Invicti Unveils Agentic Pentest to Automate Web Security Validation

Modern software development cycles often outpace traditional security audits, leaving enterprises to choose between expensive manual penetration tests or noisy, automated scans. Invicti’s new approach attempts to bridge this gap by deploying specialized AI agents that map attack surfaces and analyze authentication flows, while delegating simpler vulnerability checks to the company’s existing DAST engine. This selective use of AI avoids the high compute costs associated with models that attempt to manage every stage of security testing.

The platform’s proprietary reconnaissance engine builds a contextual map of an application, allowing agents to coordinate multi-stage attacks across classes like SQL injection and remote code execution. During early-access testing, the system identified complex business logic flaws that conventional scanners typically miss. According to an early-access participant from a SaaS firm, the tool successfully linked separate security findings into cohesive attack paths, providing developers with clear evidence to accelerate remediation efforts. Invicti intends for the platform to integrate directly into existing development workflows, offering enterprise-grade controls such as role-based access and isolated execution environments to replace or augment human-led security assessments.