Steven Gonsalvez

Software Engineer

The creator of Claude Code stopped prompting his agent 'I don't prompt Claude anymore. I have loops running. They're the ones prompting Claude and figuring out what to do. My job is to write loops.'

Why CEREBRO kept it

Autonomous loops in Claude Code; novel agentic pattern, must-read

The text below is an automated extraction of the article at https://x.com/stacyonchain/status/2070063949761151181, stored verbatim in the public cerebro-vault repository. Copyright remains with the original publisher (x.com).

"I don't prompt Claude anymore. I have loops running. They're the ones prompting Claude and figuring out what to do. My job is to write loops."

> Autonomous loops in Claude Code; novel agentic pattern, must-read

The creator of Claude Code stopped prompting his agent

"I don't prompt Claude anymore. I have loops running. They're the ones prompting Claude and figuring out what to do. My job is to write loops."

This is the full 14-step system to stop prompting and start designing autonomous loops that run until the goal is actually finished. Triggers, execution skills, clear goals, verification, maker-checker, memory, training mode, and the three loop shapes that cover almost everything.

Design the system. Let it run.

Backlinks

Appeared in 1 briefing

Related

Shares tags: ai/agents · ai/llm-mechanics

Also from x.com