Agentic-ZTA: A Multi-Agent Architecture for Autonomous Zero Trust Enforcement
Agentic-ZTA uses coordinated AI agents and retrieved policies to automate NIST zero-trust access decisions.
Agentic-ZTA operationalizes the NIST SP 800-207 zero-trust control loop with a coordinated multi-agent pipeline. Relevant policies are retrieved through RAG as top-k results and inserted into agent prompts at inference time. A policy-enforcement point enriches each request, specialized agents reason over context and constraints, and a trust algorithm aggregates scores for the final decision. In a testbed of access-control scenarios, the framework reported 95.0% accuracy, 93.9% precision, and 96.3% recall.
- Maps NIST SP 800-207 onto a multi-agent decision pipeline
- RAG supplies top-k policies embedded in agent prompts
- PEP enriches requests before specialized agents score trust
- Testbed results: 95.0% accuracy, 93.9% precision, 96.3% recall
- Organizations
- NIST
From the original article
1 min read at arxiv.orgAgentic AI is emerging as a promising paradigm for automating complex cybersecurity decisions, yet its use in enforcing zero trust introduces significant challenges in safety, reliability, and policy compliance. This paper presents Agentic AI based zero trust architecture (Agentic-ZTA) that operationalizes the NIST SP 800-207 ZTA architecture control loop through coordinated multi- agent decision pipeline.
By Shovan Roy · arxiv.org
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.