GitLab 18.8: GitLab Duo Agent Platform is now generally available.

From assistant to agent.

Whereas GitLab Duo has mainly responded to individual prompts until now, the Agent Platform introduces a different way of working. AI agents can perform tasks independently, retain context, and take action within fixed frameworks. This includes analysing code, preparing merge requests, or flagging issues in pipelines, without the need for manual instructions.

What makes this platform new?

The Agent Platform forms the technical basis for multiple Duo agents. These agents are task-oriented and designed to work securely with existing GitLab components such as repositories, issues, and CI/CD.

Key features:

• Task-oriented agents that independently execute steps within a defined goal

• Context-aware working, giving agents a better understanding of what is going on in a project

• Built-in governance, with clear boundaries around permissions and actions

• Extensible design, allowing GitLab to add new agents in the future without reinventing the platform

What does GA mean in practice?

GA status means that the platform is stable enough for production use. Organisations can deploy it in existing workflows, with support and documentation as you would expect from GitLab. It is no longer an experimental feature, but a foundation on which to build further. For teams already working with GitLab Duo, this also changes the role of AI. Fewer random suggestions, more structural support for recurring tasks.

Impact on development teams.

The introduction of the Agent Platform is not just about speed, but also about focus. By automating routine tasks, there is more room for substantive choices, better reviews and greater attention to quality. At the same time, control remains with the team: agents work within the rules that you set yourself.

AI as an integral part of the development process.

With GitLab 18.8 and the GA of the GitLab Duo Agent Platform, GitLab is demonstrating where it wants to go with AI: not as a gimmick, but as an integral part of the development process. For teams that want to deepen their workflow without adding extra complexity, this is an important step forward.