# Publish

> Build and publish Domo apps with dist workflow and first-publish ID handling.

- Skill: `stahura/publish` (Agent Skill)
- Install (CLI): `npx skillmds@latest add stahura/publish`
- Raw SKILL.md: https://api.skillmd.com/api/skills/stahura/publish/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Productivity
- Author: stahura (https://skillmd.com/u/stahura)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/stahura/publish

---


# Build & Deploy Workflow

## Prerequisites
- Node.js installed
- Domo CLI installed (`npm install -g @domoinc/ryuu`)

## Local development
```bash
npm install          # Install dependencies
npm run dev          # Start dev server (usually Vite)
```

For API calls to work locally, you need ryuu-proxy configured and `domo login` authenticated.

## Build for production
```bash
npm run build        # Outputs to dist/ (Vite) or build/ (CRA)
```

## Domo CLI Authentication
```bash
domo login           # Authenticate with your Domo instance
```

You'll be prompted for your Domo instance URL and credentials.

## Publishing
```bash
cd dist              # Change to build output directory
domo publish         # Publish to Domo
```

If publish includes new/updated Code Engine packages, run package lifecycle steps first via:

- `~/.agents/skills/code-engine-create/SKILL.md`
- `~/.agents/skills/code-engine-update/SKILL.md`

**Important - First publish:**
- On first publish, Domo generates a new `id` for your app
- This ID appears in the published `manifest.json` in your dist folder
- **You must copy this ID back to your source `manifest.json`** (e.g., `public/manifest.json`)
- If you don't, every publish creates a NEW app instead of updating the existing one

```bash
# After first publish, copy the generated ID:
# dist/manifest.json → public/manifest.json (just the "id" field)
```

## Subsequent publishes
Once the ID is in your source manifest:
```bash
npm run build && cd dist && domo publish
```

## Build Checklist

Before publishing, ensure:
- [ ] `manifest.json` has correct `dataSetId` (not `id`) for datasets
- [ ] `manifest.json` has `fields: []` in each dataset mapping
- [ ] `thumbnail.png` exists and is 300x300 pixels
- [ ] All queries use `.select()` with specific columns (never fetch all columns)
- [ ] Each visualization has its own optimized query
- [ ] No `.aggregate()` calls (use `.groupBy()` or client-side aggregation)
- [ ] `.groupBy()` calls have a grouping column (not just aggregations)