I built ZigZag, a command-line tool written in Zig that recursively scans source code directories and generates a single markdown report containing the code and metadata.

It’s designed to be fast on large codebases and uses: - Parallel directory and file processing - A persistent on-disk cache to avoid re-reading unchanged files - Different file reading strategies based on file size (read vs mmap) - Timezone-aware timestamps in reports

Each directory produces a report.md with a table of contents, syntax-highlighted code blocks, file sizes, modification times, and detected language.

Repo: https://github.com/LegationPro/zigzag

I built this mainly for auditing and documenting large repositories. Feedback, critiques, and ideas are welcome.