GitLab has launched version 19.0, marking a significant shift in how the platform approaches software development. The release centres on what the company calls 'intelligent orchestration' – a recognition that whilst AI coding assistants have accelerated code writing, the processes that follow remain frustratingly slow. Reviews, security scans, pipelines, and deployments still create manual bottlenecks that prevent teams from capitalising on faster code generation. GitLab 19.0 aims to bridge this gap by extending AI agents across the entire software lifecycle.
The update expands the GitLab Duo Agent Platform, which became generally available in January 2026. These AI agents now operate throughout the complete development cycle, from initial planning through to security remediation, running tasks simultaneously rather than waiting for human intervention at each stage. Developers can now assign Duo Developer to issues, generate merge requests directly, or mention it in discussion threads, allowing the agent to pick up work autonomously without requiring constant context-switching.
A standout feature is the SBOM-based dependency scanner, now generally available for Maven, Gradle, and Python projects. This tool provides comprehensive visibility into vulnerabilities across entire dependency trees, including transitive dependencies that aren't directly declared. This is particularly crucial given that approximately 70 per cent of critical security debt originates from third-party code, according to Veracode's 2025 report. The release also adds support for Claude Opus 4.7, Google's Gemini models, and open-source options like Devstral 2 and GLM-5.1 for self-hosted deployments.
GitLab has introduced group-level custom review instructions, eliminating the need to duplicate configurations across individual projects. This streamlines setup for organisations managing numerous repositories. Infrastructure changes include Valkey replacing Redis as the default, removal of bundled Mattermost, and dropped support for Ubuntu 20.04 – breaking changes requiring planning for self-managed customers upgrading from version 18.
The company positions this release as addressing what it calls the 'AI paradox': developers write code faster, yet overall delivery velocity hasn't improved proportionally. GitLab's answer involves orchestrating AI agents across the entire pipeline rather than focusing solely on code generation. To manage costs, GitLab Credits meter AI agent usage at one dollar per credit, with Premium customers receiving 12 credits monthly and Ultimate customers getting 24. With the AI coding tools market reaching an estimated $12.8 billion in 2026, GitLab is betting that integrated platform orchestration offers advantages over standalone specialist tools.
Fuente Original: https://thenextweb.com/news/gitlab-19-intelligent-orchestration-agentic-devops
Artículos relacionados de LaRebelión:
- Cloudsmith Secures 72M for AI Software Supply Chain Security
- CPUID Supply Chain Attack Delivers Trojanised Software
- IA Clona Software de Codigo Abierto Instantaneamente
- AI Revolutionising Software Development Lifecycles
- Malware Autopropagable Ataca Software de Codigo Abierto
Artículo generado mediante LaRebelionBOT
No hay comentarios:
Publicar un comentario