Developed and integrated a CPU-bound matrix multiplication microbenchmark for the nanvix/nanvix repository, focusing on expanding the project's benchmarking capabilities. The microbenchmark was implemented in Rust using a cache-oblivious algorithm to better reflect realistic performance characteristics and support accurate baseline measurements. Parameters for the benchmark are configurable through environment variables, allowing flexible performance testing across different environments. Integration with the existing build system was achieved by updating both Cargo.toml and Makefile configurations, ensuring reproducibility and ease of use. This work leveraged skills in benchmarking, build systems, and systems programming, providing a foundation for regression tracking of matrix-multiplication workloads.