September 2025 Summaries

GitLab has introduced Claude Sonnet 4.5, the latest advanced coding model from Anthropic, into its GitLab Duo model selector, allowing users to choose from a range of leading models to enhance their development experience with improved tool orchestration, context editing, and domain-specific capabilities. This model, noted for its strong performance in areas such as cybersecurity, finance, and research-heavy workflows, offers GitLab users enhanced insights and deeper context in their development tasks. Claude Sonnet 4.5, when used with the GitLab Duo Agent Platform, facilitates smarter assistance and integrates seamlessly throughout the software lifecycle, providing faster development and more precise outcomes. GitLab Duo Pro and Enterprise customers can access this model immediately, with availability in supported IDEs anticipated shortly, and interested users are encouraged to explore GitLab's documentation for further details.

Agentic AI represents a significant advancement in artificial intelligence, enabling systems to act independently by utilizing advanced language models and natural language processing, unlike traditional AI tools that require constant human oversight. GitLab, in its recent updates, has integrated agentic AI into its platforms to enhance software development processes, such as automating complex workflows, improving code quality, and facilitating seamless human-AI collaboration. Their GitLab Duo Agent Platform and integration with Amazon Q exemplify the fusion of DevSecOps with AI, offering features like AI-driven code analysis, automated test generation, and optimized code reviews. These developments promise to transform software engineering by promoting intelligent, adaptive workflows and scaling platform engineering potential, ultimately reshaping how developers interact with AI to boost productivity and maintain security.

Part 7 of the guide on the GitLab Duo Agent Platform focuses on leveraging the Model Context Protocol (MCP) to enhance AI capabilities within development workflows. MCP, an open standard introduced by Anthropic in 2024, acts as a secure interface allowing AI tools to access and interact with internal systems and data sources beyond their original training data. The integration of MCP into GitLab enables developers to seamlessly incorporate AI into their workflows by using natural language within the IDE, without needing to switch contexts. GitLab supports MCP both as a client, accessing external services, and as a server, allowing AI tools to connect securely to GitLab instances for comprehensive project interaction. This dual support facilitates AI's meaningful assistance across development tools while maintaining security and simplifying integration. As part of this guide, practical examples and interactive demos illustrate the application of MCP in real-world scenarios, emphasizing the potential for streamlined workflows and enhanced productivity.

GitLab has been recognized as a Leader for the third year in a row in the 2025 Gartner Magic Quadrant for DevOps Platforms, excelling in categories such as Agile Software Delivery, Cloud-Native Application Delivery, Platform Engineering, and Regulated Delivery. This acknowledgment highlights GitLab's comprehensive platform strategy, which is particularly relevant as organizations strive to integrate AI capabilities while ensuring security and operational excellence. GitLab's platform offers a unified approach that reduces integration overhead, enhances security, and accelerates innovation without disrupting existing workflows. Key features include AI-powered automation for code reviews and vulnerability management, embedded security measures, and flexible deployment options that meet various regulatory requirements. This versatility allows for consistent innovation and improved productivity, as evidenced by customer testimonials from companies like NatWest. As companies seek to maximize developer productivity and speed up innovation, GitLab's platform approach is becoming increasingly critical, underscored by its recent leadership position in the AI Code Assistants Magic Quadrant.

GitLab 18.4 enhances the software development lifecycle by integrating AI tools and features that promote efficient collaboration, security, and governance. New capabilities include the GitLab Duo AI Catalog for creating and sharing custom agents, the Agentic Chat for seamless interaction with AI agents, and the Knowledge Graph for improved navigation and understanding of codebases. The release also introduces business-aware pipeline maintenance with the Fix Failed Pipelines Flow, allowing for strategic prioritization of fixes based on business impact. Developers can benefit from model selection features and GitLab Duo Context Exclusion for data protection, while expanded MCP tools enhance integration possibilities within the GitLab environment. These advancements are designed to streamline workflows, accelerate development, and maintain strategic alignment, all within a secure and compliant framework.

GitLab has been acknowledged as a Leader in the 2025 Gartner Magic Quadrant for AI Code Assistants, highlighting its GitLab Duo's generative AI code assistance capabilities within the broader AI strategy. GitLab Duo, originally an AI add-on, has evolved into a native component of the GitLab DevSecOps platform, enabling developers to collaborate with multiple AI agents that automate tasks across the software lifecycle. These agents, operating with full project context via GitLab's Knowledge Graph, handle tasks such as code generation and security analysis, thus allowing human developers to focus on higher-value tasks. The GitLab Duo Agent Platform emphasizes security and compliance, ensuring that AI enhances productivity without compromising governance through secure integrations and interoperability with external tools. This platform supports human-agent collaboration through natural-language chat and customizable workflows, maintaining oversight and control. GitLab's continued innovation, including new agents and advanced workflows, aims to amplify productivity and revolutionize AI-native DevSecOps, as indicated by their ongoing commitment to expanding the GitLab Duo Agent Platform.

Modern businesses increasingly rely on web-based platforms for critical operations, making them attractive targets for cybercriminals as digital transformation and remote work expand their attack surfaces. Dynamic Application Security Testing (DAST) becomes essential in this context, as it identifies runtime security vulnerabilities that static code analysis cannot detect. GitLab's integrated DAST solution allows for automated security testing within CI/CD pipelines, facilitating continuous security validation without hindering development workflows. By testing applications in their actual operating environments, DAST can uncover issues like authentication flaws, input validation vulnerabilities, and configuration weaknesses. The integration of DAST into shift-left security workflows not only reduces the cost of fixing vulnerabilities by addressing them early in the development lifecycle but also accelerates time-to-market and empowers developers with immediate security feedback. Additionally, DAST helps organizations comply with regulatory standards like PCI DSS, SOC 2, and ISO 27001 by providing consistent, automated security testing. GitLab DAST supports both passive and active scanning methodologies, allowing for comprehensive vulnerability detection and seamless integration into existing security strategies. It offers flexible scanning options, including on-demand and scheduled scans, and can be governed by centralized security policies to ensure consistent security standards across all projects.

The text details a walkthrough of using the GitLab Duo Agent Platform to automate the generation of dbt models for managing Reddit Ads data within an enterprise data platform, Snowflake. The process involves extracting data from the Reddit Ads API to Snowflake's raw layer through Fivetran, then automating the creation of dbt models to transform this data through the prep and prod layers without manual coding. The GitLab Duo platform efficiently generates comprehensive dbt models, including source and workspace models, complete with tests and documentation, in under 10 minutes—a task that would typically take hours manually. Key steps include preparing data structures, setting up GitLab Duo with Visual Studio Code, and using specific prompts to automate model generation. The walkthrough emphasizes the significance of metadata preparation, clear context provision, thorough validation, and AI utilization for follow-up tasks, showcasing a significant increase in developer efficiency and code quality.

GitLab and Accenture have formed a global reseller agreement, making Accenture an authorized reseller and Professional Services Provider for GitLab's complete DevSecOps platform, available through channels like the AWS Marketplace. This collaboration aims to merge GitLab's DevSecOps platform with Accenture's expertise in digital transformation and implementation services to help organizations build secure software at scale. The partnership focuses on enterprise-scale DevSecOps transformation, mainframe modernization, and AI-driven software development with GitLab Duo and Amazon Q, targeting organizations looking to enhance development velocity while ensuring security and compliance. This agreement provides a flexible global framework adaptable to local conditions, and both companies are committed to accelerating innovation, improving development processes, and enhancing security for their customers.

GitLab has significantly enhanced its CI job status updates by reducing API calls by 92.56%, dropping from 45 million to 3.4 million daily, thanks to implementing WebSockets and GraphQL subscriptions. This shift from traditional polling, which involves frequent network requests regardless of data changes, to event-driven WebSockets allows instant updates only when data changes, thus improving efficiency and reducing latency. By refactoring the job header component to use GraphQL for data and employing ActionCable for real-time updates, users now experience immediate job status changes, and network traffic is drastically cut. This optimization has not increased CPU usage, marking a win for both the software's performance and user experience. GitLab aims to extend this real-time capability across its entire CI/CD workflow, replacing remaining polling mechanisms to enhance user feedback and system efficiency further.

Git Much Faster is an optimization script designed to dramatically reduce the time it takes to clone large Git repositories, addressing a common bottleneck in workflows involving extensive codebases or binary-heavy repositories. By employing strategies such as disabling compression, increasing HTTP buffer sizes, and utilizing shallow and partial clones, the script achieves up to a 93% reduction in clone times and significantly decreases disk space usage. These optimizations are particularly beneficial for CI/CD environments, where reducing clone times can alleviate infrastructure costs and improve productivity. Benchmarks demonstrate that tailored optimizations consistently outperform standard Git and Scalar methods, offering substantial improvements in repository handling across various scenarios, including embedded development, enterprise monorepos, and media-heavy projects. The tool provides a practical approach to enhancing Git performance by focusing on network transfer efficiencies, CPU utilization, and storage patterns, ultimately transforming how teams interact with large codebases.

Retailers face significant challenges in application security due to the complex nature of modern commerce, which includes mobile apps, AI-driven personalization, and omni-channel platforms that expand the attack surface. Traditional security methods, often applied as an afterthought, struggle to keep pace with the rapid innovation and diverse threat vectors characteristic of the retail industry. This complexity is compounded by factors such as supply chain fragility, legacy systems, AI compliance requirements, and customer-facing automation risks. The solution lies in integrating security directly into the development lifecycle via a DevSecOps platform like GitLab, which offers comprehensive security scanning tools and ensures vulnerabilities are addressed before reaching production. This approach not only accelerates secure innovation but also eliminates the inefficiencies of managing multiple disconnected security tools. By transforming security into a shared responsibility across development, security, and operations teams, retailers can maintain high security standards while delivering seamless customer experiences and reducing the costs associated with fragmented toolsets.

GitLab Duo Agent Platform, currently in Beta, facilitates AI-driven interactions with GitLab resources, such as issues and merge requests, to streamline complex tasks throughout the software development lifecycle. It offers two main experiences: conversational agentic chat and automated agent Flows, both of which assist in code generation, security vulnerability resolution, and project analysis while maintaining enterprise-grade security and customizable controls. The "Issue to MR" Flow is a feature that efficiently transforms a well-defined issue into a draft merge request by analyzing the issue's details, creating a development plan, and proposing an implementation, all within the GitLab UI. This Flow aims to reduce development overhead by minimizing the time spent on locating files and navigating complex review processes, thus accelerating application updates. GitLab Duo Agent Platform enhances collaboration by providing a unified data model and built-in security, supporting interoperability and extensibility across various tools, and enabling seamless multi-agent collaboration. It is designed to improve quality and efficiency in development processes by keeping context tight and reducing unnecessary handoffs, with the ability to monitor progress and validate changes throughout the workflow.

Artificial intelligence is increasingly pivotal across industries, necessitating strong governance frameworks to ensure ethical and accountable implementation. GitLab has achieved the ISO/IEC 42001 certification, marking a significant milestone as the first international standard for Artificial Intelligence Management Systems. This certification covers GitLab's AI offerings, including GitLab Duo and the GitLab Duo Agent Platform, which provide AI-driven solutions throughout the software development lifecycle. These solutions include features like asynchronous collaboration between developers and AI agents, code suggestions, vulnerability explanations, and test generation. The certification enhances trust and transparency by adhering to globally recognized AI governance standards, supports strategic risk management, and aligns with evolving global AI regulations. GitLab commits to continuous improvement of its AI capabilities through regular audits and assessments, reinforcing its position as a leader in responsible AI innovation within the DevSecOps community.

GitLab Duo Agent Platform is now generally available, emphasizing transparency, independence, and developer-first principles in AI-driven DevSecOps. With a focus on vendor independence, GitLab offers expanded AI model support, enabling flexibility and avoiding lock-in. The platform integrates comprehensive security and compliance features, reducing reliance on third-party tools and minimizing potential risks. GitLab's commitment to open source and community engagement fosters collaboration and innovation, while their robust data governance policies ensure customers retain control over their data. The AI Transparency Center, launched in 2024, exemplifies GitLab's dedication to clear data usage policies and ethical AI practices. Through flexible deployment options and a cloud-neutral stance, GitLab allows organizations to navigate complex regulatory landscapes effectively while maintaining strategic control over their technology stack.

Rust has become a favorite among developers for its performance, memory safety, and concurrency features, and GitLab's CI/CD platform complements these attributes with robust support for Rust projects. GitLab offers integrated DevSecOps tools, allowing developers to set up automated testing, cross-platform builds, and documentation generation with ease. A key demonstration involves a Rust-based mortgage calculator, which illustrates the platform's capabilities in building, testing, packaging, scanning, and deploying applications. GitLab’s caching mechanisms optimize Rust's lengthy compilation times, while its container-first approach enables seamless deployment across diverse environments, including Kubernetes clusters. Security is enhanced with GitLab's Static Application Security Testing and other integrated tools, providing comprehensive protection and compliance support. Moreover, GitLab Duo AI features improve the developer experience by offering intelligent code suggestions and vulnerability explanations, streamlining the development process and enhancing code quality. Through features like GitLab Pages, the platform also facilitates automatic documentation deployment, ensuring that project documentation remains up-to-date with code changes. Overall, GitLab's ecosystem aligns well with Rust's philosophy, creating an efficient, secure, and productive environment for Rust developers.