| name | duplication-audit |
| version | 0.8.0 |
| description | Deterministic, advisory copy-paste clone audit for Python. Uses jscpd to detect duplicated token sequences and emits EXTRACT (cross-file) / MERGE (same-file) findings to the shared code-health finding schema. Never mutates source.
|
duplication-audit
Overview
Reports duplicated code blocks as advisory findings. It does not refactor.
Quick Start
python3 scripts/duplication_audit.py \
--root /path/to/repo \
--source-prefix src/pkg/ \
--out-dir /tmp/duplication
Output
duplication_findings.json — sorted findings (shared schema).duplication_report.md — summary grouped by signal.
Exit Codes
0 clean, 1 advisory findings present, 2 tool/config error (e.g. jscpd/node missing).
Tools and Thresholds
jscpd via npx. Override with --config thresholds.json (min_tokens,
min_lines); see skills/duplication-audit/references/rubric.md.
Notes
--source-prefix is repeatable and filters product code.- Findings are deterministic: identical input yields byte-identical
duplication_findings.json.
- One finding per clone pair (cross-file → EXTRACT, same-file → MERGE).
Limits
- Advisory only — emits EXTRACT/MERGE findings and never refactors or mutates
source.
- Requires
jscpd (run via npx) and a working Node.js toolchain; a missing
tool is a config error (exit 2), not a finding.
- Token-sequence based: detects copy-paste clones above
min_tokens/min_lines
thresholds and will miss structurally similar but tokenwise-different code.
- Tuned for the Python source surface; non-product code is filtered via
--source-prefix.
- Deterministic and offline; no network access.
