Find Journal

Find Journal

Use this skill to help researchers choose submission journals from the local Journal Finder project. The skill is Codex-native: do not rely on Claude Code slash commands or Claude-only tools.

Locate the Project

Prefer the first existing path:

1. A repository path the user gives explicitly.
2. journal-finder-config.json in this installed skill directory, using its project_dir or data_dir.
3. The current workspace if it contains skill/scripts/query_db.py.
4. /Users/a86166/journal-finder.
5. ~/journal-finder.

The expected project contains:

• skill/scripts/query_db.py for keyword + semantic/hybrid journal search.
• skill/scripts/semantic_search.py for direct semantic search.
• data/journals_ssci.json as the canonical database.
• Optional semantic assets: data/journal_index.faiss, data/journal_embeddings.npy, data/journal_index_map.json.

If the project cannot be found, ask the user for the journal-finder repository path.

Intake

Extract these from the user request:

• Title and abstract, or at least a topic paragraph.
• Keywords, explicit or inferred.
• Preferences: topic fit, prestige, speed, budget, CN-friendly, annual volume/capacity, review-data reliability.
• Constraints: maximum APC, OA-only, minimum JCR quartile, maximum review days, review-data required, whether the manuscript is a review article.

If there is no usable title, abstract, or topic paragraph, ask for one before running a query.

If the user provides a usable title, abstract, or topic but does not state submission preferences, ask one concise preference question before running a query: 你更看重哪些方面？可选：主题贴合、求档次、求快、预算有限、国人友好、容量稳；也可以说“均衡推荐”。 Continue after the user answers. If the user says balanced, unspecified, or no preference, use --sort balanced --priorities fit.

Infer 3-8 scope-defining keywords. Keep countries, methods, and samples as context, but do not let them replace the main field/problem concepts. For law papers, include legal-field concepts such as law, legal regulation, human rights, environmental law, criminal justice, constitutional law, international law, legal sociology, or law and society when relevant.

Query

Run query_db.py from the Journal Finder project. Use python3 unless the project documents another environment.

Basic query:

python3 skill/scripts/query_db.py --discipline all --keywords "keyword1,keyword2,keyword3" --top 15

Useful options:

--sort balanced|prestige|speed|cn_friendly
--priorities fit,prestige,speed,budget,cn,volume,data
--max-apc 3000
--oa-only
--min-quartile Q1|Q2|Q3|Q4
--max-review-days 180
--require-review-data
--include-review-only
--data-dir /path/to/journal-finder/data

Preference mapping:

• Topic fit: --sort balanced --priorities fit
• Prestige: --sort prestige --priorities prestige
• Speed: --sort speed --priorities speed
• Speed with evidence only: --sort speed --priorities speed --require-review-data
• Budget limited: --sort balanced --priorities budget --max-apc
• CN-friendly: --sort cn_friendly --priorities cn
• Balanced prestige + speed: --sort balanced --priorities prestige,speed
• Review manuscript: add --include-review-only

If the user asks for separate lists, run separate queries for each priority and compare the top results. If they ask to balance goals, run one combined query.

Interpret Results

Parse the JSON output. Check these fields before presenting recommendations:

• query.semantic_search: whether semantic search was used.
• query.semantic_status and query.semantic_error: fallback reason when semantic search is unavailable.
• quality.status: ok, limited_matches, or no_good_match.
• Each result's _final_score, _fit_scores, _recommendation_notes, _risk_flags, _review_confidence, and _review_evidence.

If semantic_search is false, tell the user the run used keyword matching only and is weaker for cross-disciplinary topics.

If quality.status is limited_matches or no_good_match, do not pad the table with weak journals. Explain that constraints or keywords should be relaxed.

Treat review time carefully:

• credible: usable review-speed evidence.
• limited or very_limited: mention that the sample is sparse.
• missing: do not imply the journal is fast or slow.

Output

Match the user's language. For Chinese users, use Chinese.

Present 5-8 recommended journals unless the user asks for a different count. Use this table:

| 期刊名称 | JCR分区 | 中科院分区 | 影响因子 | OA/APC | 国人占比 | 年发文量 | 审稿周期(天) | 避坑提示 |
|----------|---------|-----------|---------|--------|---------|---------|-------------|---------|

Use — for missing values. After the table, briefly explain the top 3 journals:

• Why the journal fits the paper.
• What tradeoffs or risks exist.
• Whether it is a reach, solid-fit, or safer option.

Always prioritize topic fit over prestige, speed, or low APC. Flag prestigious but weak-fit journals as risky rather than recommending them confidently.

Scope Notes

The database covers SSCI/AHCI humanities and social sciences broadly, including law. It can support law and adjacent areas such as criminology, legal sociology, governance, human rights, international law, environmental law, law and economics, and public policy.

The database is not designed for pure natural sciences, clinical medicine, environmental chemistry, engineering, or ecology. For out-of-scope topics, ask for the closest social science, humanities, or environmental-health angle.

Common Failure Modes

• Missing database: ask for the project path or pass --data-dir.
• Missing semantic dependencies or FAISS index: proceed with keyword fallback, then mention that semantic search needs the project's dependency setup and scripts/build_embeddings.py.
• Overly narrow constraints: relax APC, quartile, review-data, or OA filters.
• Keywords too methodological: replace method-only terms with journal-scope concepts.

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: yennhi85219-glitch
• Source: yennhi85219-glitch/journal-finder
• License: MIT

Install and usage instructions live in the source repository linked above.