mflow

This repo teaches you to take your Machine Learning project from your laptop to production

mflow syncs local worktrees so AI agents and humans stop overwriting each other. It runs a small daemon per worktree, watches file changes, and relays encrypted updates through a shared room. Without it, parallel agents drift silently until git finds the conflict too late.