Writing SOP logo

Writing SOP

Community
wentorai
Writing SOP

Four-phase iterative protocol for academic writing and document production. Covers outline → draft → self-review → polish with Stop Points, material supplementation loops, version comparison, and mechanical self-check. Includes IMRaD, LaTeX, docx/markdown, citation formatting, and md2pdf.

Overview

Publisherwentorai
RepositoryResearch-Claw
Skill nameWriting SOP
Stars
859
Forks
117
Bundled files
Instructions only
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • Self-contained

    Everything the model needs lives in the instructions — no extra files to sync.

  • Open source

    Published by wentorai on GitHub. Read the source before you install it.

Installation

Install the Writing SOP AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/wentorai/Research-Claw/tree/main/skills/writing-sop
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/wentorai/Research-Claw.git /tmp/Research-Claw
mkdir -p .claude/skills
cp -r /tmp/Research-Claw/skills/writing-sop .claude/skills/wentorai-writing-sop
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Writing SOP in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Writing SOP on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Writing SOP is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Academic Writing SOP — 学术写作标准操作规程

When to Read This Skill

Read this skill when the user asks to:

  • Draft or edit academic text (papers, reports, theses, proposals)
  • Format citations or bibliographies
  • Export/import references (BibTeX, RIS, CSV, JSON, PDF, DOI list)
  • Write or compile LaTeX documents
  • Convert markdown to PDF, docx, or other formats
  • Polish academic language or adjust tone
  • Prepare a manuscript for submission

FOUR-PHASE ITERATIVE WRITING PROTOCOL

CRITICAL — READ THIS FIRST

You MUST follow this four-phase protocol for ALL writing tasks longer than one paragraph. NEVER produce a complete draft in a single pass and call it done. Each phase has explicit steps. Complete them IN ORDER.

所有超过一段的写作任务必须遵循四阶段迭代流程。不得一次性写完就结束。

Phase 1  材料审查与大纲  Material Review & Outline    ⏸ STOP → wait for user
Phase 2  初稿生成        First Draft Generation       (continuous, mid-draft search if needed)
Phase 3  自检与修订      Self-Review & Revision        ⏸ STOP → present review results
Phase 4  精修与导出      Polish & Export               ⏸ STOP → present final + diff

Version Tagging: Every workspace_save MUST use version tags: {topic}-v1.md, {topic}-v2.md, {topic}-final.md.


Phase 1: 材料审查与大纲 (Material Review & Outline)

Goal: Confirm requirements, audit materials, produce outline with per-section citation needs. Identify gaps BEFORE writing.

Step 1: Gather Requirements

MUST ask (skip items already answered by user):

  1. Target format — journal / conference / thesis / report / proposal?
  2. Venue/style — specific journal, conference, or template?
  3. Citation style — APA 7 / MLA 9 / Chicago / IEEE / other? (Default: APA 7)
  4. Language — English / Chinese / bilingual?
  5. Length target — approximate word/page count?
  6. Existing materials — notes, data, figures, prior drafts in workspace?

Step 2: Audit Available Materials (审查已有材料)

MUST execute:

  1. workspace_list — check outputs/drafts/ and outputs/reports/ for existing work
  2. library_search — search local library for relevant papers
  3. workspace_read — read any existing notes/outlines the user mentions

Step 3: Generate Section Outline (生成大纲)

MUST produce this table for EACH section:

SectionKey PointsCitations NeededAvailableGap?
IntroductionResearch gap, RQ, contributions5-83YES
Methods............

Step 4: Material Supplementation Loop (材料补充循环)

IF any section has Gap? = YES:

  1. Tell user: "Sections X, Y need more citations. I will search."
  2. Trigger search-sop — use web_search or academic search tools
  3. Add papers via library_add_paper or library_import_bibtex
  4. Update outline table
  5. REPEAT until gaps filled or search exhausted (document remaining gaps)

NEVER proceed to Phase 2 with known gaps unless user explicitly approves.

⏸ STOP POINT 1 — Present Outline (展示大纲,等待确认)

MUST present: (1) outline table, (2) materials summary, (3) remaining gaps, (4) estimated length.

MUST say: "Please review this outline. Reply with changes, or confirm to proceed."

MUST WAIT. Do NOT proceed to Phase 2 automatically.


Phase 2: 初稿生成 (First Draft Generation)

Goal: Write section by section per outline. Save versioned draft.

Step 1: Write Section by Section (逐节撰写)

For EACH section in order:

  1. Write following outline key points
  2. Embed citations per agreed style (see Appendix A)
  3. Per-section check:
    • Every factual claim has citation? If NO → add or flag
    • All cited papers verified/in library? If NO → search or flag
    • All outline key points covered? If NO → expand
  4. If citation missing mid-section: pause → search (search-sop) → add to library → continue
  5. If section needs a figure: pause → load Plotting SOP → generate figure → workspace_save to outputs/figures/ → embed reference ("See Figure N") → continue

Step 2: Save v1

workspace_save → outputs/drafts/{topic}-v1.md
library_export_bibtex → outputs/exports/bibliography-{topic}.bib   (if BibTeX)

Step 3: Progress Note

Tell user: "First draft (v1) saved. Moving to self-review." List any sections with improvised material.

Do NOT wait for input here. Proceed to Phase 3.


Phase 3: 自检与修订 (Self-Review & Revision)

Goal: Execute mechanical checklist. Fix issues. Save v2. Present results.

Step 1: Self-Check Checklist (自检清单)

MUST check ALL 10 items. Record PASS or FAIL for each.

#Check ItemVerify By
1Structure complete — follows IMRaD or target format?Compare headings vs outline
2Every claim cited — no unsupported assertions?Scan each paragraph
3Citation format consistent — same style throughout?Spot-check 5+ citations
4No fabricated references — all papers real?library_search cross-check
5Logical flow — transitions between paragraphs?Read endings→beginnings
6Academic tone — no casual language?Scan for colloquialisms
7Data/figures labeled — numbered, captioned, referenced?Check each table/figure
8Abstract ≤250 words (if applicable)?Word count
9No orphaned sections — every header has content?Scan structure
10Length ±20% of target?Word count

Step 2: Fix Every FAIL

For EACH failed item:

  1. Describe the problem
  2. Fix it
  3. Re-verify → confirm PASS

If fixing requires new citations → trigger search (same as Phase 1 Step 4).

Step 3: Save v2

workspace_save → outputs/drafts/{topic}-v2.md

⏸ STOP POINT 2 — Present Self-Review (展示自检结果)

MUST present:

  1. Checklist results (10 items, PASS/FAIL)
  2. Change summary (v1 → v2)
  3. Remaining concerns
  4. Ask: "Review {topic}-v2.md. Reply with changes, or confirm for final polish."

MUST WAIT. Do NOT proceed to Phase 4 automatically.


Phase 4: 精修与导出 (Polish & Export)

Goal: Apply feedback, final polish, export, show version diff.

Step 1: Apply User Feedback

For each piece of feedback: locate section → make change → verify consistency. If "no changes" → skip to Step 2.

Step 2: Final Language Polish (最终润色)

MUST apply these 5 passes (details in Appendix F):

  1. Conciseness — cut redundancies
  2. Precision — vague terms → exact values
  3. Flow — smooth transitions, add signposts
  4. Terminology — consistent terms for same concepts
  5. Grammar — agreement, tense, articles

Step 3: Save & Export

workspace_save → outputs/drafts/{topic}-final.md

Export per user request:

FormatAction
docxworkspace_export({ source: "…-final.md", format: "docx" })
PDFworkspace_export(…, format: "pdf") or md2pdf pipeline (Appendix E)
LaTeXSave as .tex with preamble (Appendix D)
BibTeXlibrary_export_bibtex

If no format specified, ask user.

Step 4: Version Comparison (版本对比)

MUST present diff summary: structural changes, content changes, quality fixes, word counts (v1 / v2 / final).

⏸ STOP POINT 3 — Final Deliverables (展示最终成果)

MUST present: (1) all saved file paths, (2) version diff, (3) all 10 checks PASS, (4) caveats.

Writing task complete.


APPENDICES — Reference Material for the Phases Above

Appendix A: Citation Formatting

StyleIn-textReference Format
APA 7(Author, Year)Author, A. A. (Year). Title. Journal, vol(issue), pp.
MLA 9(Author page)Author. "Title." Journal, vol. X, no. Y, Year, pp. Z.
ChicagoFootnote / (Author Year)Author. Title. Place: Publisher, Year.
IEEE[1][1] A. Author, "Title," Journal, vol. X, no. Y, pp. Z, Year.
Vancouver(1)1. Author AA. Title. Journal. Year;Vol(Issue):Pages.
Harvard(Author Year)Author, A.A. Year. Title. Journal, Vol(Issue), pp.
NatureSuperscript^11. Author, A. A. Title. Journal Vol, Pages (Year).
ACM[Author Year]Author. Year. Title. In Proc. Conf., Pages.

Default: APA 7. For CSL/GB/T 7714-2015, read Citation Styles skill.

Appendix B: Citation Integrity Rules

  1. Every cited paper MUST exist in local library or be verifiable via DOI.
  2. Cite primary sources. Secondary only if primary genuinely inaccessible.
  3. Paraphrase → cite. Quote → quotation marks + page numbers.
  4. Cannot verify a paper? Say so. NEVER fabricate references.

Appendix C: Writing Quality Standards

  1. Clarity: Active voice. One idea per sentence. Define acronyms on first use.
  2. Precision: Exact numbers, not "many." Specify units.
  3. Structure: IMRaD unless user specifies otherwise.
  4. Tone: Academic 3rd person default. "We" for multi-author. Adjust per user.
  5. Transitions: Signpost phrases ("However," "In contrast," "Building on this,").

IMRaD Section Guidelines

  • Introduction — literature gap, research question, contribution summary
  • Methods — materials, procedures, statistical tests, sample size justification
  • Results — data first, figures/tables referenced, no interpretation
  • Discussion — compare prior work, limitations, implications, future work
  • Abstract — standalone summary (background, methods, results, conclusion; ≤250 words)

Other sections: Literature Review (humanities), Theory/Model (econ/physics), Related Work (CS), Case Study (law/business).

Appendix D: LaTeX Writing SOP

Compilation: xelatex → biber/bibtex → xelatex ×2 → output.pdf Use exec: cd <dir> && xelatex -interaction=nonstopmode main.tex

TemplateUse CasePackage
article/revtexJournalrevtex4-2, elsarticle
beamerPresentationsbeamer
thesisGraduationTHUThesis, SJTUThesis, ustcthesis
IEEE confCS/EEIEEEtran
ACMCSacmart

Formulas: Inline $...$, display \[...\] or equation, align align/aligned. Numbering: equation auto, equation* suppresses. Cross-ref: \label{eq:x} + \eqref{eq:x}.

Overleaf: Use \input{} to split; switch to xelatex for CJK.

Appendix E: Export & Conversion

Export

FormatToolUse Case
BibTeX .biblibrary_export_bibtexLaTeX, Overleaf, JabRef
RIS .rislibrary_export_bibtex(format:'ris')EndNote, Mendeley, Zotero
CSV/JSON/MDworkspace_saveSpreadsheet, API, docs

Import

FormatTool
BibTeXlibrary_import_bibtex
RISlibrary_import_ris
PDFlibrary_add_paper (extract DOI)
DOI listlibrary_batch_add
CSVParse + library_batch_add

Pandoc

bash
pandoc input.md -o output.docx --reference-doc=template.docx          # md→docx
pandoc input.md -o output.pdf --pdf-engine=xelatex -V CJKmainfont="Noto Serif CJK SC"  # md→pdf
pandoc input.docx -o output.md --extract-media=./media                 # docx→md
pandoc input.md -o output.pdf --citeproc --bibliography=refs.bib --csl=apa.csl  # +bib

For CJK: always specify CJK fonts. Use --reference-doc for consistent docx styling.

md2pdf

Use md2pdf-export skill's Puppeteer pipeline. Setup: bash skills/md2pdf-export/scripts/setup-env.sh. Convert: node skills/md2pdf-export/scripts/md2pdf.js input.md [--format A4 --toc --theme github]

Appendix F: Language Polishing Guidelines

  1. Preserve meaning — NEVER alter scientific content or conclusions
  2. Grammar — agreement, tense consistency, articles, parallel structure
  3. Conciseness — "in order to" → "to", "it is important to note that" → "notably"
  4. Hedging — "may suggest," "appears to indicate" (avoid over-hedging)
  5. Terminology — same term for same concept throughout
  6. CJK → English — restructure topic-comment → subject-verb-object

Present changes as tracked edits (original vs. revised).

Appendix G: RC Local Tools Reference

TaskToolPattern
Save draftworkspace_saveoutputs/drafts/{topic}-v1.md
Save revisionworkspace_saveoutputs/drafts/{topic}-v2.md
Save finalworkspace_saveoutputs/drafts/{topic}-final.md
Review draftworkspace_readoutputs/drafts/{file}
List workspaceworkspace_listCheck existing materials
Export Wordworkspace_export{ source: "…", format: "docx" }
Export PDFworkspace_export{ source: "…", format: "pdf" }
Export biblibrary_export_bibtex.bib subset
Search librarylibrary_searchCheck available citations
Add paperlibrary_add_paperAdd found paper
Link to tasktask_link_fileLink file to task
Compile LaTeXexeccd {dir} && xelatex -interaction=nonstopmode main.tex
Pandocexecpandoc {in} -o {out} [opts]

Appendix H: Related Research-Plugins Skills

  • writing/composition/ — academic structure, ML paper writing, abstract crafting
  • writing/citation/ — Zotero, BibTeX management, APA/MLA/Chicago
  • writing/latex/ — LaTeX syntax, Overleaf, formula rendering
  • writing/templates/ — thesis templates (THUThesis, SJTUThesis), conference
  • writing/polish/ — language polishing, tone, grammar tools
  • tools/document/ — PDF parsing (GROBID), format conversion, large docs

Frequently asked questions

What does the Writing SOP AI skill do?

Four-phase iterative protocol for academic writing and document production. Covers outline → draft → self-review → polish with Stop Points, material supplementation loops, version comparison, and mechanical self-check. Includes IMRaD, LaTeX, docx/markdown, citation formatting, and md2pdf.

Why use Writing SOP on TypingMind?

Because you install it once and use it with any model. Writing SOP is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Writing SOP in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/wentorai/Research-Claw/tree/main/skills/writing-sop. TypingMind reads its SKILL.md and installs it as a skill you can enable per chat.

Which AI models can use Writing SOP?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Writing SOP?

As many as you like. As long as a model supports skills, you can use Writing SOP with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Writing SOP AI skill free?

It is published on GitHub by wentorai. Check the repository for licensing terms. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇