audit-duplication

Pre-computed context

Current branch: !git branch --show-current 2>/dev/null || echo "unknown"

Purpose

Duplication is the measure most likely to report noise, because some replication is deliberate: a repository that vendors one helper into seventeen plugins on purpose has seventeen copies and zero debt. This skill reports the clone groups a detector found, then subtracts the groups the target repository has already declared about itself, and stops there: no pass or fail, no severity, no finding. What to do about a surviving clone is the reader's call.

Collectors, tried in ladder order per lane:

No collector resolves means the lane's row says so with the tool's install hint and the run continues. This plugin never installs, downloads, or npx-fetches a detector.

Run it

"${CLAUDE_SKILL_DIR}/scripts/audit-duplication.sh"                       # the change: diff from the merge-base plus uncommitted files
"${CLAUDE_SKILL_DIR}/scripts/audit-duplication.sh" src/ lib/            # explicit paths (a missing one is a usage error)
"${CLAUDE_SKILL_DIR}/scripts/audit-duplication.sh" --all                # every tracked or untracked-but-not-ignored file
"${CLAUDE_SKILL_DIR}/scripts/audit-duplication.sh" --registry scripts/cross-plugin-source-registry.txt --all
"${CLAUDE_SKILL_DIR}/scripts/audit-duplication.sh" --json --all src/    # the code-metrics/v2 document instead of markdown
"${CLAUDE_SKILL_DIR}/scripts/audit-duplication.sh" --json --keep --all   # the document, also kept as the next run's trend baseline

Present the markdown report as printed. It opens with the scope and a "Coverage of this run" table (lane, collector, status, reason), then one row per clone group, largest first, listing every instance as file:start-end, then a rollup per lane and per directory, then the summary lines: files with clones, the duplicated-line total, how many groups a registry excluded, and which lanes were partial. Files with clones leaves out any file holding nothing but excluded groups; the scope's count keeps it. When an earlier report of this project with the same scope mode was kept, the last line is the clone trend: this run's class count, the previous run's, and the signed delta, then how many classes are new since that run and how many gained copies (trend in the JSON); the first run in a project prints none. When the report opens with No clone detector ran in any lane, offer the user the install command that headline carries (npm install -g jscpd, or a devDependency) and run it only when they confirm; never install silently and never npx-fetch it. Keep the --json document when the numbers feed a comparison: /verification:measure metrics consumes it when the verification plugin is installed (treat a report whose status is empty on either side as INCONCLUSIVE); otherwise keep the JSON beside your notes and compare by hand.