In a nutshell: AI systems can discover and evaluate zero-days in real time through automated code analysis and LLM-based vulnerability detection.
Security company Intruder has developed an AI-powered system that automatically uncovers software vulnerabilities by combining code analysis and large language models. The system has already identified and exploited a previously unknown vulnerability in a WordPress plugin.
Intruder has developed a method that combines Large Language Models with code-slicing techniques to automatically identify complex software vulnerabilities. The system works through tokenization and exploratory traversal of code graphs, with AI models recognizing patterns that indicate security flaws.
The company demonstrated the practical applicability of the system using a zero-day vulnerability in a WordPress plugin, which the system independently discovered, verified, and successfully exploited. Additional vulnerabilities were coordinated with affected vendors under responsible disclosure procedures.
The approach has immediate implications for CISOs: on one hand, attack speed is increasing, as adversaries can employ similar techniques for fully automated exploitation. On the other hand, defensive teams now have the opportunity to proactively identify vulnerabilities before attackers find them — provided such systems are established in their own software supply chain assessments.
Source: www.bleepingcomputer.com · Published July 15, 2026
Lumi AI News — AI-assisted curation in accordance with Art. 50 EU AI Act. Paraphrase and classification by Lumi News Pipeline v1.7.3.