AWS has announced Continuum, a new security platform that automates the entire lifecycle of code vulnerabilities.
AWS introduces Continuum, a security platform that addresses software vulnerabilities for customers. The platform handles discovery, prioritization, validation, and remediation using AI models and business context, and is now available in a gated preview.
AI is increasingly being used for security purposes or to detect (code) vulnerabilities. For example, OpenAI also recently launched an AI agent to resolve vulnerabilities in code.
Vulnerability lifecycle
The platform initially focuses on code vulnerabilities, moving through four continuous phases. In the first phase, a full scan is performed, identifying existing vulnerabilities and new issues via frontier AI models. Continuum then enriches and prioritizes these findings based on context: reachability, production path, and business impact are all taken into account.
This is followed by validation, where false positives are filtered out. Continuum places findings within the context of the environment and generates reproducible exploit examples in a sandbox. Finally, the platform advises on mitigation or remediation based on code analysis and existing defense mechanisms. Recommended patches are validated by the system itself, including visibility into potential consequences.
Capacities
Continuum is model-agnostic and integrates various frontier models, ensuring it always benefits from the latest developments. In addition to code vulnerabilities, the platform now also includes pentesting and code scanning, as well as threat modeling that automatically generates STRIDE analyses from source code or design documents.
The system starts in learning mode by default, involving human oversight. As confidence grows, Continuum can perform more actions automatically, depending on the risk profile.
The gated preview of AWS Continuum is now available for companies in sectors including finance and technology.