Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

MobiLLM: An Agentic AI Framework for Closed-Loop Threat Mitigation in 6G Open RANs

Prakhar Sharma, Haohuang Wen, Vinod Yegneswaran, Ashish Gehani, Phillip Porras, Zhiqiang Lin

TL;DR

The paper addresses the lack of closed-loop, automated threat mitigation in 6G O-RAN by introducing MobiLLM, a knowledge-grounded, multi-agent AI framework. It combines Threat Analysis, Threat Classification anchored to MITRE FiGHT via Retrieval-Augmented Generation, and a two-layer Response Planning/Execution pipeline that uses safe, predefined network APIs with human oversight. Key contributions include the first end-to-end agentic architecture for cellular threat analysis, planning, and response; evaluation across five real-world 6G threat scenarios; and a discussion of guardrails and safety to prevent hallucinations or unsafe actions. Findings show strong FiGHT retrieval (Top-3 accuracy ~94%) but mixed remediation action success (~64%), underscoring the need for telecom-specific fine-tuning; the work demonstrates the feasibility of autonomous security operations in 6G and provides a blueprint for trustworthy AI-driven network defenses.

Abstract

The evolution toward 6G networks is being accelerated by the Open Radio Access Network (O-RAN) paradigm -- an open, interoperable architecture that enables intelligent, modular applications across public telecom and private enterprise domains. While this openness creates unprecedented opportunities for innovation, it also expands the attack surface, demanding resilient, low-cost, and autonomous security solutions. Legacy defenses remain largely reactive, labor-intensive, and inadequate for the scale and complexity of next-generation systems. Current O-RAN applications focus mainly on network optimization or passive threat detection, with limited capability for closed-loop, automated response. To address this critical gap, we present an agentic AI framework for fully automated, end-to-end threat mitigation in 6G O-RAN environments. MobiLLM orchestrates security workflows through a modular multi-agent system powered by Large Language Models (LLMs). The framework features a Threat Analysis Agent for real-time data triage, a Threat Classification Agent that uses Retrieval-Augmented Generation (RAG) to map anomalies to specific countermeasures, and a Threat Response Agent that safely operationalizes mitigation actions via O-RAN control interfaces. Grounded in trusted knowledge bases such as the MITRE FiGHT framework and 3GPP specifications, and equipped with robust safety guardrails, MobiLLM provides a blueprint for trustworthy AI-driven network security. Initial evaluations demonstrate that MobiLLM can effectively identify and orchestrate complex mitigation strategies, significantly reducing response latency and showcasing the feasibility of autonomous security operations in 6G.

MobiLLM: An Agentic AI Framework for Closed-Loop Threat Mitigation in 6G Open RANs

TL;DR

The paper addresses the lack of closed-loop, automated threat mitigation in 6G O-RAN by introducing MobiLLM, a knowledge-grounded, multi-agent AI framework. It combines Threat Analysis, Threat Classification anchored to MITRE FiGHT via Retrieval-Augmented Generation, and a two-layer Response Planning/Execution pipeline that uses safe, predefined network APIs with human oversight. Key contributions include the first end-to-end agentic architecture for cellular threat analysis, planning, and response; evaluation across five real-world 6G threat scenarios; and a discussion of guardrails and safety to prevent hallucinations or unsafe actions. Findings show strong FiGHT retrieval (Top-3 accuracy ~94%) but mixed remediation action success (~64%), underscoring the need for telecom-specific fine-tuning; the work demonstrates the feasibility of autonomous security operations in 6G and provides a blueprint for trustworthy AI-driven network defenses.

Abstract

The evolution toward 6G networks is being accelerated by the Open Radio Access Network (O-RAN) paradigm -- an open, interoperable architecture that enables intelligent, modular applications across public telecom and private enterprise domains. While this openness creates unprecedented opportunities for innovation, it also expands the attack surface, demanding resilient, low-cost, and autonomous security solutions. Legacy defenses remain largely reactive, labor-intensive, and inadequate for the scale and complexity of next-generation systems. Current O-RAN applications focus mainly on network optimization or passive threat detection, with limited capability for closed-loop, automated response. To address this critical gap, we present an agentic AI framework for fully automated, end-to-end threat mitigation in 6G O-RAN environments. MobiLLM orchestrates security workflows through a modular multi-agent system powered by Large Language Models (LLMs). The framework features a Threat Analysis Agent for real-time data triage, a Threat Classification Agent that uses Retrieval-Augmented Generation (RAG) to map anomalies to specific countermeasures, and a Threat Response Agent that safely operationalizes mitigation actions via O-RAN control interfaces. Grounded in trusted knowledge bases such as the MITRE FiGHT framework and 3GPP specifications, and equipped with robust safety guardrails, MobiLLM provides a blueprint for trustworthy AI-driven network security. Initial evaluations demonstrate that MobiLLM can effectively identify and orchestrate complex mitigation strategies, significantly reducing response latency and showcasing the feasibility of autonomous security operations in 6G.

Paper Structure

This paper contains 14 sections, 2 figures, 1 table.

Table of Contents

  1. Introduction
  2. System Design and Implementation
  3. Threat Analysis Agent
  4. Threat Classification Agent
  5. Response Planning and Execution Agent
  6. Implementation Details
  7. Evaluation
  8. Evaluation Setup
  9. Quantitative Results
  10. Overall Remediation Validity
  11. Discussion and Future Work
  12. General purpose vs. specialized agents
  13. Accessible and robust O-RAN security
  14. Conclusion

Figures (2)

  • Figure 1: Overview of MobiLLM's agentic architecture. The config tuning agent can be invoked autonomously, but the effects are only applied after human approval is gathered.
  • Figure 2: This figure illustrates the interaction flow of MobiLLM, where specialized agents sequentially process incident telemetry to classify a threat and generate a response plan. The resulting configuration change is then presented for mandatory human approval before being executed by the appropriate tool.