Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Deploy code to Railway using "railway up". Use when user wants to push code, says "railway up", "deploy", "ship", or "push". For initial setup or creating services, use railway-new skill. For Docker images, use railway-environment skill.
.claude/skills/davila7-railway-deploy/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 24% | 0% |
| case-02 | ✗→✓ | ▲ Improved | 115% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 23% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -12% | 0% |
| case-06 | ✗→✓ | ▲ Improved | -30% | 0% |
Deploy code from the current directory to Railway using railway up.
Starts deploy and returns immediately. Use for most deploys.
bashrailway up --detach
Streams build logs until complete. Use when user wants to watch the build or needs to debug issues.
bashrailway up --ci
When to use CI mode:
Default is linked service. To deploy to a different service:
bashrailway up --detach --service backend
Deploy to a project without linking first:
bashrailway up --project <project-id> --environment production --detach
Requires both --project and --environment flags.
| Flag | Description | |------|-------------| | -d, --detach | Don't attach to logs (default) | | -c, --ci | Stream build logs, exit when done | | -s, --service <NAME> | Target service (defaults to linked) | | -e, --environment <NAME> | Target environment (defaults to linked) | | -p, --project <ID> | Target project (requires --environment) | | [PATH] | Path to deploy (defaults to current directory) |
Railway CLI walks UP the directory tree to find a linked project. If you're in a subdirectory of a linked project, you don't need to relink.
For subdirectory deployments, prefer setting rootDirectory via the railway-environment skill, then deploy normally with railway up.
Deploying to <service>...Use railway-deployment skill to check build status (with --lines flag).
Build logs stream inline. If build fails, the error will be in the output.
Do NOT run railway logs --build after CI mode - the logs already streamed. If you need more context, use railway-deployment skill with --lines flag (never stream).
No Railway project linked. Run `railway link` first.No service linked. Use --service flag or run `railway service` to select one.The build logs already streamed - analyze them directly from the railway up --ci output. Do NOT run railway logs after CI mode (it streams forever without --lines).
Common issues:
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-19 | pass→pass | 9,086 | 3,788 | -58% | 1 | 1 | 0% | 1,417 | 1,301 | -8% | 0 | 0 | — |
case-01 | fail→pass | 5,031 | 1,498 | -70% | 1 | 1 | 0% | 781 | 965 | +24% | 0 | 0 | — |
case-02 | fail→pass | 3,303 | 1,834 | -44% | 1 | 1 | 0% | 443 | 954 | +115% | 0 | 0 | — |
case-03 | fail→pass | 4,600 | 1,550 | -66% | 1 | 1 | 0% | 768 | 948 | +23% | 0 | 0 | — |
case-04 | pass→pass | 6,245 | 1,678 | -73% | 1 | 1 | 0% | 995 | 897 | -10% | 0 | 0 | — |
case-05 | fail→pass | 6,927 | 1,504 | -78% | 1 | 1 | 0% | 1,074 | 940 | -12% | 0 | 0 | — |
case-06 | fail→pass | 8,267 | 1,521 | -82% | 1 | 1 | 0% | 1,312 | 912 | -30% | 0 | 0 | — |
case-07 | pass→pass | 3,482 | 1,432 | -59% | 1 | 1 | 0% | 477 | 927 | +94% | 0 | 0 | — |
case-08 | fail→fail | 5,772 | 7,305 | +27% | 1 | 1 | 0% | 941 | 1,234 | +31% | 0 | 0 | — |
case-09 | pass→pass | 9,989 | 5,139 | -49% | 1 | 1 | 0% | 1,743 | 1,552 | -11% | 0 | 0 | — |
case-10 | fail→fail | 4,837 | 3,648 | -25% | 1 | 1 | 0% | 680 | 1,407 | +107% | 0 | 0 | — |
case-11 | fail→pass | 7,981 | 4,769 | -40% | 1 | 1 | 0% | 1,376 | 1,540 | +12% | 0 | 0 | — |
case-12 | pass→pass | 4,051 | 2,992 | -26% | 1 | 1 | 0% | 663 | 1,282 | +93% | 0 | 0 | — |
case-13 | pass→pass | 9,249 | 3,980 | -57% | 1 | 1 | 0% | 1,513 | 1,431 | -5% | 0 | 0 | — |
case-14 | pass→pass | 4,428 | 2,503 | -43% | 1 | 1 | 0% | 690 | 1,103 | +60% | 0 | 0 | — |
case-15 | pass→pass | 6,918 | 2,453 | -65% | 1 | 1 | 0% | 1,153 | 1,038 | -10% | 0 | 0 | — |
case-16 | pass→pass | 14,468 | 7,211 | -50% | 1 | 1 | 0% | 2,212 | 1,974 | -11% | 0 | 0 | — |
case-17 | fail→pass | 3,061 | 1,960 | -36% | 1 | 1 | 0% | 476 | 936 | +97% | 0 | 0 | — |
case-18 | fail→pass | 6,867 | 1,669 | -76% | 1 | 1 | 0% | 1,066 | 966 | -9% | 0 | 0 | — |
case-20 | pass→pass | 8,348 | 5,979 | -28% | 1 | 1 | 0% | 1,471 | 1,695 | +15% | 0 | 0 | — |
case-21 | pass→pass | 7,176 | 3,905 | -46% | 1 | 1 | 0% | 1,135 | 1,384 | +22% | 0 | 0 | — |
case-22 | fail→pass | 7,544 | 1,901 | -75% | 1 | 1 | 0% | 1,320 | 961 | -27% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted, and 21 counted toward the lift figure. The other 1 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +41 percentage points is the difference between those two pass rates over the 21 comparable cases.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.