- Home
- Skills
- DevOps & Cloud
- No-code to Cloudflare Migration Planner
More screenshots
Works with the AI Agents you already use
No-code to Cloudflare Migration Planner
This skill takes a structured app-spec.json and gives you a kit to build from.
$8.99
No-code to Cloudflare Migration Planner
Example session with this skill installed
Use the nocode-to-cloudflare-planner skill on ./app-spec.json and the records in ./export. Write the kit to ./plan.
- Read your context and instructions
- Compiled the no-code to cloudflare
Summary
Read app-spec.json (spec 0.1): 6 entities, 4 option sets, 8 privacy rules, 11 pages. Wrote to ./plan: 0001_schema.sql (12 tables, loaded into SQLite to confirm it runs), policy.ts, routes.md, redirects.json (24 rules) with test-urls.sh, and generator-report.md. The loader turned the export into ordered 0002_NN_data_*.sql files.
Access policy
All 8 non-default privacy rules translated, 0 MANUAL. Drivers read their own check rows, admin and workshop read all, everyone else is denied. Writes are deny-by-default and listed for handler decisions.
Load report
- user: 1 row
- vehicle: 1 row
- check: 1 row, 1 orphan reference set to NULL (check.vehicle points to a vehicle that is not in the export)
- 3 entities had no export file and were skipped
Decisions needed
- Money-like and number fields: keep as REAL or change to cents
- Magic link or reset email for the sign-in announcement
- Cutover hour
- Who may write to vehicles and checklist items (the audit says several actions rely only on a page redirect)
- 6 redirect items need review (slug URLs and dynamic routes)
Next steps
- Answer the decisions list.
- Apply 0001_schema.sql to a staging D1 database.
- Load the data files in order.
- Build the write handlers from the checklist.
- Parallel run on a subdomain, then cutover.
Connects securely to your tools. The creator never sees your data.
About this skill
Moving a no-code app to Cloudflare goes wrong in the same places every time: lost constraints, orphaned references, access rules that only existed as page redirects, broken old URLs, and users who cannot log in on day one. This skill takes a structured app-spec.json and gives you a kit to build from.
What you get
- D1 (SQLite) schema SQL generated from your spec and tested in SQLite: source ids kept as primary keys, join tables for lists, option set lookup tables, file key and legacy URL columns, indexes for delta syncs.
- An access policy skeleton (policy.ts) translated from your privacy rules, default deny. Rules it cannot translate are listed as MANUAL with the original rule text.
- A redirect map (redirects.json) and a curl test script, built from your pages.
- A data loader: put your exported records in a folder and get ordered, re-runnable INSERT files. Orphan references become NULL and are listed in a report. Nothing is dropped silently.
- A route list, write-handler checklist, file and user move plan, and a cutover and rollback plan with the decisions the owner must make.
Where the spec comes from Bubble apps: the free bubble-app-audit skill writes app-spec.json. Lovable, Bolt, Base44, Replit and other apps: a guide in the package shows how to write the spec from a schema export. This route is less proven than the Bubble one.
What it does not do It does not connect to any platform or Cloudflare, download files, move records over the network or change DNS. The scripts read local files only. The audit does not read UI conditionals, and privacy rules cover reading only, so write permissions are marked for manual decisions. Apps above roughly 500,000 records or 50 GB of files need a different approach.
Includes a worked example from a real six-table app and a synthetic sample export. Platform names are trademarks of their owners; this skill is not affiliated with them.
Does not connect to any platform, download files or change DNS. UI conditional visibility is not covered and privacy rules are read-only, so write permissions need manual decisions. Apps over about 500k records or 50 GB need a different approach. Non-Bubble specs must be written by hand or by your agent.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed today
Requires