New approaches for finding business logic flaws via annotation-based sanitisation and detecting site isolation bypasses in modern browsers.

Background Autonomous malware analysis requires moving beyond static, predefined containment environments. Modern malware actively fingerprints its execution context, checking CPUID leaves, measuring execution time with RDTSC, and probing for virtualization-specific hardware devices [6, 13]. Traditional sandboxing methodologies rely on a one-size-fits-all containerization approach, which typically embeds an agent inside the guest OS to monitor API calls and system behavior [1, 12]. This in-guest presence provides a highly visible artifact for evasive malware to detect and subvert. ...

New findings on the attack surface of LLM agents, from cross-channel fragmentation in tool-calling pipelines to adversarial task contamination and supply chain poisoning of self-modifying code.

A layered analysis of threats to Kubernetes environments, exploring container escapes, control plane vulnerabilities, and real-world case studies.

A systematic evaluation of cache side-channel attacks on modern Intel microarchitectures, and a new security architecture to mitigate integrity violations in LLM-integrated app systems.