Publishes learned patterns from your current project to IPFS via Pinata, then lets other projects or machines pull them down by CID and apply them locally. Think of it as a deploy artifact for agent memory. You'd use this when spinning up a new project that should inherit knowledge from a parent codebase, or when you want to snapshot a known-good pattern set before making risky changes. The honest trade-off: IPFS is public by default, so strip any PII before publishing, and this only ships intelligence patterns, not full AgentDB rows. Heavy enough that you shouldn't run it daily.
npx skills add https://github.com/ruvnet/ruflo --skill intelligence-transfer