I’ve started asking Claude Design for multiple directions on a canvas, then sending the promising ones through another pass before I choose. In a conversation between Michael Riddering and designer/builder Kyle Zantos on Dive Club, Zantos builds the missing part of that workflow: an interface for reviewing what the agents made.
Zantos uses a standalone HTML artifact to render the proposals and collect decisions in the same place:
You act as the orchestrator and the quality-control reviewer and send out Opus 4.8 sub agents to execute these audits as each skill does it. Compile the results, but at the end of it, compile it all, remove the duplicates, and then present me an HTML doc that has all the proposed changes and then options for me to approve, deny, or discuss them, and then have a way at the bottom to copy all of what I put in here and just paste it right back into my session with Claude.
Zantos on combining decisions across different parts of a product:
Then be like, “Oh, no, not that. That’s terrible. I hate that.” Or just like, “Yes, but add this other thing.” Or, “I really like what you did in number two on number six that’s in a totally different area of the app or the site or whatever that I’m working on, but grab the technique that we used in the second point of this doc.” So just visualizing everything from your agents has been huge.
