OpenHack: AI-Powered Vulnerability Research by Hadrian

A new open-source project called OpenHack offers AI-powered vulnerability research. This tool, developed by the Dutch security firm Hadrian, uses AI coding harnesses to review application code. OpenHack is a set of agents and tools that mimics how Hadrian's research team finds vulnerabilities. It operates within a coding harness and keeps durable state in plain files. A human operator approves every phase transition in the process. Hadrian's CEO, Rogier Fischer, states that OpenHack's effectiveness proves security teams can find critical vulnerabilities without needing other tools. The system uses a checkpointed, scenario-first review model. It defines 12 expert families aligned with OWASP and MITRE standards. These cover categories like Broken Access Control, Injection, and Cryptographic Failures. This tool could help security teams more efficiently discover critical vulnerabilities in application code.