I’m particularly interested in the disable-model-invocation: true safety on /exec and /cleanup. It addresses the biggest hurdle for AI in infra: the fear of an agent hallucinating a delete or exec command in the wrong context.
In addition the skills teach the agent about which kubectl commands are mutating so it doesn’t treat them and normal Bash() commands and asks you for permission first.
I wonder whether the future of software is like this project, where computer capabilities come in the form of skills that you can (purchase?/rent? to) download and use.
At first, it's really fun to make your own bases. Eventually, it seems you either lose interest, or you get impatient and copy in others' blueprints. A blueprint is just a plain-ish text representation of a part of a factory, and late game (or creative mode), the robots just build it. All you do is copy and paste it, not much creativity required.
At first, it's really fun to make your own bases. Eventually, it seems you either lose interest, or you get impatient and copy in others' blueprints. A blueprint is just a plain-ish text representation of a part of a factory, and late game (or creative mode), the robots just build it. All you do is copy and paste it, not much creativity required.