It's a pipeline engine for Kubernetes[1]. The idea is to build a tool that can be flexible enough for CI/CD, but also used for other things such as for example data processing.
The pipelines are just designed as serial and parallel execution of pods with no fancy features (i.e. loops or conditions) and the idea is to let other tools generate more complex pipelines using code.
Responsibility of triggering the pipelines is decoupled to standalone controllers such as github-screener[2] which enables triggering of pipelines used for CI/CD.
The project currently builds itself, but the docs are a bit outdated and there's quite a lot progress to be made.
[1] https://github.com/kuberik/engine