beads-agent-village

Multi-agent coordination with Beads starts by running bd init to initialize shared memory in your project, then bd doctor to configure git hooks. This establishes a centralized state that multiple agents can read and write to, enabling synchronized workflows across your agent village.

Beads provides shared memory and state synchronization for agent coordination. Adding MCP Agent Mail enables direct inter-agent messaging for real-time communication. Use Beads alone for state tracking; add Agent Mail when agents need to pass messages and coordinate work dynamically.

Yes, Beads supports swarm-style coordination across multiple machines. After initializing with bd init, use bd sync to keep agent work synchronized across distributed systems, and bd ready --json to monitor readiness states across your agent fleet.

Beads tracks work across sessions by maintaining shared memory initialized with bd init. Use bd sync to synchronize updates between sessions and bd cleanup to manage state. This enables agents to resume or coordinate tasks without losing context between runs.

You need Beads installed and a git repository initialized. The core setup requires running bd init to create shared memory, then bd doctor to configure hooks. Optional: install MCP Agent Mail if your agents need direct inter-agent messaging.

Beads handles planning state directly within its shared memory. The metadata describes planning and planning-outside patterns as built-in strategies for agent villages, so you coordinate work through Beads' state without requiring a separate planning tool.