What problem does it solve? Duplication audits often produce vague, unverifiable findings that teams cannot act on. This Skill runs an independent detection pass over a defined scope against a named checklist and returns only findings backed by a runnable verifier command and a concrete drift hypothesis. ## Core Features & Use Cases - Checklist-driven detection: Scans a directory, glob, or file list against 6-10 named duplication patterns, capping findings per pattern to avoid noise. - Verifier-burden enforcement: Every finding must include a runnable command (grep, rg, find) that re-confirms the pair; findings whose verifiers fail are dropped. - Drift-hypothesis discipline: Each finding states the concrete consequence of continued drift, rejecting vague claims like "would be cleaner to share". - Use Case: During a drydry audit of a large repository, the orchestrator routes to this Skill to walk the scope, downsample oversized file sets by language, and return structured markdown findings the orchestrator assembles into the audit artefact. ## Quick Start Ask the agent to run the drydry audit pipeline over a target directory so the sweep detection pass returns verified duplication findings.
