Claude Code quietly picked up a feature that turns it from an on-demand coding assistant into something closer to an automated background worker. The /loop command lets you run any prompt or slash command on a recurring interval - defaulting to every 10 minutes, but configurable down to whatever cadence you need.
The use cases are straightforward: monitor a deployment and flag errors, re-run a test suite periodically during a long refactor, poll a CI pipeline for status, or keep checking if a PR has been reviewed. Instead of manually re-running commands or context-switching to check dashboards, you set it and let Claude Code handle the repetition.
The syntax is minimal. /loop 5m /your-command runs every five minutes. /loop 30m "check the deploy logs for errors" runs a natural language prompt on a half-hour cycle. It works with any existing slash command or freeform instruction.
This sits in a growing pattern of AI coding tools moving beyond single-shot interactions toward persistent, ongoing assistance. Cursor has background agents. GitHub Copilot has its agent mode. Claude Code's approach is more manual - you explicitly tell it what to watch and how often - but that explicitness is arguably a feature when you want predictable behavior from your tools.
For developers already living inside Claude Code, /loop removes one more reason to leave the terminal.