swarm-coordination

Swarm coordination uses file locking and Beads as a source of truth to manage concurrent agent workflows. Agents claim files for editing, document handoff plans, and synchronize state through atomic changes and frequent commits, ensuring clean transitions between parallel work streams.

A handoff is a clean transition of responsibility between agents working on shared tasks. Use handoffs when dividing a feature into parallel streams that must sync on completion, preventing conflicts and ensuring one agent's output becomes the next agent's input safely.

File locking uses Beads as the authoritative source of truth, allowing agents to claim exclusive editing rights to files. Locks prevent concurrent modification conflicts, and agents release locks after atomic changes and commits, enabling the next agent to safely take over.

Yes. Swarm coordination applies to environments where agents manage shared resources and perform concurrent operations. It provides state synchronization patterns, unique session IDs, and explicit bd commands to keep agent state consistent during parallel execution.

Swarm coordination works best when agents operate with Beads as truth source and can perform frequent commits. It requires explicit synchronization via bd commands and may not suit workflows needing real-time streaming or where agents cannot maintain atomic change records.

Testing is listed as a functional requirement for swarm coordination workflows. Ensure tests are in place to validate that handoffs are clean, state synchronization works correctly, and parallel streams integrate properly after agents complete their work.