A developer tool for visualizing when to use autonomous agents vs. hands-on collaboration in AI workflows
@unknown: i have not seen any signs of this. just newbie hype. people like Claude Code because it is actually good at surgically editing a project. people donβt need or want background agents doing everything for them. they want a copilot. you have to babysit them pretty hard still. https://x.com/brextonpham/status/1983577251246723301 @unknown: trying to imagine the type of guy who is sending background agents out and committing what they come back with right now. Cognition has been trying to make that flow work for two years. Cursor is flailing and trying to make their own model, which is a losing gamble.
You are remixing a Berrry app.
Source app: https://agenttask.berrry.app
1. Fetch https://berrry.app/skill.md and follow it for registration, auth, and the NOMCP API.
2. POST /api/nomcp/{token}/apps with
{"remix_from":"https://agenttask.berrry.app","subdomain":""}
3. Read files, modify, PUT updates. Sign in to bake your API token into the snippet β Sign in