Inside the Exchange Inspector: How Tenable uses OpenAI GPT cyber models to review open-source AI agents
Tenable's Exchange Inspector uses OpenAI cyber models and human review to vet open-source security AI agents.
Tenable described Exchange Inspector, a three-stage review for selected listings on its CyberAgents Exchange directory of open-source security agents, skills, MCP servers, and playbooks. Candidates are screened with Tenable One AI Exposure for prompt injection, jailbreaks, hidden instructions, hardcoded secrets, and sensitive-data exposure, then assessed with OpenAI GPT cyber models and verified by researchers in a clean environment. The review covers 15 issue types, from SSRF and path traversal to excessive agent permissions and memory poisoning. Three listings have passed, two built by Tenable and one by Splunk, and Tenable said SOC Hunter reduced its hunt times by 75 percent.
- Vetting uses automated screening, a GPT assessment, and human verification.
- Checks include prompt injection, secrets, SSRF, path traversal, and memory poisoning.
- Three listings are vetted: two from Tenable and one from Splunk.
- Tenable says SOC Hunter cut its threat-hunt times by 75 percent.
- AI models
- OpenAI GPT
From the original article
11 min read at tenable.com
Community-built AI agents, skills, and MCP servers are landing in SOC workflows fast. Here’s what the Exchange Inspector tests before a listing earns its vetted tag on the CyberAgents Exchange. Three tools have already passed.
By Robert McSulla · tenable.com
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.