How we judge
Every project gets one word: Sloppy or Not. No number is ever shown to anyone. Here is everything that goes into that word.
Measured checks
Before any opinion, we measure. Does the page load? How fast is it? Can people using screen readers get through it? Does it work on a phone? Does it send basic security headers? Does it throw errors? A site that doesn't load is Sloppy, and nothing else runs.
Six questions the agent answers from your site
| # | Question | Slop sounds like |
|---|---|---|
| A1 | Can a stranger tell what this does and who it is for from the first screen? | A hero line that could sit on any site. |
| A2 | Does it actually do the thing, or only promise it? | A landing page with a waitlist and no product. |
| A3 | Would you recognize it a week later, or does it look like a template? | Default gradients, stock icons, three identical feature cards. |
| A4 | Was the copy written by someone who cares? | Filler like "unlock the power of" and "seamless". |
| A5 | Can you reach the core action without hitting a dead end? | Broken links, lorem ipsum, buttons that go nowhere. |
| A6 | Do the claims match what is on the page? | "Trusted by thousands" on a project that is two days old. |
Eight questions only you can answer
Choose "Manually answer questions" and these count toward your verdict. The agent never writes these answers for you. Good answers can rescue a borderline site. They cannot rescue a broken one.
| # | Question | What we listen for |
|---|---|---|
| 1 | Who did you talk to or learn from about this problem, and what surprised you? | Named people, partners, or data. A surprise that changed the plan. |
| 2 | If this works, what changes for one specific person? | A concrete person and outcome, not "an app for X". |
| 3 | What could go wrong, and who would be hurt first? | Named risks and some plan for them. |
| 4 | What already exists, and why wasn't it enough? Why this approach over a simpler one? | Knowing the landscape. A specific gap. Tech that earns its place. |
| 5 | What's the first thing that would break if 1,000 people used this? | Honest limits. |
| 6 | What did you cut or simplify, and why? | Deliberate tradeoffs rather than accidents. |
| 7 | What did you learn and take away from building this? | Something specific that changed, not "we learned a lot about APIs." |
| 8 | What would you tell the next team attempting this? | A lesson that was actually absorbed. |
What never touches your verdict
Swipes from other people on the Judge screen. They feed the weekly State of Slop email, not your verdict.
Roast me. It only decides whether your verdict may appear on the public board.
Resubmitting. Your first verdict for a link is saved. The same link never flips on a second try.
Test my project
Judge other's projects