pruning-stale-and-nonfunctional-material

What problem does it solve? Repositories accumulate stale files, duplicate content, dead paths, and unreachable documents that add noise without changing behavior. This Skill provides a disciplined workflow to reduce that bulk without losing unique functionality, evidence, or backward compatibility. ## Core Features & Use Cases - Structured Decision Workflow: Follows an inventory → classify → preserve-unique → prune-approved → verify-no-loss pipeline with explicit KEEP, MERGE, MOVE_TO_REFERENCE, COMPATIBILITY_STUB, ARCHIVE, DELETE, and UNVERIFIED verdicts. - Decision Matrix Reference: Ships a references/pruning-decision-matrix.md file defining each verdict and the pre-deletion checks (canonical sources, back-references, generators, CI, external links). - Safety Gates: Blocks deletion when usage, consumers, or authority are unverified, and requires re-verification of canonical freshness, links, registry, and tests after pruning. - Use Case: When cleaning up a documentation repository before a release, use this Skill to classify every candidate file, preserve unique evidence, and produce an auditable report of what was removed and why. ## Quick Start Ask the AI to audit this repository for stale, duplicate, or unreachable files and produce a KEEP/MERGE/ARCHIVE/DELETE classification with rollback notes.