distributed-mesh

You can synchronize files across machines using git by materializing remote state locally through a shared state repository. This approach uses git as the transport layer, copying sync scripts into your project root to deterministically pull remote squad files before agents start work.

A git mesh topology defines which squads participate, their zones, and connection details in a mesh.json file. You need it when squads span multiple machines or organizations and local agents require access to remote files to reduce handoffs and context-switching.

Set up cross-machine coordination by generating a mesh.json from your topology, initializing a shared state repository with per-squad directories and placeholder summaries, and running the bundled sync-mesh.sh or sync-mesh.ps1 scripts to synchronize state.

Yes, you need jq, curl, and git installed to run the cross-machine synchronization scripts. These dependencies are required to process the mesh.json topology configuration and execute the state materialization commands.

Git mesh synchronization uses git as the transport layer to materialize remote state locally, differing from other distributed approaches by scaffolding per-squad directories and creating decision inbox entries for team governance rather than relying on real-time network connections.

You should avoid using git as a transport layer for cross-machine coordination when squads require real-time file access or when the overhead of managing a shared state repository and sync scripts outweighs the benefit of local state materialization for your team size.