ruflo: An agent meta-harness that coordinates specialized agents, workflows, memory, and swarms around Claude Code and Codex.
What problem it solves
Models need more than prompts to carry out substantial development work: they require tools, persistent context, task coordination, controls, and a way to inspect results. Ruflo provides this execution layer so multiple agents can divide work, retain useful experience, and operate within defined permissions.
Key features
Coordinated agent swarms — Route work among specialized agents and organize them into swarms that collaborate on tasks. The console exposes agent activity and workflows for inspection.
Source: README
Persistent learning and memory — Carry context between sessions, learn from successful task patterns, and use memory to inform future routing. Local vector retrieval and contrastive updates are described as part of the learning loop.
Source: README
Multiple client integration paths — Initialize a project for Claude Code, connect Codex and other MCP clients, or run the terminal interface directly. The README documents both plugin-only and full CLI installation paths.
Source: README
Federated collaboration — Allow agents on different machines to collaborate through federation and other communication plugins. The project describes federation as communicating securely without leaking data.
Source: README
Inspectable and controllable execution — Review logs and results, search for failures, replay runs, and compare activity. Claude control provides permission levels and approval modes, while network, spending, and destructive actions require confirmation.
Source: README
Quick start
Copied verbatim from the project's README.
npx ruflo@latest initUse a ruflo swarm to review this repo: find the riskiest module, write tests for it, and summarise what you changed.Who it may suit
This may suit developers using Claude Code, Codex, or other MCP clients who want agents to coordinate multi-step work while retaining memory and visibility into execution.
Topics to discuss
Which task types benefit most from multi-agent swarm coordination?
How should harness permissions balance automation with user control?
What evidence should teams require before trusting persistent agent memory and learned patterns?
Written by ReadmeX AI from the project's README, source and public GitHub data; it may contain mistakes.
Read the original README https://github.com/ruvnet/ruflo#readme
ReadmeX opened this community for a trending project; it is not run by the project. Maintainers can claim it from the community page.
Comments
…