SuperAPI AI Directory
Partner with us中文AI newsFavorites
Skillsstyle-journal-rewrite

style-journal-rewrite

Writing & research

View repository

Updated 10-15 16:45

About this skill

Source: https://github.com/aipoch/medical-research-skills

Style and Journal Format Adaptation Skill

Trigger Conditions

Triggers

  • Style change: "rewrite in XX style", "write like XX", "match this tone"
  • Journal format: "format for XX journal", "rewrite following this format", "convert to NC format"
  • Adjust reference style or abstract structure

Does Not Trigger

  • Factual additions, content expansion
  • Proofreading typos
  • Complete structural rewrite (unrelated to style/format)
  • Adding references only (use find-paper-references)

Requirements

  • python-docx: pip install python-docx (required for B4/C4 visual formatting, B1/B2 scripts)

Process

0. Define Boundaries

Determine three things:

  • Input format: DOCX / MD / TXT / conversation text?
  • Task type: Style only / format only / both?
  • Output format: Same as input or specific format requested?

Then branch into the corresponding path below.

1. Extract Target Profile

Read extraction guides from references/:

  • Style adaptation → references/style-profile-extraction.md
  • Journal format adaptation → references/journal-format-extraction.md

Do not fabricate rules not found in the user's materials.

Path A: Pure Style Adaptation (any input format)

Read references/rewriting-criteria.md directly, rewrite according to profile. Change only expression style, not formatting.

Output matches input format. DOCX: change text only, not styles; MD/TXT: output same format; conversation text: output directly.

After completion, skip to Section 5 QC.

Path B: DOCX + Format Adaptation

Input is .docx, requires journal format changes (may also include style changes).

B1. Structure Layer

  • Heading numbering:
    • Reference article has no numbering → run scripts/remove_heading_numbers.py <input> <output>
    • Reference article has numbering but different format → manually edit heading text
  • Abstract structure, section order: AI modifies directly

B2. Citation Layer

  • End-of-text list numbering: use convert_ref_format.py --ref-list-style to switch between [N] / N.
  • In-text citations: use convert_ref_format.py --style to switch superscript (Unicode / font format)
  • --range-joiner and --group-separator must be extracted from the reference article's actual writing style, do not assume
  • How the reference article combines consecutive citations: observe multi-citation sequences in text, then set parameters

B3. Field Code Detection

Scan w:instrText in DOCX XML:

  • Contains ADDIN ZOTERO_ITEM → load format-references-zotero skill, search CSL style → install → Refresh
  • Contains {Author, Year, Title} → load format-references-endnote skill, check .ris in same directory → deliver
  • Plain text citations (no field codes) → skip this step, B2 already handled

B4. Visual Formatting

python-docx adjusts font, size, line spacing, paragraph spacing, indentation, alignment per paragraph. Use para.runs to set font.name / font.size, use para.paragraph_format for spacing.

*/!* Do NOT use para._element.clear() to clear paragraphs containing field codes — this destroys Zotero/EndNote field codes. B4 only modifies w:rPr and w:pPr.

Path C: MD + Format Adaptation

Input is .md, requires journal format changes (may also include style changes).

C1. Check PMID Markers

MD must contain [PMID:xxxxxxxx] markers to use this path.

  • Has markers → continue
  • No markers → prompt user: run find-paper-references first to insert PMID markers, then return to this skill

C2. Style and Structure Adjustment

AI directly modifies MD text:

  • Remove heading numbers, adjust abstract structure, adjust section order
  • Change writing style (if needed)

/!\ Critical constraint: When moving or rewriting sentences with citations, [PMID:xxxxxxxx] markers must follow the text — do not lose, change numbers, or modify spaces.

C3. Hand Off to Citation Formatting Skill

Load the corresponding skill based on user's tool:

  • Zotero → Skill("format-references-zotero")
  • EndNote → Skill("format-references-endnote")

Both skills do: parse [PMID:xxxxx] → fetch PubMed metadata → download target journal CSL style → output DOCX (in-text citations and bibliography formatted per journal style).

This step replaces all work done by convert_ref_format.py in Path B, with more professional results — citation formatting is driven by Zotero/EndNote engines, avoiding inconsistencies from manual Unicode superscript assembly.

C4. Visual Formatting

python-docx applies journal-required font, size, line spacing, margins, and figure/table caption positioning on the DOCX produced by C3.

Same as B4: only modify w:rPr and w:pPr, do not affect Zotero/EndNote field codes.

C5. Output

Final DOCX = C2 (structure+style) → C3 (citation formatting) → C4 (visual formatting)

4. Output and Delivery

  • DOCX → doc.save() → deliver_attachments
  • MD → Write() → deliver_attachments
  • TXT → Write() → deliver_attachments
  • Conversation text → output directly to conversation
  • After delivery, summarize changes in one sentence

5. Quality Check

Verify each item:

  • Original key information fully preserved, no example content mixed in
  • If style target: consistent tone throughout, not just first few paragraphs
  • If journal format target:
    • Heading numbering matches reference article
    • Abstract structure matches reference article
    • (DOCX path) Field codes detected, corresponding citation skill executed
    • (MD path) PMID markers intact, citation skill executed
    • Visual formatting (font/size/line spacing) applied
  • No fabricated "journal rules" unsupported by user materials
  • File opens normally, no formatting issues (MD path: final DOCX opens with working Refresh/Update Citations)

Error Handling

  • If required inputs are missing, state exactly which fields are missing and request only the minimum additional information.
  • If the task goes outside the documented scope, stop instead of guessing or silently widening the assignment.
  • If execution fails, report the failure point, summarize what can still be completed safely, and provide a manual fallback.
  • Do not fabricate files, citations, data, search results, or execution outcomes.

Input Validation

This skill accepts requests that match the documented purpose of style-journal-rewrite and include enough context to complete the workflow safely.

Do not continue the workflow when the request is out of scope, missing a critical input, or would require unsupported assumptions. Instead respond:

style-journal-rewrite only handles its documented workflow. Please provide the missing required inputs or switch to a more suitable skill.