Hi Ketan, the VSCode plugin sounds like an interesting feature, looking forward to it! Though, I think it addresses a slightly different problem - we sometimes have very long running productive workloads, which produce a lot of files. If such a task fails at upload or just after a long period of time, we are only able to reproduce the problem by rerunning the task and waiting for hours or hypothetically even days, to arrive at the exact same state, while, in fact, we could debug the task by simply looking at the created files. Kubernetes directly does not provide a way to run a shell in a completed pod, but maybe someone was already in a similar situation? A low-level solution would absolutely be enough for the time being.