GitLab launches beta AI Governance Dashboard to track coding agent activity

One view for agents inside and outside GitLab

GitLab says AI coding agents now write and change code in many places, and that most of their activity stays invisible to the people accountable for it. Teams often struggle to know whether a developer used a tool or model that is approved. GitLab describes the result as a shadow AI problem: real work happening in projects, from many tools, that no one can see in one place or tie to a team member.

The AI Governance Dashboard is meant to address that. It shows which agents ran, who used them and which projects they touched, along with the tools and sessions associated with each agent. GitLab Duo Agent Platform agents run inside GitLab, while externally connected agents run on developers' own machines. Users can toggle between all agents, only Duo Agent Platform agents, or only externally connected agents.

Three tiles at the top show how many agents ran, how many sessions they started and how many tool calls they made over the past seven or 30 days. Each tile includes the daily trend and the change from the previous period, so a spike, such as a new tool the team started using last week, stands out. Cards below help teams check whether the agents developers rely on match the approved ones, see which projects carry the most agent activity, and know who to talk to when usage shifts.

An audit trail for every agent action

GitLab notes that traditional audit trails were built for human changes, so records of agent work are often scattered across tools or missing. The company now records every agent action as an audit event, detailing the model used, the tools called and the project it ran in.

Opening any GitLab Duo Agent Platform session shows the prompt the developer sent, the tools the agent called, the MCP servers it used, the model that ran and the project it changed. From there, users can download an immutable copy of the session as evidence, or stream the events to a security information and event management (SIEM) system. GitLab says this lets teams answer auditors on who ran an agent, what it did and who is accountable, without stitching together multiple data sources.

Setup steps and requirements

GitLab lists four steps to get started:

• Check access. The dashboard needs the Owner role on the top-level group, or a custom role with the read_agent_artifacts ability.
• Turn on AI audit event storage. Storage is off by default, and events are only recorded after it is enabled. In the top-level group, go to Settings > GitLab Duo, select Change configuration, then Store AI audit events in the Data and privacy section. This requires an Owner or Security Manager role.
• Open the dashboard. Go to AI > Governance and select the Dashboard tab to see GitLab Duo Agent Platform activity.
• Connect external agents. Sign in with GitLab CLI (glab) 1.122.0 or later on macOS or Linux. Running glab govern setup installs the Claude Code hooks to capture Claude Code sessions. Running glab govern setup --agents codex,cursor also captures Cursor and Codex sessions through a background job that sends them every 30 minutes. Running glab govern doctor confirms the setup works.

What it means for teams

The AI Governance Dashboard is in beta for GitLab Premium and Ultimate customers. For organizations where developers use a mix of built-in and third-party coding agents, it offers a way to connect agent activity to individual developers and projects, and to produce evidence for audits from one place. Because audit event storage is off by default, teams need to enable it before any events are recorded. GitLab also offers a free trial of GitLab Ultimate for those not yet on the platform.