Tenable Adds Ability to Vet Open Source AI Agents for Cybersecurity Teams
Security Boulevard, Monday, October 19th, 2026
Tenable and OpenAI now vet community-contributed security agents for prompt injection, hardcoded secrets and data exposure.
Tenable, working with OpenAI, added an Exchange Inspector that evaluates open source AI agents contributed to its CyberAgents Exchange directory.
An inspection engine checks for malicious content such as prompt injection and for hardcoded secrets or exposed customer data, followed by an assessment with OpenAI's GPT cyber models and a review by Tenable researchers.
Three skills have been vetted so far, including a SOC Hunter agent Tenable says cuts threat-hunting time by up to 75% and a daily remediation planning agent.
Michael Vizard notes adoption is unclear and that shared agents still need strict guardrails and access controls.