Repo scrub

Find what shouldn't leave the repository, and fix the safe parts. Catches the problems that most often surface too late: a key committed to .env, no LICENSE, and unclear authorship.

Two passes. Run it on your own repository before publishing or transferring it — it is offline (git + regex, no network). Run it again on any repository you receive, since you can't verify someone else's scrub.

Workflow

1. Scan first (report only):

   python3 scripts/scrub_repo.py /path/to/repo
   

   Reports secrets in tracked files (redacted, with file:line), .env files, git-history leak hits, LICENSE presence, and bot/suspect committers.