Deployment

You're staring at Railway dashboard logs trying to figure out why your deployment failed at 3am. This handles the forensics: pulls recent logs with line counts, checks deployment status, triggers redeploys when you need to bounce a service, and removes deployments that shouldn't have shipped. It's the subset of Railway operations you reach for when something's on fire or you need to clean up after a bad push. The commands are straightforward, bash based, nothing clever. If you're doing the full Railway setup dance with buckets and variables, you want the parent use-railway skill. This is just the deployment lifecycle stuff.