ruvnet/ruflo

Designed specifically around Anthropic's Claude, this orchestration layer lets you spin up coordinated networks of agents that can tackle complex, multi-step workflows without you wiring together every handoff manually. The core pitch is that instead of building one monolithic Claude prompt, you decompose work across specialized agents that communicate, share context through RAG pipelines, and self-organize based on task demands. Native Claude Code integration means agents can write and execute code as part of their reasoning loop, which is genuinely useful for developer automation use cases. The architecture targets teams that want enterprise patterns like persistent memory, structured agent roles, and audit-friendly workflow coordination, without leaving the Claude ecosystem. That narrow focus is both its appeal and its risk. If your stack is already Claude-heavy, the tight integration removes friction. If you need model flexibility, you are boxed in from day one. The honest reservation is that the repo is early and the documentation does not yet match the ambition of the feature list. Expect rough edges and plan to read source code. -> Best for: Claude-committed founders building internal automation tools or agent-heavy SaaS who want a structured starting point rather than a blank framework.