recursive-dep-trace-parallel

What problem does it solve?

This Skill provides an exhaustive, line-by-line traversal of every file reachable from a single entry script so users can obtain a complete code and data dependency closure without making ad hoc assumptions or skipping data artifacts.

Core Features & Use Cases

• Breadth-first recursive tracing: Seed with an entry file and discover imports, path literals, subprocess-invoked scripts, glob patterns, and data-file references across the repository.
• Parallel worker dispatch: Dispatch batches of workers to read files line-by-line, report every external reference, and push new items back to the root queue for resolution.
• Data-inspect support: Open and inspect .npz, .pkl, .json, .yaml, and .csv artifacts to extract path-shaped strings and push them into the closure; record unreadable files as HOLES.
• Resumable, auditable state: Maintain artifacts/task_<name>/_state.md and worker_reports/ to resume interrupted runs, append incremental closure.zh.md entries, and produce a final third-party and holes summary.
• Use case: Ideal for security audits, reproducibility checks, migration planning, or any scenario that requires an exhaustive map of both code and data dependencies starting from one entry point.

Quick Start

Point the skill at your repository entry script and request a complete recursive dependency closure and worker report.