Anyone here ever do Rust-focused benchmarking for LLMs or agents? I'm testing a framework right now but there's not a lot out there re: established benches. Building a framework around the rust cases in multi-SWE-bench, but there's only a few crates tested and I'm looking to find/scrape/create more evals for Rust-focused coding.