It helps answer questions like: where is this file used, what are the entry points, are there circular dependencies, and which files or node_modules are unused.
I recently rewrote it from scratch in Go to focus on performance and low memory usage. On large codebases (500k+ LOC), circular dependency detection runs in a few hundred milliseconds and is up to 175x faster than Madge in my benchmarks.
Feature-wise it overlaps partially with tools like madge, dpdm, skott, and dependency-cruiser, but aims to provide fast, list-based, actionable output rather than graphs.
Feedback and performance comparisons on other large codebases are very welcome.