GitLab Joins Open Secure AI Alliance, Says AI Security Should Be Built in the Open

Why GitLab is joining the alliance

According to GitLab, the move follows an argument made by CEO Bill Staples in a recent essay: when code becomes abundant, trust becomes scarce. AI is collapsing the cost of producing software, GitLab says, and everything around production, including context, verification, governance and evidence, becomes proportionally more important.

GitLab says security and collaboration are both key to producing trust. In its view, AI widens what has to be trusted. It now covers the agents that write software as well as the software that gets written.

The Open Secure AI Alliance brings together organizations to develop and share open technologies, techniques and tools that help safeguard software and agents in the age of AI. GitLab says it hopes to share practical work already underway across its AI and security teams and to learn from the broader community.

Open work already in place

GitLab describes its participation as an extension of work it already does in the open. Its AI Gateway is a standalone service that provides access to AI features across GitLab.com, GitLab Dedicated and self-managed instances. According to the company, the source code is public.

So are the agent workflow configurations that run on the AI Gateway. GitLab lists configurations for:

• code creation
• code review
• secrets false-positive detection
• security review

The company links to examples including a Developer Flow configuration, a Code Review Flow configuration and a Security Review Flow. It says these reflect a principle: an AI system is more than the model behind it. Trust, GitLab argues, depends on the surrounding harness, workflows, controls, context and evaluation. Open tooling makes it easier for defenders to understand how those pieces work, test them against real risks and improve them over time.

Model choice and transparency about risks

GitLab also says enterprises need choice. Organizations should be able to use proprietary and secure open weight models, select the right model for the job, and decide where and how AI systems operate. The company says model and cloud agnosticism preserves that flexibility, while an open ecosystem gives customers more options for building, governing and securing AI.

On risk, GitLab says transparency has to extend to the threats themselves. Its documentation on security threats in agentic systems sets out the attack paths that apply to its own agents and the controls that apply in each environment.

The company says it treats its agents as privileged actors. Each one runs with an explicit identity, constrained tools and permissions, and actions that can be audited. GitLab points to documentation on composite identity, sandboxed execution, per-flow tool restrictions and agent tool governance.

What it means

The announcement is a statement of position more than a product launch. GitLab says its goal is not to set open weight and closed models against each other, but to make sure defenders have the transparency, control and flexibility they need to build trustworthy AI systems.

For GitLab customers, the practical point is that the AI Gateway code and several agent workflow configurations are already public and can be inspected. For the wider market, GitLab's stated view is simple: AI security should be built in the open.