Operate Railway infrastructure through CLI, MCP, and GraphQL workflows.
npx skills add https://github.com/Chia1104/agent-air --skill use-railway-chia1104
Please help me install this Agent Skill. Skill: use-railway Source: https://github.com/Chia1104/agent-air/tree/main/skills/shared/use-railway Command: npx skills add https://github.com/Chia1104/agent-air --skill use-railway-chia1104
railway up to sign up, create projects and services, and deploy from the current directory in one chain, including headless device-code sign-in handling. - Database analysis: Introspect PostgreSQL, MySQL, Redis, and MongoDB services over SSH with per-database metric thresholds, tuning guidance, and log analysis. - Operations and configuration: Manage environments, variables, domains, feature flags, buckets, sandboxes, usage limits, and troubleshoot build or runtime failures. - Use Case: A user pastes a Railway dashboard URL and says "analyze this database" — the Skill extracts the project, service, and environment IDs, detects the database type, runs the matching analysis script, and produces a full performance report. ## Quick Start Ask the assistant to deploy the current directory to Railway or analyze a specific Railway database service by pasting its dashboard URL.High-intent search queries and answers about installing and using this skill.
Run `railway up` from the project directory. It authenticates you if needed, creates a project and service when none is linked, and deploys the code. Use `railway up -y` for non-interactive runs and verify the deployment reaches SUCCESS status.
Provide the Railway dashboard URL or service name, and the Skill runs the matching analysis script (analyze-postgres.py, analyze-mysql.py, analyze-redis.py, or analyze-mongo.py) over SSH. It collects metrics, logs, cache ratios, and query statistics, then produces a full performance report.
Yes. The CLI auto-detects headless environments and switches to a device-code flow that prints a sign-in link and code. Run the command in the background and relay the link immediately, since the code expires after 10 minutes.
Use the CLI for workflows needing local state like `railway up`, SSH, or database scripts. Use remote MCP (https://mcp.railway.com) for OAuth-scoped platform reads, feature flags, and simple redeploys. Use `railway api` for GraphQL operations without a dedicated tool.
`railway up --detach` returns after upload at QUEUED status, not success. Poll `railway deployment list --json` with the same project, environment, and service scope until the deployment shows SUCCESS, and triage FAILED or CRASHED states via logs.
Analysis requires SSH access with a key registered via `railway ssh keys add`. If introspection fails, only infrastructure metrics and logs are available, and no performance conclusions can be drawn. Metrics history is capped at 168 hours.
nansen-aiAuthenticate Nansen CLI, verify installation, and introspect command schemas.
DevelopersGlobalImplement structured logging, distributed tracing, and alerting with OpenTelemetry.
rubrical-worksConfigure Railway preview, staging, and production deployments with GitHub Actions integration.
effiraidDetect deployment platforms and health checks to write CLAUDE.md deploy configurations.