Blog

Running agents inside devcontainers make sense from security perspective. But how to balance power of agents with security concenrns?