swarm-advanced

The swarm-advanced skill provides a robust framework for managing complex distributed AI agent operations within the Ruflo/Claude-Flow ecosystem. Designed for engineers and researchers, this skill allows users to move beyond single-agent interactions by deploying coordinated swarms that utilize specialized topologies to solve multifaceted problems. Whether you are conducting deep literature reviews, managing full-stack software development lifecycles, or automating quality assurance pipelines, this skill provides the structure required to scale AI capabilities efficiently.

• Swarm Topologies: Implements Mesh for decentralized research, Hierarchical for structured dev workflows, Star for centralized testing and validation, and Ring for sequential processing pipelines.

• Specialized Agent Orchestration: Dynamically spawn and manage agents with custom capabilities, including researchers, analysts, documenters, and developers, orchestrated through adaptive or balanced strategies.

• Workflow Automation: Features powerful task orchestration commands that support parallel execution, sequential dependencies, and cognitive analysis of findings.

• Knowledge Management: Integrates neural pattern recognition and HNSW-indexed vector memory to store, retrieve, and map knowledge across research domains.

• Integration & Compatibility: Fully compatible with Claude Code and MCP (Model Context Protocol), allowing seamless injection of orchestration tools into existing development environments.

• Utilize the swarm_init command to define your topology and maxAgents constraints before spawning your specialized workforce.

• For research projects, leverage the parallel_execute command to gather information from multiple sources simultaneously, followed by cognitive_analyze for data synthesis.

• Ensure tasks are defined with clear strategy parameters—such as adaptive or parallel—to optimize agent load distribution and operational latency.

• Practical constraint: Maintain clear namespace separation in the memory layer to avoid context collisions between different swarm instances.

• Use the CLI fallback options for rapid deployment of common swarm patterns without manually constructing the entire MCP command sequence.