New infosec products of the week: October 2, 2026

Vega II introduces security-trained AI and lasting memory to the Security Operations Center by integrating frontier AI trained for security operations into SOCs, providing teams with a record of learned information, and removing legacy SIEM barriers. Genea has released Genea MCP, a Model Context Protocol server that allows AI agents direct access to the Access Control platform, streamlining tasks like onboarding or setting permissions via natural language commands. BlackFog launched ADX Vision 2.0 to govern generative and agentic AI across the enterprise by introducing seven layers of protection against prompt injection and malicious prompts. Thales has introduced Sentinel Envelope Plus, which hardens compiled applications against AI-assisted reverse engineering, automated vulnerability discovery, and exploit generation without altering source code.

The convergence of security tooling and generative AI reveals a dynamic shift in how enterprise defense mechanisms are being architected, moving beyond static defenses to incorporating learned context and agentic control. The developments from Vega focus on embedding operational memory and advanced AI directly into the SOC workflow, suggesting a trajectory where situational awareness is automated rather than manually aggregated. Simultaneously, Genea’s integration of AI agents via MCP indicates that access management—a core security function—is being redefined from a procedural hurdle into an agentic capability, automating complex administrative tasks. BlackFog and Thales are addressing the emergent threat landscape posed by AI itself: prompt injection and reverse engineering. This signals a necessary defense pivot where the primary adversarial surface shifts from traditional network perimeter exploits to the integrity of the AI systems themselves and the software artifacts they generate. The underlying pattern suggests an increasing tension between the velocity of AI deployment and the slower, deliberate process of establishing security and governance protocols. What structures are being built to manage the inherent risks introduced by autonomous capabilities, and whose agency is being redefined when access controls become language-driven?