Cloudflare Introduces AI-Powered Vulnerability Remediation Service

Cloudflare has launched a new service, Vulnerability Discovery and Remediation, currently in early access as part of its Managed Defense offering. The service leverages OpenAI Daybreak models, including GPT-5.6 Cyber, to identify vulnerabilities in customer-authorized codebases, propose code patches, and suggest edge mitigations. By integrating data from Cloudflare's Web Application Firewall (WAF) and production traffic, the service prioritizes vulnerabilities based on their exposure to active routes, traffic volume, and associated security events. This context-aware approach helps security teams focus on the most critical threats first. Importantly, customers retain full control over whether to implement proposed patches or mitigations. The service operates semi-automatically, with every proposed solution undergoing rigorous validation before being presented for review. Vulnerability Discovery and Remediation is available by invitation to select customers, with investigations limited to authorized applications and controlled access to relevant data. This initiative addresses the growing challenge of prioritizing vulnerabilities in an era where attackers increasingly use AI to accelerate exploitation.