GitLab Enhances AI Software Development with New Security Features

New Tools for AI-Driven Software Development

GitLab has rolled out innovative features aimed at enhancing the management of AI-driven software development. These updates focus on various critical areas such as workflow automation, credential management, software supply chain security, and monitoring AI-related costs. This initiative aims to assist organizations in effectively governing AI-generated code as it progresses from development to production.

Significant Growth in AI Development Activity

In the past three months, GitLab has witnessed a remarkable surge in activity among its users engaged in agentic software development, with a substantial 200% year-on-year increase. Other significant growth metrics include a 100% rise in secure repositories, an 80% boost in user namespaces, and a 40% increase in continuous integration and delivery (CI/CD) pipelines.

Streamlining the Development Process

Organizations often rely on multiple disconnected tools for various aspects of software development, including coding, source code management, and security. This fragmented setup can complicate maintaining consistency in identity and policy controls, as well as tracking changes throughout the development lifecycle. To address these challenges, GitLab's new features facilitate integration across different stages through standardized workflows, policies, and audit trails.

Automating Workflows with GitLab Duo Agent

The introduction of goal-driven workflows through the GitLab Duo Agent Platform is a significant enhancement. Features such as the /goal command in Duo CLI, headless mode, and Duo Agentic Chat, combined with the GitLab for Slack app, allow for greater automation of multi-step tasks. This streamlined workflow operates under shared identity and policy controls, effectively tracking movement through various development stages.

Centralized Management of Software Artifacts

GitLab Artifact Central, now in beta, will soon allow users to manage containers and software packages alongside source code management and CI pipelines. By establishing organization-level policies for software artifacts, teams can better track published items and potentially reduce overall ownership costs by up to 50% compared to other tools.

Strengthening Supply Chain Security

To bolster security measures in AI-driven software development, GitLab is introducing new features aimed at managing dependencies, safeguarding credentials, and evaluating security risks.

Enhancements to Dependency Management

The GitLab Dependency Firewall, currently in early access, inspects packages against established organizational policies before they are integrated into a build. Teams can tailor rules based on factors like dependency vulnerabilities, package age, and license compliance, ensuring a secure environment from code inception to deployment.

Centralized Secrets Management

GitLab Secrets Manager, now generally available on GitLab.com, centralizes the management of build-time secrets and restricts access to essential jobs. This capability leverages existing permissions and maintains a secure audit trail, enabling teams to efficiently revoke any compromised credentials.

Improving AI Workflow Efficiency

GitLab is focused on enhancing the efficiency of AI workflows by providing contextual insights and visibility into AI usage.

Introducing GitLab Orbit

Having entered beta in June, GitLab Orbit has already been utilized by over 3,500 organizations and managed more than 280,000 queries from coding agents. By mapping software lifecycle data into a knowledge layer, Orbit helps streamline task completion, significantly reducing retries and token consumption.

Impact Analytics for AI Spending

Additionally, the early access of Duo Agent Platform Impact Analytics provides insights into AI expenditures and outcomes by team and task. This feature enables administrators to set spending limits and effectively track adoption metrics and workflow success, supporting diverse environments that use a variety of models.

Conclusion

With these enhancements, GitLab is positioning itself as a leader in facilitating secure and efficient AI-driven software development. The new capabilities not only streamline processes but also ensure organizations can govern their AI-generated code more effectively as it transitions through various stages of development.