How production-grade is your team's AI engineering?

25 questions on how consistently your whole team ships with AI — and the fastest way to turn one power-user into a team that ships products, not prototypes.

Vibe-coding gets you a demo fast — but is the output production-grade? See exactly where you stand, and the fastest path to shipping products, not prototypes.

What you get
  • Your score (0–75) and tier — instantly, on the next screen
  • A per-section breakdown of where you’re still prototype-grade
  • A personalized 30/60/90-day playbook + a fix-it guide for every miss
~8 minutesOne tap per questionInstant score

Your score is instant and free — no account needed. Add your email after you finish to unlock your full 30/60/90 playbook and fix-it guides.

Not you? Try another scorecard: Take the AI Developer / Solo Builder ScorecardTake the C-Level / Founder Scorecard
Michał Jaskólski
Created by @jaskol_ski
Creator of wondel.ai Skills — 1.1k+ stars on GitHub

26 years building web apps · 2 IPOs · AI Lecturer at SWPS University · Using Cursor since 2023, Claude Code and Codex since early beta · CTO at TasteRay (built with AI)

Used by engineering teams at

Grupa Morizon-GratkaRingier Axel SpringerTasteRayWondel.ai
Sample report
Your score
42out of 75· 56%
Tier
Intermediate

You ship fast, but context engineering and extensibility are still prototype-grade — that’s what stands between you and production.

Section breakdown
Tooling & subscription78%
Context engineering↓ weakest33%
Extensibility (skills / MCP / hooks)50%
Parallelism & automation40%
Context engineering
Every session starts blind.
→ Author a CLAUDE.md this week, then layer .claude/rules/.
Your 30 / 60 / 90 plan
30 days
  • •Switch to an agent-first tool
  • •Install the Context7 MCP
60 days
  • •Write your first CLAUDE.md
  • •Try Plan mode on a real change
90 days
  • •Author 2–3 custom commands
  • •Add the GitHub MCP