Cron jobs at (well organized companies) are often monitored so when they fail someone is alerted.
Yeah, I've got tons of esoteric Unix minutiae in my head but the full crontab syntax somehow feels unknowable.
I guess it is a big deal.
That's awesome. I love the cron pattern.
I was just recently explaining to someone non-technical how most of what we think of as "always-on" with computers is actually implemented with some form of a sleep/wake pattern.
The core feature of systems like OpenClaw is that they make it possible to run full coding agents asynchronously in the background with sufficient access to actually do things for you.
For all the other features and the hype, we're basically just talking about what I'm calling 'agentic cron jobs'.
Some of those issues can be addressed by your system for running claws and agents are amazingly good at helping with the rest.
But it's still true that that people with a technical background need to be involved for the most complicated/security sensitive claws.
Sad to say, decades later, I'm *sure* that the clever crontab syntax is not the right long-term solution to the problem.
It's just too hard for humans to parse even with tons of experience.
The non-standard attempts at adding "@hourly" and /etc/cron.hourly did solve this problem to a large extent.
Even with agents, I still insist on reviewing/understanding all my code so I want something I can parse.
But after years of reading/writing regular expressions it became increasingly easy and productive, in a way crontab syntax never did despite being much simpler.
This is an open source effort, not in any way proprietary, and completely open to contributions from any organization.
