The Slop Risk Test

Seven yes/no questions about how your team actually works with AI. No email required, no tracking, no “subscribe to see your results.”

Your Result

See the Slop Prevention Checklist →

1. Before prompting AI to write code, is there a written spec or acceptance criteria?

Yes No

2. Does a human review the architecture AI suggests before implementing it?

Yes No

3. Are acceptance criteria clear before prompting — not invented after the code lands?

Yes No

4. Do AI-generated changes come with tests that someone actually trusts?

Yes No

5. Can a developer on your team explain the architecture of AI-generated code — components, data flow, failure modes — well enough to direct future changes?

Yes No

6. Are prompts, decisions, and tradeoffs captured somewhere durable — not just in chat history?

Yes No

7. Is your repo free of repeated half-solutions to the same problem?

Yes No

Each “no” is one slop point. Lower is better. The scoring lives client-side — your answers never leave this page.

Slopapalooza

A field guide by Doug Kerwin. No code was harmed in the making of this site — allegedly.

Anti-Patterns Risk Test Checklist Horror Stories About Contact