With the current improvements in agentic AI we're seeing the benefits in "busy work" like tracking build failures in the package addition or rebuilds of the RoboStack distros.
The holy grail would be that those agents then run unattended and would be triggered automatically. This is however expensive to setup as you can't use your basic token for these kind of automated jobs.
One of the ideas is to create a really good set of skills that agents can use to do this work. Lots of this knowledge is hidden within the brains of the existing contributors. These skills should then be added to the distro repositories and can be used by anyone contributing to the repo.
We could even start a RoboStack/skills repo with these skills listed and easily modifiable.
With the current improvements in agentic AI we're seeing the benefits in "busy work" like tracking build failures in the package addition or rebuilds of the RoboStack distros.
The holy grail would be that those agents then run unattended and would be triggered automatically. This is however expensive to setup as you can't use your basic token for these kind of automated jobs.
One of the ideas is to create a really good set of skills that agents can use to do this work. Lots of this knowledge is hidden within the brains of the existing contributors. These skills should then be added to the distro repositories and can be used by anyone contributing to the repo.
We could even start a
RoboStack/skillsrepo with these skills listed and easily modifiable.