As hokage mentioned the blueprint, I added a simple explanation of how I understand this. The hierarchy of ROMA on @SentientAGI shows how open, coordinated intelligence actually works. At the base is the 𝑨𝒕𝒐𝒎𝒊𝒛𝒆𝒓, it takes a complex goal and breaks it into smaller, actionable sub-tasks. This is where big problems become manageable. Next is the 𝑷𝒍𝒂𝒏𝒏𝒆𝒓, which builds the roadmap. It connects those smaller tasks, defines priorities, and decides how they’ll be executed efficiently. Then comes the 𝑬𝒙𝒆𝒄𝒖𝒕𝒐𝒓, the one that gets things done. It performs each task or delegates to specialized sub-agents, turning strategy into action. Finally, the 𝑨𝒈𝒈𝒓𝒆𝒈𝒂𝒕𝒐𝒓 collects and refines all outputs, evaluating results and feeding them back into the system for continuous improvement. Together, the flow — Atomizer → Planner → Executor → Aggregator — forms a recursive loop that lets agents reason, act, and evolve. That’s how ROMA turns coordination into intelligence.