Your code says
it’s finished.
Let’s verify.
Deterministic checks for the patterns AI-generated code uses to perform doneness. Runs entirely in your browser.
Scan code
Bring your own agent
CHECKING…This page is a WebMCP host. It hands an agent the same doors the buttons use — so an agent can scan a repository, read the receipts and explain a rule, while the verdict on each finding stays yours. Nothing runs in a hidden channel: a tool-driven scan fills the visible input, flips the visible tab, and moves the real card in the deck.
What counts as slop
Twelve deterministic checks for code that performs doneness — the comment that describes the validation instead of doing it, the catch that files the error somewhere quiet, the config that moved in and started receiving mail. Every hit is a real line, quoted, with a weight you can see.
Score is weighted hits per thousand lines, so a big repository is not punished for being big. A single habit repeated forty times is charged once and then damped — one stylistic decision should not fail a codebase on its own.
It cannot tell you who wrote the code. Nothing in source text reliably identifies an author, and anything claiming otherwise is guessing. What this measures is whether a codebase carries patterns that AI-assisted code carries more often. A human can write every one of them, and frequently does. Read the grade as a description of the code, never as an accusation about the person.