Hashing file contents to replace subjective 'same-origin' guesses in a project audit

A developer reviewing about 55,000 files and 30+ projects flagged three groups of repositories as likely sharing a common origin based on naming and theme similarity. Instead of accepting these impressions, they built a Python tool using only the standard library that computes content fingerprints and file-name overlap across the projects. The measurement showed that a large raw duplicate count of 5,684 reduced to 551 unique content matches under a consistent counting method, changing several suspected pairs into cases with no shared code. The author identified their earlier reasoning as premature pattern matching and committed to running this check before future monthly audits. Planned extensions include AST and import-graph similarity to catch renamed but logically identical modules.