Capital One Launches Open-Source VulnHunter Agentic AI Code Security…

Capital One has released VulnHunter, an open-source agentic AI tool that identifies exploitable software vulnerabilities by reasoning like an attacker rather than scanning for patterns, reducing false alerts and enabling developers to focus on confirmed, remediable defects. The tool uses a falsification engine to validate findings and maps complete attack paths before presenting results to engineering teams. If you build code security, vulnerability scanning or developer security tooling, Capital One just raised the bar: a major bank has released a production-tested agentic AI tool under open source that cuts false positives and maps real attack paths. Your customers now have free proof that AI-first reasoning beats pattern-matching; you need to compete on speed, integration or depth, not on whether AI can do this.