T-MAP: Red-Teaming LLM Agents with Trajectory-Aware Evolutionary Search

While prior red-teaming efforts have focused on eliciting harmful text outputs from large language models LLMs, such approaches fail to capture agent-specific vulnerabilities that emerge through multi-step tool execution, particularly in rapidly growing ecosystems such as the Model Context Protoc...

Introducing the Generative Application Firewall (GAF)

This paper introduces the Generative Application Firewall GAF, a new architectural layer for securing LLM applications. Existing defenses -- prompt filters, guardrails, and data-masking -- remain fragmented; GAF unifies them into a single enforcement point, much like a WAF coordinates defenses fo...

The vulnerability of the Firefox browser, which allows a malicious actor to compromise the integrity and accessibility of protected information

The vulnerability exists in Mozilla Firefox due to an incorrect limitation on event handling, which replaces events related to configuration changes. Exploiting this vulnerability allows malicious actors to remotely alter the positions of icons on the user interface by using specially crafted...