Builds a code-review prompt that produces findings with a reproduction instead of a list of naming preferences. Pick the language, the stance, the scope and what to look for; it writes the prompt, including the instruction that forces every finding to come with the input that breaks it.
Without this, the review is guesswork. One or two sentences is enough.
Review prompt
Paste your code at the bottom where the prompt says to. Keep the code last — models weight the end of the context heavily, and you want the code fresh in mind when it answers.
Nothing leaves your browser. Your code never enters this tool — only the prompt around it is built here.
How it works
- Twelve review dimensions, each expanded into what it actually means rather than a one-word label.
- Three stances: ship-blocking only, balanced, or thorough including nits.
- House rules and known issues go in, so you stop getting told about your tab-versus-space choice.
- Every generated prompt requires file, line, failure scenario, severity and a fix as code.
What it does not do
Every tool on this site says this plainly, because a tool that pretends to do more than it does costs you more than it saves.
- A model reviewing code cannot run it, so anything depending on code you did not paste is guesswork.
- It catches local defects and misses whether the change was the right one.
- Treat it as the pass before a human review, not instead of one.
Nothing you type here leaves your browser
There is no account, no server and no logging. The whole tool is JavaScript running on your own machine — close the tab and it is gone. You can check: open your browser’s network tab and use it.
Questions people ask
Because nothing in a default prompt makes a finding costly to invent. Requiring a concrete reproduction removes most of them.Why does AI code review produce so many false positives?
A diff when you are reviewing a change, a file when you are reviewing the code. Tell it which — the review is different.Should I paste a diff or the whole file?
Prompts that go with this
-
Code Review
Reviews that find defects with a reproduction, not a list of naming preferences.
7 prompts
-
Testing & QA
Tests written against the spec, not against the implementation’s bugs.
7 prompts