How do I install the Deploy skill?

Run npx skillmds@latest add skillist-io/deploy in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is Deploy safe to use?

This skill has not completed SkillMD's automated safety review yet. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Deploy?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Deploy free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

Who published Deploy?

skillist-io (@skillist-io) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Deploy by skillist-io — Agent Skill

Name: Deploy
Author: skillist-io

File contents

Deploy & release

Default: you do not deploy manually. CI owns it.

Normal path (automatic)

.github/workflows/ci.yml runs on push to main: check → typecheck → test → playwright → build, and only if all pass, deploys:

api → wrangler deploy --config apps/api/wrangler.production.jsonc
web → built with VITE_API_URL=https://api.skillist.io, then deployed
docs → wrangler deploy

Then a smoke job runs pnpm smoke against production. So: land it on main via a merged PR and let CI deploy. Don't run production wrangler deploy by hand unless explicitly asked to hotfix.

Manual deploy (only if explicitly requested)

Requires Cloudflare credentials/secrets already configured. Use the production config: apps/api/wrangler.production.jsonc (custom domains, real resource IDs, larger containers). Never deploy the base wrangler.jsonc (that's local-dev bindings) to production.
Worker secrets are managed separately: pnpm setup:secrets. Never inline secrets into vars.
Schema ships before the code that reads it. deploy-api runs pnpm db:migrate against the PRODUCTION_DATABASE_URL secret before wrangler deploy. If you deploy the API by hand, apply pending migrations first — skipping that shipped a Worker querying a column production did not have, and 500'd the whole observability endpoint.
Locally, .env's DATABASE_URL is the dev branch. Production is DATABASE_URL_PROD, and it must be passed explicitly: DATABASE_URL="$(grep -m1 '^DATABASE_URL_PROD=' .env | cut -d= -f2-)" pnpm db:migrate

npm packages (`@skillist/skill-format`, `@skillist/cli`)

Published via the manual workflow_dispatch workflow .github/workflows/publish-packages.yml (builds skill-format then cli, skips versions already published, uses NPM_TOKEN). Locally: pnpm publish:packages. Bump the package version first; the workflow no-ops if the version already exists.

Before any deploy

Run /preflight (or pnpm check && pnpm typecheck && pnpm test). Deploys should only follow green gates.

---name: deploydescription: How Skillist deploys (CI-driven on main) and the guardrails for any manual Cloudflare deploy. Use when asked about deploying, releasing, wrangler deploy, or publishing the npm packages.---# Deploy & release**Default: you do not deploy manually.** CI owns it.## Normal path (automatic)`.github/workflows/ci.yml` runs on push to `main`: `check → typecheck → test → playwright → build`, and only if all pass, deploys:- **api** → `wrangler deploy --config apps/api/wrangler.production.jsonc`- **web** → built with `VITE_API_URL=https://api.skillist.io`, then deployed- **docs** → `wrangler deploy`Then a **smoke** job runs `pnpm smoke` against production. So: land it on `main` via a merged PR and let CI deploy. Don't run production `wrangler deploy` by hand unless explicitly asked to hotfix.## Manual deploy (only if explicitly requested)- Requires Cloudflare credentials/secrets already configured. Use the **production** config: `apps/api/wrangler.production.jsonc` (custom domains, real resource IDs, larger containers). Never deploy the base `wrangler.jsonc` (that's local-dev bindings) to production.- Worker secrets are managed separately: `pnpm setup:secrets`. Never inline secrets into `vars`.- **Schema ships before the code that reads it.** `deploy-api` runs `pnpm db:migrate` against the `PRODUCTION_DATABASE_URL` secret before `wrangler deploy`. If you deploy the API by hand, apply pending migrations first — skipping that shipped a Worker querying a column production did not have, and 500'd the whole observability endpoint.- Locally, `.env`'s `DATABASE_URL` is the **dev** branch. Production is `DATABASE_URL_PROD`, and it must be passed explicitly: `DATABASE_URL="$(grep -m1 '^DATABASE_URL_PROD=' .env | cut -d= -f2-)" pnpm db:migrate`## npm packages (`@skillist/skill-format`, `@skillist/cli`)Published via the **manual** `workflow_dispatch` workflow `.github/workflows/publish-packages.yml` (builds skill-format then cli, skips versions already published, uses `NPM_TOKEN`). Locally: `pnpm publish:packages`. Bump the package `version` first; the workflow no-ops if the version already exists.## Before any deployRun `/preflight` (or `pnpm check && pnpm typecheck && pnpm test`). Deploys should only follow green gates.

Frequently asked questions

Deploy