Advancements in Cybersecurity and Reverse Engineering

The field of cybersecurity and reverse engineering is rapidly evolving, with a growing focus on proactive mitigation of phishing campaigns and the application of Large Language Models (LLMs) to improve software security. Recent developments have seen the introduction of adaptive multi-agent systems, such as PhishLumos, which can identify entire attack campaigns before they are confirmed by cybersecurity experts. LLMs are also being used to enhance reverse engineering tasks, including binary decompilation and vulnerability detection. Notable papers in this area include PhishLumos, which demonstrated a 100% success rate in identifying phishing campaigns, and SK2Decompile, which achieved a 21.6% average re-executability rate gain over existing baselines. Other significant contributions include the development of AGNOMIN, a novel architecture-agnostic approach for multi-label function name prediction, and CORTEX, a multi-agent LLM architecture for high-stakes alert triage. These advancements have the potential to significantly improve the security and reliability of software systems, and are expected to have a major impact on the field in the coming years. Noteworthy papers include: PhishLumos, which introduced an adaptive multi-agent system for proactive phishing campaign mitigation. SK2Decompile, which presented a novel two-phase approach to decompile from the skeleton to the skin of programs.

Sources

PhishLumos: An Adaptive Multi-Agent System for Proactive Phishing Campaign Mitigation

SoK: Potentials and Challenges of Large Language Models for Reverse Engineering

SK2Decompile: LLM-based Two-Phase Binary Decompilation from Skeleton to Skin

A Global Analysis of Cyber Threats to the Energy Sector: "Currents of Conflict" from a Geopolitical Perspective

Boosting Pointer Analysis With Large Language Model-Enhanced Allocation Function Detection

Binary Diff Summarization using Large Language Models

Characterizing Event-themed Malicious Web Campaigns: A Case Study on War-themed Websites

AGNOMIN - Architecture Agnostic Multi-Label Function Name Prediction

Black-box Context-free Grammar Inference for Readable & Natural Grammars

CORTEX: Collaborative LLM Agents for High-Stakes Alert Triage

MAVUL: Multi-Agent Vulnerability Detection via Contextual Reasoning and Interactive Refinement

Semantics-Aligned, Curriculum-Driven, and Reasoning-Enhanced Vulnerability Repair Framework

POLAR: Automating Cyber Threat Prioritization through LLM-Powered Assessment

Built with on top of