What happened
Tenable announced (Sept 9, at OpenAI's Intelligence at Work: Cyber Summit) the CyberAgents Exchange AI Inspector, a new review process combining OpenAI GPT cyber models with Tenable One AI Exposure skills-inspection and human researcher review to vet AI agents, skills, MCP servers, and multi-agent playbooks on the CyberAgents Exchange (100+ community components) before enterprise deployment, checking for prompt injection, hidden Unicode instructions, supply-chain risk, and undisclosed egress.
Why it matters
Distinct from the previously-covered Tenable One Adversary View (Claude Mythos 5 integration), this establishes a pre-deployment vetting/certification gate specifically for third-party agentic components — addressing a supply-chain risk vector (malicious skills/MCP servers) increasingly exploited per recent CVE disclosures.
Applicability
Security teams sourcing community-built agents/skills/MCP servers should require Exchange Inspector review status before internal deployment.