Can I use the base pack first and add Jev screening later?

Yes. The installer only adds the 12 base skills by default. You add the extension later with --only-jev, or together at first install with --with-jev, and upgrades keep backups without touching the other set.

What setup turns on the optional cnki-jev extension?

You copy the example config into your research directory, set title and abstract send permission, call limits, and budget, and export TYPESAFE_API_KEY. The default config permits zero paid calls and never creates an account for you.

Does Jev write the final include or exclude decision?

Modes are off, shadow, and assist. Multiple conditions per paper merge into one request, first release never writes final conclusions, and exclusions plus uncertain items go back to model or human review with source evidence.

Which CNKI entry does it use when the network reroutes?

It starts from the domestic CNKI entry and falls back to an overseas mirror only after telling you why, tagging which platform produced the results. Do not use overseas mirrors as the start address.

What is in the 12-skill base pack?

Search, advanced search, result parsing and pagination, paper detail, journal browse, download, export, screening, an evidence table, task resume, and a researcher entry point; it does not call Jev.

What does the cnki-jev shadow mode do?

It runs the paid screening alongside the original workflow and compares the two judgements, while assist mode generates paid screening suggestions and off keeps the original workflow unchanged.

What output does a screening run produce?

An Excel workbook with five sheets for the screening table, evidence comparison, source evidence, task progress, and rules, plus CSV files and an audit snapshot backed by a SQLite ledger.

CommunityJev Agents & MCP

Cnki Skills

Skill pack for Codex, Claude Code, and WorkBuddy that drives CNKI topic search, result parsing, detail reading, downloads, citation export, screening tables, and resumable task ledgers.

See highlights

Open github.com

Stars: 21
Language: Python
License: MIT
Published by: Community
Author: longkou1988
Last verified: Oct 2026

longkou1988/cnki-skills

User repository on GitHub

View on GitHub

Experimental CNKI literature workflow skills for Codex and Claude Code, with tested offline BibTeX conversion

Forks: 2
Last pushed: Updated Sep 2026

Repo stats from the GitHub API, cached Oct 2026.

Tagscommunity jevbest catalog-import

Works forJev for Classification

Used inJev for Software & DevTools

Languages: Python

Highlights

Twelve base skills cover keyword and advanced search, paging, detail extraction, journal browse, PDF-first download, and BibTeX export.
Screening and evidence tables record include, exclude, and pending rows with source evidence, plus Excel and CSV output.
Installer targets Codex, Claude, and WorkBuddy paths, refuses to overwrite without --upgrade, and keeps timestamped backups.
Optional cnki-jev extension runs layered pre-screening with caching and budget caps, defaulting to off with no key bundled.

Watch out

CNKI access needs your own authorized browser session, result download links may serve CAJ, and detail pages can show puzzle checks needing manual pass. Copy reviewed against the repository README on 2026-10-04, not from running it.

Questions about this entry

Short answers drawn from the linked project or page.

More Jev resources like this

jev-recruiter

Local LinkedIn recruiting workspace on Browser Use's Jev Ultrafast: Jev screens titles, drives navigation, and assesses profile excerpts against your brief while a live browser stays visible; matches and evidence export as JSON.