hive-mind-advanced

Multi-agent coordination uses a queen-led hierarchy where specialized worker agents collaborate through shared memory and consensus mechanisms. This Skill orchestrates that coordination by implementing Byzantine fault-tolerant decision-making, role-based specialization, and persistent collective memory, enabling agents to align on strategy and execution across distributed tasks like building platforms or microservices architectures.

Byzantine consensus is a fault-tolerant decision protocol that requires a 2/3 majority agreement, allowing multi-agent systems to reach robust decisions even when some agents fail or behave unpredictably. This Skill applies Byzantine consensus to critical approvals in agent swarms, ensuring coordination reliability in large-scale distributed AI tasks without requiring perfect agent behavior.

Collective memory in multi-agent systems provides a shared, persistent knowledge base accessed by all agents through SQLite WAL persistence and LRU caching with memory pressure handling. This Skill implements that system so agents store design decisions, progress, and learnings in real time, avoiding information silos and enabling faster consensus.

Yes. Queen-led hive-mind architecture is designed for large-scale distributed AI tasks requiring persistent memory and role-based specialization. Use cases include building e-commerce platforms, microservices architectures, and other complex systems where a strategic coordinator directs frontend, backend, database, and testing agents in parallel.

Without consensus, agents may diverge on critical decisions, create conflicting outputs, or fail to recover from faulty agent behavior. Byzantine consensus mechanisms prevent these failures by enforcing agreement thresholds, but they require sufficient agent quorum and add decision latency compared to single-agent approaches.

Yes. This Skill depends on claude-flow for underlying agent orchestration. The Skill layers queen-led hierarchy, consensus, and collective memory on top of claude-flow's agent-execution foundation, enabling structured multi-agent workflows with governance and shared state.