evaluate a Rust rewrite for the shim

Writing the eBPF shim in Rust would give us some benefits:

- smaller runtime overhead for "wrapped" binaries (~5MB for the Go version, 1MB for Rust)
- faster startup
- get rid of some dependencies (LLVM, Clang, bpftool, bpf2go) by using directly Rust to write both the shim and the eBPF code . See https://github.com/aya-rs/aya

Cons:
- ecosystem has less maturity
- steep learning curve

Start by reviewing the existing Go-based eBPF shim and the Aya project linked in the issue. Compare the proposed Rust approach against the stated runtime overhead, startup speed, and dependency reduction goals, while accounting for ecosystem maturity and the learning curve. Done means documenting whether the rewrite is feasible and deciding whether to pursue it.

Written by the indexing model from the issue text.