Reads the diff
Questions are tied to the changed lines in the PR, not generic trivia.
SlopBlock posts a short quiz based on the actual diff. If the pull request matters, the author has to pass it before the merge goes through.
Your second pair of eyes
Moving fast is fine. Blindly merging is the problem. When you're deep in a feature, especially with AI in the loop, it's easy to ship code you only half read. SlopBlock slows you down just long enough to prove you actually know what changed.
The questions come from the diff, not from a generic list of best practices. They tend to hit the parts you'd otherwise skim.
If Copilot or Cursor wrote half the PR, you should still be able to explain it. This gives you a quick gut check before it lands.
Docs, renames, and low-risk changes pass through automatically. The quiz shows up when a PR looks worth slowing down for.
Enforce understanding at merge
Code review can catch bugs. It does not always catch the author nodding along to code they barely understand. SlopBlock adds one more check before merge: can the person who opened this PR explain what changed and where it might break?
If AI wrote part of the pull request, the author still has to own it. Passing the quiz is a simple way to prove that.
Routine changes can get a lighter touch. Auth, payments, and infra can get stricter checks. You control that per repo.
SlopBlock runs as a status check. Reviewers still review. This just makes sure the author came prepared.
How it works
Questions are tied to the changed lines in the PR, not generic trivia.
Docs, renames, and trivial changes pass through without extra noise.
The author gets a short multiple-choice quiz about behavior, risk, and implementation details in the PR.
Pass the quiz and the status check goes green. Miss a question and retry based on your team's policy.
For solo builders and teams