- Openclaw tutorials and step-by-step guides - is there actually anything in the classroom? Seem just a titles and empty content?

Hello, just a basic architecture, mainly how to build workflows so agents are working together and it's not a bluff but real advantage of them breaking tasks between each other. Real life example - I want to type my agent something like "Hey, there is a new openclaw update, can you research and find release notes and let me know what possible benefits it will give us?". I'd want to build a system when this main agent delegates a task to a chain of other agents, they check my current setup, another one searches for issues we had in the past or features I've planned or discussed, writer then present it nicely back to main agent who verifies it and I get a nice output back. Something better than just a generic web_fetch and rephrased mix of words without any real meaning. In general I'm struggling in making my agents smarter, they just seem to talk a lot but changing goalposts every 5 minutes and nothing meaningful comes out of this. I'm a beginner and although I've spent a good few week on this, I have no AI background so perhaps that's why I'm struggling. I'd like agents to think ahead and proactively search for solutions online, skills that we may need, evaluate them, propose... right now it feels just like ChatGPT ;) Thanks!