Worked on the k3s-io/k3s repository to enhance the reliability of the Airgap image workflow by addressing a bug in the Airgap Image Watcher. Focused on backend development using Go and containerd, the solution modified the watcher logic to trigger image re-imports whenever either the modification timestamp or file size changed, rather than requiring both conditions. This adjustment ensures that updated airgap images are consistently detected and processed by the agent, improving workflow stability. The code change was deliberately scoped to minimize impact, adhering to repository conventions and supporting safer, faster deployments while maintaining alignment with established project practices.