| name | cnki-collect-details-codex |
| description | Use when Codex needs CNKI search results enriched in batch with abstracts, keywords, fund information, or thesis detail metadata through an existing Chrome CDP session. |
CNKI Collect Details (Codex)
Use this skill when plain result rows are not enough and the user wants detail-page fields returned in batch.
Run
Normal literature:
python3 _shared/cnki/cli.py collect-details \
--query "干细胞治疗关节炎" \
--count 10
Force serial detail collection:
python3 _shared/cnki/cli.py collect-details \
--query "干细胞治疗关节炎" \
--count 10 \
--concurrency-mode serial
Doctoral theses only:
python3 _shared/cnki/cli.py collect-details \
--query "干细胞治疗关节炎" \
--scope theses \
--degree doctoral \
--count 10
Notes
- Normal literature mode always resets CNKI to
总库 + 中文 before searching.
- Thesis mode reuses dissertation search and supports
both, doctoral, and master.
- Adaptive detail collection is the default. It starts conservatively, increases concurrency only after repeated healthy detail fetches, and falls back to single-page mode when CNKI shows risk signals.
- Optional tuning flags are
--concurrency-mode, --max-concurrency, --min-delay-ms, and --max-delay-ms.
- Each returned item keeps the result-row fields and adds
abstract, keywords, fund, classification, affiliations, detailAuthors, pubInfo, and nested detail.
- If a specific detail page fails while batch collection continues, the item includes
detailError.
