GitHub App · Contents: read · Issues: write
iterative reads your whole repository, finds bugs that are actually there, and opens one issue for each one: path, line, evidence, and the fix. It never writes code and it never opens a pull request.
41const repo = repos.find((r) => r.id === id);42return repo.fullName; <— BUG 🔴 undefinedfind() returns undefined for an unknown id; line 42 reads it.
42if (!repo) return null; <— FIX 🟢 guard43return repo.fullName;How it works
What it finds
A value that can be missing is used as if it is always there.
A loop, slice or limit that is wrong by exactly one.
Two parts of the code that must agree, and do not.
A failure that is dropped, swallowed or treated as success.
Concurrent work that can interleave into a wrong state.
A handle, timer or listener that is never released.
Pricing
Up to 750k characters of code. Your first scan is free.
The boundary
Two permissions, and neither of them can change a line of your code.
FAQ
Last step
Pick one repository and read the issues its first scan opens. Your first scan is free.
iterative.sh · one GitHub issue per real bug · never a pull request