aidlessaidless/research-writing-skill/CLAUDE.md

research-writing-skill

CLAUDE.md0 starsChanged 39 days agoMITView on GitHub

This repository is a multi-platform skill/plugin for research-paper writing workflows (not a conventional app/service). It provides a routed skill system for brainstorming, chapter writing, literature review, figures, and LaTeX output. The main execution model is: 1. Entry/routing via skills/using-research-writing/SKILL.md 2. Medium/full-paper orchestration via skills/paper-orchestration/SKILL.md 3. Mandatory brainstorming via skills/brainstorming-research/SKILL.md 4. Evidence-driven sections via skills/evidence-driven-writing/SKILL.md 5. Chapter-by-chapter writing via skills/writing-chapters/SKILL.md 6. Optional specialized skills (literature, experiments, figures, review, LaTeX, environment setup) When editing skills or workflow logic, maintain these repository-level invariants:

2 other projects use a version of this.

Adopted by

Most recently taken up by

Norman-bury/articlewriting-skillword for word, first seen Oct 8, 2026
Norman-bury/research-writing-skillword for word, first seen Oct 8, 2026

Other versions2 projects use a version of this.
Word for word · 2
Norman-bury/research-writing-skill★ 3.4k100% same
Norman-bury/articlewriting-skill★ 3.4k100% same
Edited · 0
None found.
Wrote their own version · 0
None found.
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository21 other files it gives its agents.
- AGENTS.mdAGENTS.md
- research-writing-assistantSkill
- brainstorming-researchSkill
- environment-setupSkill
- evidence-driven-writingSkill
- experiment-results-planningSkill
- figures-diagramSkill
- figures-pythonSkill
- latex-outputSkill
- literature-reviewSkill
- paper-orchestrationSkill
- peer-reviewSkill
- prompts-collectionSkill
- statistical-analysisSkill
- using-research-writingSkill
- verificationSkill
- writing-chaptersSkill
- writing-coreSkill
- writing-humanitiesSkill
- writing-lawSkill
- writing-medicalSkill

The file

Show its rules and text

We split this CLAUDE.md into the rules it gives an agent

13 rules. About 0 are common, 8 are distinctive, and 5 are very specific to this project.

≈0
Common · estimate
Many files like this have a rule of this kind.
≈8
Distinctive · estimate
This author's own choices. Worth a look.
5
Very specific
Tied to their project's names, files or tools. Probably skip.

Common
Distinctive
Very specific

Distinctive:
When editing skills or workflow logic, maintain these repository-level invariants:
line 40
Distinctive:
Do not bypass brainstorming for writing tasks.
line 42
Very specific:
Do not bypass paper-orchestration for medium/full-paper work.
line 43
Distinctive:
Do not draft Introduction or Related Work without an evidence map.
line 44
Very specific:
Keep plan/ as persistent project memory (project-overview.md, progress.md, notes.md, outline.md, stage-gates.md).
line 45
Distinctive:
Keep chapter flow sequential and confirmation-driven (one chapter at a time, user confirmation before moving on).
line 46
Distinctive:
Do not fabricate citations or research facts.
line 47
Distinctive:
Do not present mock/synthetic planning data as real results.
line 48
Distinctive:
Keep user instructions and process notes out of manuscript body.
line 49
Very specific:
run style_check on modified markdown outputs
line 107
Very specific:
run scripts/check_skill_integrity.ps1 after skill workflow edits
line 108
Very specific:
run scripts/research_quality_gate.ps1 -ProjectPath <paper-project> on manuscript workspaces
line 109
Distinctive:
run target Python scripts with representative inputs to verify behavior
line 110

≈ “Common” is an estimate: at least 1 in 5 of the 26,980 CLAUDE.md files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.

Files

CLAUDE.md

This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.

Project purpose

This repository is a multi-platform skill/plugin for research-paper writing workflows (not a conventional app/service). It provides a routed skill system for brainstorming, chapter writing, literature review, figures, and LaTeX output.

The main execution model is:

Entry/routing via skills/using-research-writing/SKILL.md
Medium/full-paper orchestration via skills/paper-orchestration/SKILL.md
Mandatory brainstorming via skills/brainstorming-research/SKILL.md
Evidence-driven sections via skills/evidence-driven-writing/SKILL.md
Chapter-by-chapter writing via skills/writing-chapters/SKILL.md
Optional specialized skills (literature, experiments, figures, review, LaTeX, environment setup)

Key architecture (big picture)

skills/ — Primary modular skill system (authoritative behavior).
- using-research-writing/ is the workflow router and rule gate.
- paper-orchestration/ handles medium/full-paper task packets, subagent routing, review gates, and capability-use audits.
- brainstorming-research/ enforces conversational requirement collection before writing.
- evidence-driven-writing/ forces evidence maps and paragraph blueprints before Introduction/Related Work drafting.
- writing-chapters/ enforces one-chapter-at-a-time writing with two-stage review.
- experiment-results-planning/ separates real results from mock planning data and defines experiment/table/figure handoff.
hooks/ — Session-start injection for platforms (especially Claude/Cursor), including loading entry skill context.
- hooks/session-start injects using-research-writing content at session start.
- hooks/hooks.json wires SessionStart for Claude plugin.
.claude-plugin/plugin.json, .cursor-plugin/plugin.json — Plugin metadata + hook wiring per platform.
plan-template/ — Canonical template files copied into user project plan/.
modules/ — Legacy compatibility docs; keep in sync conceptually with skills/ behavior.
scripts/ — Operational tooling:
- init_plan.sh / init_plan.ps1: bootstrap plan/ from plan-template/
- style_check.sh / style_check.ps1: writing-style checks for markdown
- scholar_search.py: multi-source literature retrieval + citation formatting
- pdf_parser.py: PDF text/section/metadata extraction

Workflow invariants to preserve

When editing skills or workflow logic, maintain these repository-level invariants:

Do not bypass brainstorming for writing tasks.
Do not bypass paper-orchestration for medium/full-paper work.
Do not draft Introduction or Related Work without an evidence map.
Keep plan/ as persistent project memory (project-overview.md, progress.md, notes.md, outline.md, stage-gates.md).
Keep chapter flow sequential and confirmation-driven (one chapter at a time, user confirmation before moving on).
Do not fabricate citations or research facts.
Do not present mock/synthetic planning data as real results.
Keep user instructions and process notes out of manuscript body.
Preserve compatibility across platform entry points (SKILL.md, AGENTS.md, GEMINI.md, plugin manifests, hooks).

Common commands

Run from repository root (research-writing-skill-main).

Plan bootstrap

bash scripts/init_plan.sh

Initialize plan for a specific project directory:

bash scripts/init_plan.sh /path/to/project

PowerShell:

powershell -ExecutionPolicy Bypass -File scripts/init_plan.ps1

Style/lint-like checks for markdown output

bash scripts/style_check.sh chapters/01-introduction.md

PowerShell:

powershell -ExecutionPolicy Bypass -File scripts/style_check.ps1 -FilePath chapters/01-introduction.md

Literature search utility

python scripts/scholar_search.py "deep learning" --sources crossref --year 2020-2024 --limit 20

Export BibTeX:

python scripts/scholar_search.py "transformer attention" --format bibtex --output refs.bib

PDF parsing utility

python scripts/pdf_parser.py paper.pdf --sections --json output.json

Test/verification status

There is no dedicated unit-test framework in this repository. Validation is script-level:

run style_check on modified markdown outputs
run scripts/check_skill_integrity.ps1 after skill workflow edits
run scripts/research_quality_gate.ps1 -ProjectPath <paper-project> on manuscript workspaces
run target Python scripts with representative inputs to verify behavior

Important references

README.md / README_EN.md — user-facing positioning, workflow, and platform support.
SKILL.md — legacy/main entry for compatible runtimes.
AGENTS.md and GEMINI.md — platform-specific skill-loading entry points.

Notes on instruction files

No .cursorrules, .cursor/rules/, or .github/copilot-instructions.md were found in this repository at the time of writing this file.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

It worked
Partly
It didn't

What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

Check your setup, free