I don't want an independent userspace implementation, I want the verifier the kernel uses. I'm was looking into this and I'm not sure why it can't be pulled out. I guess depending on what you're plugging the eBPF into the environment (functions, etc) may be tightly coupled to the kernel, but surely the loop analysis and stuff could?

It basically forces you to program on a live system. No CI, no static analysis, no editor help, etc.

UBPF https://github.com/iovisor/ubpf/ and rbpf https://github.com/qmonnet/rbpf are userspace bpf interpreters.

Via https://github.com/zoidbergwill/awesome-ebpf#user-space-ebpf

Not sure if there's a full user space runtime for BPF progs, but there are lots of user space tools which are useful for working with BPF progs. For example, you can use veristat (https://github.com/libbpf/veristat) to easily see whether a BPF prog would be verified, and collect statistics, etc.