Experimental support for dynamic container engines and task runtimes per process

[abhi18av]

This internal PR highlights the tweaks we need to make sure that nf-nomad is able to utilize multiple task drivers within the same pipeline.

This is very experimental therefore, I'd appreciate if @jagedn can take a look and take this forward whenever he is available next month.

[abhi18av]

Meeting notes (07-July-2026)

1. We should have a more generic approach to see which other executors benefit from the hybrid driver approach

2. Kubernetes also benefits from this since it allows containers + virtual-machines (KubeVirt) + WASM

References

• https://dev.to/snrizvi/run-vms-alongside-containers-in-a-kubernetes-custer-with-kubevirt-20g9
• https://www.spinkube.dev
• https://github.com/containers/crun/blob/main/docs/wasm-wasi-on-kubernetes.md

========

Webassembly and Bioinformatics

• https://biowasm.com
• https://www.cncf.io/blog/2024/03/12/webassembly-on-kubernetes-from-containers-to-wasm-part-01/
• https://link.springer.com/article/10.1186/s12859-026-06431-1
• https://biowasm.com
• https://genomicx.org

=========

We can also create a wasm driver for Nomad

[jagedn]

I was playing a little with wasm and I think it will require a lot of changes

as nextflow use bash and files we need to create (or improve current) k8sexecutor with lot of changes so I'll not continue with the wasm idea at least by the moment

[jagedn]

I've changed the title of the PR as the intent is not to have a multi-executor but allow to the current executor be aware of every task to be executed