Paper Select Journal

Core Principles

Current information must be verified in real-time: scope, official website, industry recognition, CAS partition, and papers from the last 3 months are all time-sensitive information and cannot rely on old memory.
Intermediate files are only allowed to be stored in the
```
.paper-select-journal/
```
hidden directory under the current working directory; only if the user explicitly specifies another directory can the default value be overwritten.
Set1 no longer relies on fixed semantic weights. The script is only responsible for minimum hard filtering and candidate pool sorting, and the real semantic planning is completed by the current host model.
Set1 is not the final answer. The final report only retains Set3 with sufficient evidence, up to 10 journals.
Do not recommend obviously warned, predatory journals or journals with an impact factor lower than
```
3
```
; if exceptions with IF lower than
```
3
```
are indeed retained, it must be stated "why it is still a reliable choice recognized by human experts in the field".
It is better to report fewer journals than to forcefully fill up 10 journals.

Input and Workspace

User preferences are optional, but manuscript is mandatory.
Manuscript can come from pasted title / abstract / full text fragments, or local
```
.md
```
,
```
.txt
```
,
```
.tex
```
,
```
.pdf
```
,
```
.docx
```
files, or a combination of these.
Once entering the hidden workspace process, subsequent
```
analysis/*.json
```
files for script reading must be kept within the current run directory; do not point
```
manuscript_profile.json
```
,
```
set2_scope_review.json
```
,
```
final_recommendations.json
```
to outside the run directory.

First initialize the hidden workspace:

bash

python3 <skill_root>/scripts/init_workspace.py --project-root .

The script will create

.paper-select-journal/run-<timestamp>/

, which contains at least:

```
inputs/
```
```
analysis/
```
```
candidates/
```
```
pubmed/
```
```
reports/
```

All subsequent intermediate files must remain within this run directory.

Workflow

1. Write Manuscript Profile First

After fully understanding the paper, write the results into

analysis/manuscript_profile.json

Template:

templates/manuscript_profile.template.json

Writing guide:
```
references/manuscript-profile.md
```
Minimum fields:
```
title
```
```
abstract
```
```
keywords
```
```
manuscript_summary
```

The purpose of the profile is to help AI understand the manuscript, not to feed into a fixed scoring formula. If user preferences are complex, prioritize writing the preferences as natural language into

target_journal_brief

notes

, rather than forcing many hard-coded clues for the script. If manual exception journals with low IF need to be retained, they can only be used as subsequent manual supplementary candidates, and it must be explained in the final report "why it is still a reliable choice in the field despite being below the threshold".

2. Generate Set1 Candidate Pool with Built-in

2023IF.xlsx

Built-in directory:

assets/journal_catalog/2023IF.xlsx

Run:

bash

python3 <skill_root>/scripts/shortlist_journals.py \
  --workspace .paper-select-journal/run-<timestamp> \
  --profile .paper-select-journal/run-<timestamp>/analysis/manuscript_profile.json

Outputs:

```
candidates/set1_candidates.json
```
```
candidates/set1_candidates.md
```

The script only performs minimum hard filtering here:

Impact factor lower limit
Journals explicitly excluded by users
Basic metadata sorting (JIF, partition, OA ratio, citation count)

Do not misunderstand this step's output as "a final shortlist already sorted by semantics". You must read the candidate pool, then combine with the manuscript to independently plan the journals that are truly worthy of entering Set2.

3. Verify Scope, Official Website, Partition and Quality via Internet to Get Set2

Based on

candidates/set1_candidates.json

and the manuscript profile, independently decide which candidates to verify first, and verify each one via the internet:

Official website
Aims & Scope
CAS small category and its partition
Industry recognition
Whether there are warning / predatory journal signals

Priority sources:

Journal official website
PubMed / NLM
Mainstream publisher pages
Trusted partition information sources

Write the verified journals into:

```
analysis/set2_scope_review.json
```

Template:

templates/scope_review.template.json

Verification criteria:

references/journal-quality-checklist.md

4a. Fetch Raw PubMed Paper Evidence of Set2 from the Last 3 Months

Run:

bash

python3 <skill_root>/scripts/fetch_pubmed_recent.py \
  --workspace .paper-select-journal/run-<timestamp> \
  --profile .paper-select-journal/run-<timestamp>/analysis/manuscript_profile.json \
  --scope-review .paper-select-journal/run-<timestamp>/analysis/set2_scope_review.json

Outputs:

```
pubmed/recent_articles.json
```
```
pubmed/recent_articles.md
```

This only provides raw evidence and is not responsible for scoring or sorting. The script only performs API calls, XML parsing, and sorting by date.

4b. AI Assess Theme Similarity to Decide Which Journals Enter Set3

You must read simultaneously:

```
analysis/manuscript_profile.json
```
```
pubmed/recent_articles.json
```

The "AI" here refers to the host model currently executing this skill:

Completed by the current Claude session in Claude Code
Completed by the current Codex session in Codex
Do not call external AI APIs, independent model services, or separate scoring scripts for Step 4b

That is, the planning, semantic judgment, Set3 retention decision, and writing of

set3_similarity_review.json

in Step 4b must all be completed natively using the AI computing power provided in the current working environment.

Judge the semantic relevance between the last 3 months' papers of each Set2 journal and the manuscript in the following dimensions:

Whether the theme is truly relevant, not just token collision
Whether the research questions are close
Whether the methodologies are close
Whether the number and density of relevant papers are sufficient to support inclusion in the final recommendation

When executing, first quickly browse all Set2 recent papers to form a comparison framework, then make semantic judgments for each journal one by one, and finally decide Set3 retention uniformly and write reviewable reasons. Do not degrade this step into mechanical token scoring or hard-coded weighting formulas.

Write the conclusions into:

```
analysis/set3_similarity_review.json
```

Template:

templates/set3_similarity_review.template.json

Each journal must include at least:

```
journal_name
```
```
include_in_set3
```
```
similarity_assessment
```
```
relevant_articles
```
```
irrelevant_articles_count
```
```
overall_relevance_level
```

overall_relevance_level

only allows:

```
high
```
```
medium
```
```
low
```
```
none
```

5. Form Final Recommendation JSON

Based on

analysis/set3_similarity_review.json

, write up to

final journals into

analysis/final_recommendations.json

Template:

templates/final_recommendations.template.json

Field description:
```
references/report-schema.md
```

Must retain:

Impact factor
CAS small category and its partition
Industry recognition
Official website
Reasons for recommendation
Similar theme papers from the last 3 months
AI
```
relevance
```
for each evidence paper

6. Render Final Markdown Report

Run:

bash

python3 <skill_root>/scripts/render_report.py \
  --workspace .paper-select-journal/run-<timestamp> \
  --final-json .paper-select-journal/run-<timestamp>/analysis/final_recommendations.json

Final output:

```
reports/paper-select-journal-report.md
```

If using

--output

to overwrite the default file name, it can only be written into the current run directory, and the final Markdown report cannot be written outside the hidden workspace.

Final Report Requirements

All journals are written in the same Markdown file
Each journal uses
```
#
```
level, with
```
##
```
,
```
###
```
as needed below
Each journal must clearly state: impact factor, CAS small category and partition, industry recognition, official website, recommendation reasons, as well as a table of similar theme papers from the last 3 months and AI relevance explanations

Decision Rules

If scope does not match, do not force recommendation no matter how high the IF is
If there are obvious warning / predatory journal risks, eliminate directly
If there are no similar theme papers in the last 3 months, it is not necessarily eliminated, but the recommendation level should be lowered
Journals with
```
include_in_set3
```
as
```
false
```
should not be included in the final recommendation
When CAS partition cannot be reliably verified, prioritize replacing with candidates with more transparent information
When users do not have explicit preferences, independently choose the most reliable solution, and do not turn the question into a blocking issue

Command Path Description

```
<skill_root>
```
refers to the actual installation directory of the current skill.
Do not assume that the user must have a source code copy of
```
paper-select-journal/
```
in the current working directory.
If you are already in the skill root directory, you can also run
```
python3 scripts/...
```
directly.

Reference Files

```
references/manuscript-profile.md
```
```
references/journal-quality-checklist.md
```
```
references/report-schema.md
```

paper-select-journal

NPX Install

Tags

SKILL.md Content (Chinese)

Paper Select Journal

Core Principles

Input and Workspace

Workflow

1. Write Manuscript Profile First

2. Generate Set1 Candidate Pool with Built-in
`2023IF.xlsx`

3. Verify Scope, Official Website, Partition and Quality via Internet to Get Set2

4a. Fetch Raw PubMed Paper Evidence of Set2 from the Last 3 Months

4b. AI Assess Theme Similarity to Decide Which Journals Enter Set3

5. Form Final Recommendation JSON

6. Render Final Markdown Report

Final Report Requirements

Decision Rules

Command Path Description

Reference Files