowledged in the industry.
Open-source registries have a mixed track record as competitive moats. The model works when network effects accumulate faster than commercial alternatives can replicate community-contributed tooling; it stalls when governance becomes diffuse or contribution quality is uneven. Tenable's decision to provide code-level provenance and peer-review status is a direct response to that second risk. The company's existing customer base of over 40,000 organisations gives it an unusually large pool from which to seed initial contribution.
Regulatory pressure is also shaping the agentic AI conversation in security. The EU AI Act classifies certain autonomous decision-making systems as high-risk, and its conformity-assessment requirements will apply to AI tools used in critical infrastructure protection. Auditability and traceability features, which Tenable is positioning as differentiators, are likely to become baseline compliance requirements for enterprise deployments in regulated sectors over the next two to three years. NIS2, which entered enforcement in EU member states in late 2024, similarly raises the bar on supply-chain transparency for security tooling. The CyberAgents Exchange's code-level provenance model maps directly onto those expectations, giving Tenable a compliance-aligned narrative as the regulatory environment tightens.