com.llm-bus/llm-bus

Use llm-bus to design a multi-agent development workflow: a requirements agent breaks down tasks, a coding agent claims implementation work, a testing agent claims validation work, all states are written to a shared ledger, and context is passed through handoffs.

A clear multi-agent collaboration plan with task assignment, state tracking, context handoffs, and execution order.

Explain how to use atomic claims and leases in llm-bus so multiple AI agents do not claim the same task, and provide a strategy for reclaiming timed-out tasks.

A concurrency-control plan describing exclusive task claims, lease renewal, timeout recovery, and failure handling.

Define a shared-ledger structure in llm-bus for a project using multiple AI agents, recording tasks, owners, status changes, handoff logs, and final outputs, and suggest the key fields.

A suggested ledger data structure for multi-agent collaboration that tracks progress, ownership, and handoff history.