Currently, personal skills stored in ~/.cursor/skills/ are local-only and don’t follow you across machines. If you work on multiple devices or reinstall Cursor, you lose access to your skills and have to manually copy them over.
Request: Sync personal skills to the Cursor account, similar to how settings/keybindings sync today, so they’re available on any device you’re logged into.
Why it matters:
Skills encode personal workflows and shortcuts that are highly reusable
Manual copying across machines is friction that discourages building skills in the first place
It would make skills a first-class, persistent part of the Cursor experience
You can sync your personal skills across machines today by storing them in a GitHub repository and installing them via the GitHub link on each device. Cursor’s Agent Skills support installation from GitHub repos, so you can maintain one central source and pull it onto any machine.
Native cloud sync for skills (like User Rules have) isn’t available yet, but I’ll pass your request along. Thanks for the suggestion.
I ran into the same problem and ended up building a small extension for it.
Like @Ashutosh_Ranjan, I work across multiple machines and got tired of manually copying ~/.cursor/skills/ (and ~/.agents/skills/ from npx skills installs). Native account-level sync still isn’t there, and I also couldn’t find the GitHub install flow @mohitjain mentioned in my Cursor UI so I went with a GitHub repo + explicit Push/Pull instead.
It’s called SkillBridge: Cursor Skills Sync. You point it at a private GitHub repo, sync your local skills folders with Push/Pull, and on another machine you configure the same repo and Pull to restore everything. It keeps cursor and agents skills in separate remote folders (skills/cursor/ and skills/agents/) so each machine restores the same layout. Auth is a fine-grained PAT scoped to that one repo only.
Still very much early days, but it solved the problem for me. Happy to hear feedback if you try it especially around setup friction or edge cases on multi-machine workflows.
