Extract structured metadata from scholarly reference lists and generate a clean literature manifest for manual downloading or later curation. Use when the task is to parse a references section, normalize citations, recover fields like title/year/author/DOI, enrich entries with Crossref or OpenAlex hints, expose candidate article URLs that can directly open the paper landing page, detect likely duplicates, compare against an existing local PDF folder, and write `ref_manifest.json` / `ref_manifest.md` / `ref_manifest.csv`.
Use this skill when the useful output is a structured worklist of papers.
.md or .txt.scripts/normalize_reference.py.scripts/extract_ref_manifest.py.ref_manifest.json for downstream scripts.ref_manifest.md for quick human audit.ref_manifest.csv for filtering, sorting, and manual literature tracking.A good row should include, when recoverable:
scripts/normalize_reference.py — normalize one citation string into a small JSON record.scripts/extract_ref_manifest.py — parse a references file and generate ref_manifest.json + ref_manifest.md + ref_manifest.csv.Report only: