Claude Code config as infrastructure

Keeping Claude Code skills and settings in sync across devices is tedious. Switching computers only to realize you're missing a skill or preference breaks flow, and I end up spending more time setting things up instead of actually coding.

My solution: use a git repo as the source of truth for my Claude Code config. It uses an install script to symlink everything into my root ~/.claude/ directory, with an option to keep individual skills local.

Skills can be in one of four states—synced, local-only, conflict, or external symlink—and the sync script displays these states in the CLI so it's easy to see which skills live where and make sure each machine is set up correctly.

Repo: github.com/brianlovin/claude-config. I’ve only focused on syncing skills and settings so far, but over time this can expand to include rules and subagents that should sync.

DM me if you have a better approach!