AI agents have made code abundant. For government teams, the harder question is whether that code can be trusted and whether that trust can be proven to an authorizing official.

In this episode of Tanuki Trends, host Liz Burrows is joined by Omer Azaria, VP of Product for Security and Compliance at GitLab, and Joel Krooswyk, GitLab’s former Federal CTO. They discuss:

• Why the bottleneck has moved from writing code to deciding what to trust
• How risk is arriving faster than teams can respond
• How far agencies should let AI fix what it finds
• What the GitLab Security Standard offers in response

They also look at what recent Department of War and CISA publications mean for running AI in a governed software factory: giving every agent its own identity, bringing security agents into the factory alongside the ones writing code and knowing which models run where and what happens when one fails.

They close with the one question every agency should be asking about its pipeline today and what a security leader can do about it in the next 90 days. We hope you'll tune in on December 10th!