Leidos Debuts Autonomous AI Tool to Patch Zero-Day Cyber Threats

Cyber adversaries are increasingly weaponizing artificial intelligence, forcing a shift in how organizations defend their digital infrastructure. To counter this, Leidos has unveiled Parcata, a model-agnostic platform designed to autonomously identify software vulnerabilities and deploy real-time patches before breaches occur within sensitive government and commercial environments.

Corp and Tech·July 29, 2026·reads: 2,114

The Reston-based technology firm developed the tool through its Kudu Dynamics team, utilizing validation from DARPA’s AI Cyber Challenge. Unlike systems tethered to a single large language model, Parcata operates across a diverse set of models to scan both internal and third-party software. By automating the remediation process, the platform aims to neutralize zero-day threats that traditional manual security protocols might miss.

Jason O’Connor, president of Leidos Intelligence, emphasized that the tool is intended to bring order to the chaos of modern cyber warfare. The company has already begun integrating the platform into its own codebases, moving it from internal technical exercises toward broader deployment in mission-critical environments. This launch serves as a core component of the firm’s NorthStar 2030 strategy, which prioritizes scaling AI-enabled defenses for dynamic, high-stakes operational settings.