Nice that it's built-in, Claude Code needs an MCP for this at least.
> User Feedback: Intuitively integrate feedback across surfaces and artifacts to guide and refine the agent’s work.
I wish they'd just let me edit the implementation plan directly instead of me having to explain the corrections. Claude Code has the same weakness. Explaining the corrections is slower than editing the plan manually, and it still keeps the incorrect text in context as well.
> An Agent-First Experience: Manage multiple agents at the same time
Sounds nice in theory but I assume you can run multiple agents for 5 minutes or so and then you're out of credits.
As a claude code user I'm not really sold on this product.
Nice that it's built-in, Claude Code needs an MCP for this at least.
> User Feedback: Intuitively integrate feedback across surfaces and artifacts to guide and refine the agent’s work.
I wish they'd just let me edit the implementation plan directly instead of me having to explain the corrections. Claude Code has the same weakness. Explaining the corrections is slower than editing the plan manually, and it still keeps the incorrect text in context as well.
> An Agent-First Experience: Manage multiple agents at the same time
Sounds nice in theory but I assume you can run multiple agents for 5 minutes or so and then you're out of credits.
As a claude code user I'm not really sold on this product.