From flow
Guides Railway deployments: editing railway.toml/json/Procfile, configuring services/databases/workers/env vars, troubleshooting. Covers serverless patterns, multi-service setups, guardrails.
npx claudepluginhub cofin/flow --plugin flowThis skill uses the workspace's default tool permissions.
<workflow>
Manages Railway infrastructure: creates projects/services/databases/buckets, deploys code, configures environments/variables/domains, troubleshoots failures, monitors status/metrics.
Deploys Node.js, Python, and Docker apps to Railway after pre-flight checks validating project config, env vars, build files, custom domains, and monitoring.
Mandates invoking relevant skills via tools before any response in coding sessions. Covers access, priorities, and adaptations for Claude Code, Copilot CLI, Gemini CLI.
Share bugs, ideas, or general feedback.
For full project management, deployment automation, and service orchestration, we highly recommend installing the official Railway agent skills:
Installation:
npx skills add railwayapp/railway-skills
The patterns below provide additional context for Flow-specific multi-service architectures and serverless constraints.
Railway's serverless feature puts services to sleep after 10 minutes of no outbound traffic.
When to disable serverless (sleepApplication: false):
For applications with background task processing, use distinct configuration files (e.g., railway.app.json and railway.worker.json) to manage different runtime requirements.
sleepApplication: false for background worker services.${{Service.VARIABLE}} syntax (e.g., ${{Postgres.DATABASE_URL}}) instead of hardcoding values.${{PORT}} for application port injection.
sleepApplication: false for Workers: Ensure any background worker has sleepApplication explicitly set to false.${{PORT}}.
Configuration (railway.worker.json):
{
"service": {
"name": "worker",
"sleepApplication": false
},
"deploy": {
"startCommand": "python -m saq my_app.tasks.worker --workers 4"
}
}