AgriciDaniel/ace-claude-code ? reverse-engineered prompt

Build me a Claude Code helper called Ace that makes it easy to run bigger projects with multiple agents working at once.

I want it to act like a simple orchestration layer where I can start a project, create a roadmap, plan each phase, and then have the work happen in parallel with clear handoffs and saved progress between sessions. It should support natural language too, so I can just say things like “I want to build a dashboard” or “resume where we left off” and it figures out the right next step.

Please set it up so it can manage project planning files, session state, progress checks, and verification steps, and make it feel smooth for both brand new projects and existing codebases. Include sensible defaults, a config file I can edit, and a clear command style for the main workflow. If anything depends on current Claude Code behavior, look up the latest docs online first.