Agentic coding is moving from demos into production. Across Apache projects, maintainers are asking: How can we integrate LLM-powered agents into our development workflows without compromising code quality, reviewability, or the ASF’s community-driven principles? From code generation and test writing to refactoring, debugging, code review assistance, and release automation, projects across the foundation are experimenting with agents across the entire software lifecycle. These early efforts reveal not only new possibilities but also real friction points.

This track focuses on hands‑on experiences from ASF contributors who are building, integrating, or evaluating code agents. We want to hear what actually works when agents participate in discussions, propose patches, or help with large‑scale maintenance. What breaks, what becomes easier, and what repeatable patterns are emerging? We’re looking for talks that share practical lessons to help the community turn agents from flashy helpers into dependable collaborators.