Melvynx/saveit.now / .agents/skills/ns-setup-r2
ns-setup-r2
VerifiedSet up Cloudflare R2 image/file uploads for NowStack Mobile - create the bucket, get S3 credentials, set the R2_* env vars on the Convex deployment, and verify. Use for "set up image uploads", "configure R2", "Cloudflare storage", "enable file uploads", or "file uploads not working".
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 6.5 KB |
| scripts/setup-cloudflare-r2.mjs | script | 15.7 KB |
SKILL.md
View on GitHub---
name: ns-setup-r2
description: Set up Cloudflare R2 image/file uploads for NowStack Mobile - create the bucket, get S3 credentials, set the R2_* env vars on the Convex deployment, and verify. Use for "set up image uploads", "configure R2", "Cloudflare storage", "enable file uploads", or "file uploads not working".
---
# Setup Cloudflare R2 - NowStack Mobile
Wire Cloudflare R2 so the boilerplate's image/file uploads work. The backend is already built — `convex/storage/r2.ts` issues presigned PUT URLs (`getUploadUrl`) and reads/writes via the S3-compatible API. This skill provisions the bucket + credentials and sets the five `R2_*` env vars **on the Convex deployment** (the action runtime reads them — they do NOT go in `web-app`/`mobile-app` `.env`). For deep Cloudflare platform help, the global `cloudflare` skill covers the rest of R2/Workers/etc.
<what_the_backend_expects>
`convex/storage/r2.ts` reads exactly these (set on the Convex deployment, dev AND prod):
| Var | Value |
| --- | --- |
| `R2_ENDPOINT` | `https://<account-id>.r2.cloudflarestorage.com` (S3 API endpoint) |
| `R2_ACCESS_KEY_ID` | from an R2 API token (S3 credentials) |
| `R2_SECRET_ACCESS_KEY` | from the same token (shown once) |
| `R2_BUCKET_NAME` | your bucket name |
| `R2_PUBLIC_URL` | public base URL for reads — `https://<custom-domain>` or the bucket's `https://pub-<hash>.r2.dev` |
</what_the_backend_expects>
<automated_recommended>
**Fastest path — one script provisions everything** (creates/reuses the bucket, mints a *bucket-scoped* R2 token, derives the S3 credentials, enables the public URL, and sets all five `R2_*` on Convex). This is the same flow NowStack SaaS uses.
You need two things, then one command:
1. **Account ID** — Cloudflare dashboard → R2 (right sidebar), or any dashboard URL `dash.cloudflare.com/<account-id>/...`.
2. **A provisioning API token** — open this pre-filled link (it pre-selects the exact permission groups: R2 edit + Account API Tokens edit + Zone read + DNS edit), name it, create, copy the token:
`https://dash.cloudflare.com/?to=/:account/api-tokens&permissionGroupKeys=%5B%7B%22key%22%3A%22workers_r2%22%2C%22type%22%3A%22edit%22%7D%2C%7B%22key%22%3A%22account_api_tokens%22%2C%22type%22%3A%22edit%22%7D%2C%7B%22key%22%3A%22zone%22%2C%22type%22%3A%22read%22%7D%2C%7B%22key%22%3A%22dns%22%2C%22type%22%3A%22edit%22%7D%5D&name=NowStack%20R2%20Provisioning`
Then run (it reads `CLOUDFLARE_*` from flags or `.env`, never commits them):
```bash
CLOUDFLARE_ACCOUNT_ID=<id> CLOUDFLARE_API_TOKEN=<token> \
node .agents/skills/ns-setup-r2/scripts/setup-cloudflare-r2.mjs --bucket <app-slug>-storage
```
It's idempotent (reuses an existing bucket/token), auto-attaches a custom domain if your `site-config.ts` domain has a Cloudflare zone, else enables the managed `r2.dev` URL. Re-run with `--help` for options (`--public-url`, `--jurisdiction`, `--skip-public-url`). For **prod**, set the same five vars with `npx convex env set --prod …` (the script targets the current dev deployment). Then `npm run check-setup`.
If the script can't run (no provisioning token, or you prefer clicking), use the manual steps below.
</automated_recommended>
<step n="1" title="Create the bucket">
With wrangler (install if missing: `npm i -g wrangler`, then `wrangler login`):
```bash
wrangler r2 bucket create <app-slug>-storage
```
Or in the dashboard: Cloudflare → R2 → Create bucket. Note the bucket name → `R2_BUCKET_NAME`.
</step>
<step n="2" title="Get S3 credentials (R2 API token) — manual, shown once">
R2 S3 access keys are minted in the dashboard (wrangler cannot create them):
Cloudflare → R2 → **Manage R2 API Tokens** → Create API token → permission **Object Read & Write**, scoped to the bucket (or all buckets). On create it shows:
- **Access Key ID** → `R2_ACCESS_KEY_ID`
- **Secret Access Key** → `R2_SECRET_ACCESS_KEY` (copy now — never shown again)
- **S3 endpoint** `https://<account-id>.r2.cloudflarestorage.com` → `R2_ENDPOINT`
Never commit or print these.
</step>
<step n="3" title="Public read URL">
R2 buckets are private by default; `R2_PUBLIC_URL` is the base for serving files. Two options:
- **Custom domain (recommended for prod):** bucket → Settings → Public access → Connect Domain (e.g. `cdn.yourdomain.com`). `R2_PUBLIC_URL=https://cdn.yourdomain.com`.
- **Dev `r2.dev` URL:** bucket → Settings → Public access → enable r2.dev. `R2_PUBLIC_URL=https://pub-<hash>.r2.dev`.
</step>
<step n="4" title="Set the env vars on Convex (dev + prod)">
```bash
npx convex env set R2_ENDPOINT "https://<account-id>.r2.cloudflarestorage.com"
npx convex env set R2_ACCESS_KEY_ID "<access-key-id>"
npx convex env set R2_SECRET_ACCESS_KEY "<secret>"
npx convex env set R2_BUCKET_NAME "<bucket>"
npx convex env set R2_PUBLIC_URL "https://<custom-domain-or-r2.dev>"
# repeat with --prod before launch:
npx convex env set --prod R2_ENDPOINT "..." # ...and the other four
```
Verify: `npm run check-setup` — the R2 warnings must clear.
</step>
<step n="5" title="Verify the round-trip">
From a signed-in client, call `api.storage.r2.getUploadUrl({ key, contentType })`, `PUT` the file to the returned URL, then load `${R2_PUBLIC_URL}/${key}`. Or quick backend check:
```bash
npx convex run storage/r2:internalUploadBuffer '{"key":"test/hello.txt","base64":"aGk=","contentType":"text/plain"}'
# returns the public URL; open it — if it 404s, R2_PUBLIC_URL or public access is wrong.
```
</step>
<usage>
Client upload pattern (the backend is ready; wire it into your product UI):
1. `const url = await getUploadUrl({ key, contentType })` (public `authAction`).
2. `fetch(url, { method: "PUT", body: file, headers: { "Content-Type": contentType } })`.
3. Store `key`; read at `${R2_PUBLIC_URL}/${key}` (or `internalGetDownloadUrl` for a signed URL).
Internal helpers: `internalUploadBuffer`, `internalGetDownloadUrl`, `internalDeleteObject`.
</usage>
<failure_modes>
- **Uploads 403 / SignatureDoesNotMatch** → wrong `R2_ENDPOINT` (must be the account S3 endpoint, not the public URL) or the API token lacks write on the bucket.
- **Files upload but URLs 404** → public access not enabled, or `R2_PUBLIC_URL` is wrong (it must be the public/custom-domain base, NOT the S3 endpoint).
- **Works in dev, fails in prod** → the five `R2_*` vars were not set with `--prod`.
- **CORS errors on browser PUT** → add a CORS policy to the bucket (Settings → CORS) allowing your web origin and the `PUT` method.
- **`R2_*` not set** → file features fail; `npm run check-setup` flags them as warnings (fine if you don't use uploads yet).
</failure_modes>
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| writing-skillsobra/superpowersUse when creating new skills, editing existing skills, or verifying skills work before deployment | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| xlsxanthropics/skillsUse this skill any time a spreadsheet file is the primary input or output. This means any task where the user wants to: open, read, edit, or fix an existing .xlsx, .xlsm, .xltx, .csv, or .tsv file (e.g., adding columns, computing formulas, formatting, charting, cleaning messy data); create a new spreadsheet from scratch or from other data sources; or convert between tabular file formats. Trigger especially when the user references a spreadsheet file by name or path — even casually (like "the xlsx in my downloads") — and wants something done to it or produced from it. Also trigger for cleaning or restructuring messy tabular data files (malformed rows, misplaced headers, junk data) into proper spreadsheets. The deliverable must be a spreadsheet file. Do NOT trigger when the primary deliverable is a Word document, HTML report, standalone Python script, database pipeline, or Google Sheets API integration, even if tabular data is involved. | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |
| shipping-and-launchaddyosmani/agent-skillsPrepares production launches. Use when preparing to deploy to production, or when asking what needs to be in place before shipping. Use when you need a pre-launch checklist, when setting up monitoring, when planning a staged rollout, or when you need a rollback strategy. | Stars: 119k | Updated | Claude CodeCodex | Verified | Install |
| childhoodthedotmack/claude-memBabysit a shipped release after merge, publication, or deployment. Verify published packages, production behavior, installed plugins, and late CI or user reports until the observation window is clean; use babysit for PR readiness before release. | Stars: 118k | Updated | Claude CodeCodex | Verified | Install |