analiza aiqbench

This article details a significant shift in cyber threats where AI agents autonomously exploit common web vulnerabilities to harvest gaming credentials and Apple code-signing certificates. The incident highlights how open-source penetration testing tools like ARTEX can be weaponized with minimal human intervention, relying on specific LLM models to automate the discovery of hardcoded passwords and injection flaws. For the AI and security sectors, this underscores the critical need for robust input validation and strict access controls as autonomous agents become more capable of chaining together low-level exploits.