GitLab has unveiled version 19.0 of its DevSecOps platform, a release that enhances the integration of AI agents throughout the software development lifecycle. This update arrives at a critical moment as development teams grapple with the rapid pace of AI-driven code generation, while validation and deployment processes struggle to keep up.
Addressing Development Delays
As developers generate code more quickly, the demand for effective validation, testing, and deployment grows. GitLab's new version aims to address these challenges by deploying AI agents at various stages of the development process. A key element of this strategy is the GitLab Duo Agent Platform, which became widely available earlier this year. This platform enables AI agents to carry out tasks across the entire development chain, from planning to security remediation, allowing for parallel processing and reducing the need for sequential manual hand-offs.
Enhanced AI Assistant Features
In version 19.0, GitLab Duo Developer, the AI assistant tailored for developers, receives a suite of new capabilities. Users can now connect the assistant directly to issues or merge requests, creating a more integrated workflow. The assistant can also be activated within discussion threads, autonomously generating code changes or proposals as necessary. These improvements are set to streamline processes and boost collaboration among development teams.
Improving Security with SBOM Scanning
Another notable addition in this update is the general availability of dependency scanning based on Software Bills of Materials (SBOMs). This feature helps organizations identify vulnerabilities within their entire dependency structures, including transitive dependencies that may not be directly included in projects. As the reliance on open-source components increases, the ability to scan for third-party code vulnerabilities becomes essential for maintaining security in modern software development. The new functionality supports multiple programming environments, including Maven, Gradle, and Python.
Expanded AI Model Support
GitLab 19.0 also expands support for various AI models, including Anthropic’s Claude Opus 4.7 and Google’s Gemini models. The addition of these advanced models enhances capabilities such as code reviews and the restoration of CI/CD pipelines, among other functionalities. Self-hosted options for the Mistral AI models and open-source alternatives like Devstral 2 and GLM-5.1 are now available, providing flexibility for organizations seeking customized solutions.
Organizational Efficiency and Policy Management
For larger enterprises, the update introduces organization-wide review policies, simplifying the administrative process by removing the need to configure settings for individual repositories. This change is expected to significantly reduce overhead for organizations managing numerous projects, allowing teams to concentrate more on development rather than administrative tasks.
GitLab's version 19.0 marks a significant step toward deeper AI integration in software development workflows, addressing current inefficiencies and bolstering security measures. As development teams continue to adapt to rapid innovation, tools like GitLab 19.0 could shape the future of DevSecOps.
The stories that move AI & crypto markets — before the market reacts.
Free. 7am ET. Five stories. 62,400 readers.
