Run custom command between benchmark

[wcampbell0x2a]

I use my benchmarks to test the binaries my code creates, but to do this and test two different versions like this action allows I need to run a command before benchmarking the new branch.

See https://github.com/wcampbell0x2a/backhand/blob/fd4d8fa829de78f023082b8dc27a3ba235872f9a/.github/workflows/benchmark.yml#L15

[s-arash]

I'm also seeing errors related to submodules. I think running custom commands can help either clean up the submodule dir, or to checkout submodules between runs.
Exmaple of failed run: https://github.com/fortanix/rust-mbedtls/actions/runs/12303593121/job/34339051828?pr=373