sci-journal-selection-skill

SCI Journal Selection · Agent Skill

中文说明

Recommend SCI/SCIE target journals from a manuscript brief: verify CAS and XinRui zones, LetPub IF/reviews, and same-topic precedents by year. Defaults: prefer XinRui for ranking; skip TypeSafe Jev (no API key).

Compatible with Agent Skills on Cursor, Codex, and Claude.

Do / Do not

Defaults (ask once; silent → these)

1. Zones: prefer XinRui — rank by 2026 XinRui 大类 when it conflicts with CAS; still show both columns.
2. Ranking: no API key / skip Jev — heuristic unless the user opts in and TYPESAFE_API_KEY is set.

How it runs (text)

1. Normalize brief (title, abstract, methods, contribution, constraints). Keep an English brief for finders / optional Jev.
2. Preference gate (above).
3. Build a 15–40 journal pool from ≥2 sources (Journal Finders, MJL, LetPub, OpenAlex, …). Offline CAS lookup uses the bundled CSV (see Bundled data).
4. Verify: SCIE, CAS, XinRui (else 新锐未核实), JCR Q, LetPub page fields (secondary). Conflict: Clarivate/MJL > publisher > LetPub.
5. Precedents for final-table journals: same topic/type + year; last 3y > 4–7y > older/none. No hits → 近作先例：待核验 (never invent).
6. Rank: heuristic (scope → precedent recency → zone preference → metrics → risk) or optional Jev (scope_fit / precedent_fit / zone_fit …). Scores ≠ acceptance odds.
7. Deliver paste-ready table + no-acceptance line. No submission by this skill.

Optional illustrations

Pixar-style sketches (illustration-style-samples); decorative only.

Bundled data

skills/sci-journal-selection/data/FQBJCR2025-UTF8.csv — CAS 2025 upgrade table (UTF-8).

XinRui is not fully bundled — look up live or mark 新锐未核实. See skills/sci-journal-selection/data/README.md.

Install (Cursor / Codex / Claude)

SKILL_SRC="$(pwd)/skills/sci-journal-selection"
# Codex / Agent Skills
mkdir -p .agents/skills && ln -sfn "$SKILL_SRC" .agents/skills/sci-journal-selection
# Cursor
mkdir -p .cursor/skills && ln -sfn "$SKILL_SRC" .cursor/skills/sci-journal-selection
# Claude
mkdir -p .claude/skills && ln -sfn "$SKILL_SRC" .claude/skills/sci-journal-selection
# optional: ./install.sh --project .   or   ./install.sh --user

Environment (optional)

Only if the user chooses 有apikey / use Jev:

export TYPESAFE_API_KEY="your-key"   # placeholder — never commit a real key

Details: skills/sci-journal-selection/references/typesafe-jev.md. .gitignore ignores .env.

Quick start

1. Install (symlink above).
2. Chat: Use $sci-journal-selection / “按选刊 skill 推荐期刊”.
3. Answer the two preference questions (or accept defaults).
4. Paste title + abstract.
5. Expect CAS / XinRui / LetPub / precedent years + the no-acceptance disclaimer.

python3 skills/sci-journal-selection/scripts/openalex_precedents.py \
  --issn 0888-3270 --query "bearing PHM" --years 3 --mailto you@example.com

Layout

sci-journal-selection-skill/
├── README.md / README.zh-CN.md
├── LICENSE · install.sh · .env.example
├── examples/acm-phm-brief.md
└── skills/sci-journal-selection/
    ├── SKILL.md
    ├── data/FQBJCR2025-UTF8.csv
    ├── assets/selection-pipeline-pixar{,-zh}.jpg
    ├── references/ · scripts/ · agents/

Spec & version

• Spec: https://agentskills.io/specification
• Skill version: 1.2.0 (SKILL.md metadata)

Disclaimer

Academic reference only. Not a submission promise or acceptance forecast. Zones/IF follow official or clearly labeled secondary sources; LetPub is third-party page display. Authors must follow their unit’s rules.