Hey everyone! I'm exploring ways to utilize Claude Code in non-interactive mode since I'm struggling to fully utilize my MAX plan. I have a couple of questions:
First, can anyone share how they set it up to monitor changes in files within a folder and what actions it can take in response? Additionally, is it possible to have it execute long tasks without needing user confirmations? How would that work? Any advice or use cases you have would be super helpful!
3 Answers
I actually wrote a new provider in Cline that leverages non-interactive mode and uses the --verbose and --output-format json flags. This way, I managed to use Claude Code CLI to make calls more economically with my Max subscription. It works decently well but is a bit slower than using the API. I'm keeping it private for now since I don't want to risk Anthropic taking it down just because I’m using it differently and might potentially share it on GitHub!
I’ve been using a loop with the claude-yolo setting, no safeties, and it’s been quite a ride. Essentially, I tell it to go through my plan.md and keep running itself until it signals completion. It’s amazing how effective it can be overnight, but beware—it can definitely get carried away with scope creep and unexpected refactoring! Can’t wait for the next version to hopefully handle this better!
I was curious about setting up Cron Jobs with Claude Code too. If anyone has experience with that, I’d love to hear how it’s working out for you!
Related Questions
Online Hash Generator - String to Hash Converter
Convert CSV To HTML Table
Convert Json To Xml
Bitrate Converter
JavaScript Multi-line String Builder
GUID Generator