Search papers, labs, and topics across Lattice.
This paper introduces an agentic workflow for automating security alert investigation by combining LLMs with structured queries over Suricata logs and grep-based text search. The workflow uses LLMs to select and execute predefined queries based on an overview of available data, extract evidence, and deliver a final verdict. Results show that this agentic approach significantly improves the accuracy of alert verdicts compared to using LLMs alone, by addressing limitations of applying LLMs directly to high-volume, unstructured data.
LLMs can be surprisingly effective security analysts, triaging alerts with significantly improved accuracy when guided by structured queries and constrained tool access.
Security analysts are overwhelmed by the volume of alerts and the low context provided by many detection systems. Early-stage investigations typically require manual correlation across multiple log sources, a task that is usually time-consuming. In this paper, we present an experimental, agentic workflow that leverages large language models (LLMs) augmented with predefined queries and constrained tool access (structured SQL over Suricata logs and grep-based text search) to automate the first stages of alert investigation. The proposed workflow integrates queries to provide an overview of the available data, and LLM components that selects which queries to use based on the overview results, extracts raw evidence from the query results, and delivers a final verdict of the alert. Our results demonstrate that the LLM-powered workflow can investigate log sources, plan an investigation, and produce a final verdict that has a significantly higher accuracy than a verdict produced by the same LLM without the proposed workflow. By recognizing the inherent limitations of directly applying LLMs to high-volume and unstructured data, we propose combining existing investigation practices of real-world analysts with a structured approach to leverage LLMs as virtual security analysts, thereby assisting and reducing the manual workload.
