Home
cd ../playbooks
Academic ResearchIntermediate

Research Paper Structure Editor

Improve the structural clarity of ML/CV/NLP research papers before submission — a reverse-outlining check that maps every paragraph's topic sentence back to the section thesis, section-by-section guidance for Abstract through Conclusion, a five-dimension adversarial self-review, and a claim-evidence map that catches unsupported claims in the Abstract and Introduction before a reviewer does.

5 minutes
By Master-cai (Research-Paper-Writing-Skills)Source
#academic-writing#research-papers#machine-learning#paper-writing#peer-review

The most common reason a paper reads as unclear isn't bad sentences — it's a paragraph whose topic sentence doesn't actually map back to the section's thesis, and reverse outlining (write down the thesis, then every paragraph's opening claim, then check the mapping) catches that in minutes instead of a reviewer catching it in rejection comments.

Who it's for: graduate students and researchers preparing an ML/CV/NLP paper for submission, academic authors who want a structural editing pass distinct from grammar polishing, anyone doing a pre-submission self-review who wants a systematic adversarial pass instead of just rereading the draft, co-authors splitting section-writing who need consistent paragraph-level structure across the paper

Example

"Review my Introduction and Method sections for clarity before I submit" → A reverse outline built from each section's topic sentences and checked against its stated thesis, paragraphs failing the mapping flagged for revision, a claim-evidence map built for every claim in the Introduction checking it's actually traceable to a later result, and a five-dimension adversarial self-review (contribution, clarity, experimental strength, evaluation completeness, method soundness) before the draft is called done

CLAUDE.md Template

New here? 3-minute setup guide → | Already set up? Copy the template below.

# Research Paper Structure Editor

Improve academic paper writing quality for ML/CV/NLP-style papers — clear section structure, paragraph flow, and reviewer-facing presentation. Use when drafting or revising the Abstract, Introduction, Related Work, Method, Experiments, or Conclusion; polishing figures/tables; checking claim-support alignment; or performing a self-review before submission.

## Overview

Rewrite a research paper into a reviewer-friendly, high-clarity draft. Prioritize first-impression quality (figures, tables, layout), logical flow, and evidence-backed claims.

## Core Workflow

1. Clarify the paper's story before doing any sentence-level edits.
2. Work section by section, applying that section's specific guidance.
3. Rewrite paragraph-by-paragraph — one message per paragraph, not a bulk rewrite.
4. Run reverse outlining after writing each section.
5. Check every major claim in the Abstract and Introduction against the actual experimental evidence.
6. Run a final adversarial review before calling the draft done.

## Global Principles

1. **One paragraph, one message.** Never let a paragraph try to carry two separate points.
2. **State the paragraph's message in the first sentence.** A reader should know what a paragraph is for before finishing its first sentence.
3. **Make nouns self-contained.** Define new terms before reusing them — don't assume the reader remembers a term from three pages back.
4. **Maintain sentence-to-sentence flow.** Every sentence should connect to the previous one via a clear relation: cause, contrast, consequence, or refinement.
5. **Iterate with adversarial self-review.** Read the draft as a skeptical reviewer looking for reasons to reject, not as the author looking to confirm it's fine.
6. **Treat visual quality as core content, not decoration.** A teaser figure and a clean pipeline diagram carry real information; don't treat them as an afterthought.
7. **Use readable, minimal-ink tables.** Every line and shaded cell should earn its place.
8. **Keep formatting consistent and tidy** throughout the whole document.

## Paragraph Clarity Check

Use this whenever asked whether a paragraph "flows" or is clear.

**1. Read as an external reader:**
- Does this paragraph have one explicit message?
- Does the first sentence state what the paragraph will do?
- Are all key nouns/terms readable without hidden context from earlier?
- Does each sentence connect to the previous one with a clear relation (cause, contrast, consequence, refinement, example)?

**2. Run reverse outlining for the section:**
- Write down the section's thesis or main claim.
- Write down each paragraph's topic sentence.
- Write down the evidence or explanation points under each paragraph.
- Check the mapping: does each topic sentence support the thesis, and does the evidence support its topic sentence?
- Revise or remove any paragraph that can't be mapped cleanly onto this structure.

**3. If flow is still weak**, add temporary section headers and explicit transition phrases during revision, then remove any headers that aren't actually needed before finalizing.

## Section-by-Section Focus

Work on one section at a time rather than trying to hold the whole paper's guidance in mind at once:

- **Abstract** — state the problem, the gap, the method, and the headline result in that order; every claim here must be traceable to an actual result later in the paper.
- **Introduction** — motivate the problem before describing the solution; end with a clear, numbered list of contributions.
- **Related Work** — organize by dimension of comparison (what makes this work different), not a flat list of "X did Y, Z did W."
- **Method** — motivation, design, and the technical advantage of each design choice, in that order, for every subsection.
- **Experiments** — state the question each experiment answers before presenting its results.
- **Conclusion** — summarize contributions and their significance without introducing new claims.

## The Final Review

Before finalizing, append and answer a five-dimension self-review question list, then revise based on any unresolved items:

1. **Contribution** — is what's novel here stated clearly and is it actually novel?
2. **Writing clarity** — does every section pass the paragraph clarity check?
3. **Experimental strength** — do the experiments actually support the claims made about them?
4. **Evaluation completeness** — are the standard baselines, ablations, and failure cases covered?
5. **Method design soundness** — would a skeptical reviewer accept the design choices as justified, not just described?

Treat claim-evidence alignment as a hard constraint, especially in the Abstract and Introduction — perform an adversarial pass specifically hunting for claims the results don't actually support, and resolve every high-risk question before calling the draft final.

## Execution Rules

1. Build a mini-outline before drafting any prose.
2. For each subsection, explicitly include motivation, design, and technical advantage where applicable.
3. Avoid a writing style that reads like incremental patching of a naive baseline — state the actual contribution, not a sequence of small fixes.
4. Keep terminology stable across the entire paper — the same concept should never get two different names.
5. If a claim can't be supported by the actual results, weaken or remove it rather than leaving it as written.

## Output Contract

When asked to rewrite or draft sections, return:

1. A compact section outline (3–7 bullets).
2. Revised paragraphs with explicit paragraph roles noted (opening/challenge/method/advantage/evidence/limitation).
3. A short self-review checklist covering clarity, flow, terminology consistency, unsupported claims, and missing evidence.
4. A claim-evidence map for each major claim in the revised text: `Claim: ... | Evidence: ... | Status: supported/needs evidence`.

Get new playbooks like this one

One email a week with new Claude Code workflows. Free, like everything here.

No spam. Unsubscribe anytime.

README.md

What This Does

A structural editing methodology for ML/CV/NLP-style research papers, distinct from sentence-level grammar polishing — it treats paper clarity as a reverse-outlining problem: for each section, write down the thesis, write down every paragraph's topic sentence, write down the evidence under each paragraph, and check whether the whole thing actually maps together. A paragraph that can't be cleanly mapped back to the section's thesis gets revised or removed, which catches the single most common source of "this reads unclear" feedback — a paragraph technically well-written in isolation that doesn't actually connect to what the section is arguing.

It pairs that structural check with section-specific guidance (an Abstract's claims need to be traceable to later results; Related Work should organize by dimension of comparison rather than a flat "X did Y" list; a Method subsection should state motivation, design, and technical advantage in that order) and a genuinely adversarial final pass: a five-dimension self-review (contribution, clarity, experimental strength, evaluation completeness, method soundness) and an explicit claim-evidence map that traces every major claim in the Abstract and Introduction to the actual result supporting it — weakening or removing anything that can't be traced, rather than leaving an overclaim in place. It works one paragraph at a time rather than bulk-rewriting a whole section at once, keeping each edit reviewable.


Quick Start

Step 1: Create a Project Folder

mkdir paper-structure && cd paper-structure

Step 2: Download the Template

Click Download above, then:

mv ~/Downloads/CLAUDE.md ./

Step 3: Review or Draft a Section

claude

Point Claude at a paper draft (or a section of one) and ask for a structural review or help drafting. It will build a reverse outline to check paragraph-to-thesis mapping, apply the relevant section-specific guidance, and produce a claim-evidence map for anything that reads as a claim needing support.


Tips & Best Practices

  • Run the reverse-outlining check on the Introduction and Abstract before anywhere else — claim-evidence misalignment there is what reviewers notice first and penalize hardest.
  • Ask for one paragraph revised at a time rather than a whole section rewritten at once; it's much easier to verify a single paragraph's new topic sentence still supports its section's thesis than to audit a bulk rewrite.
  • Take the "avoid incremental-patching style" rule seriously in the Method section specifically — describing a contribution as a sequence of small fixes to a baseline reads weaker than stating the actual unified design rationale, even when the underlying work is identical.

Limitations

  • A structural and clarity editor, not a fact-checker or citation verifier — it doesn't check whether cited claims about prior work are accurate, or whether references actually exist; pair it with citation-verification tooling for that.
  • Tuned for the conventions of ML/CV/NLP-style papers (contributions lists, ablations, baselines); a humanities or theoretical-math paper's structural conventions differ enough that some section-specific guidance won't transfer directly.
  • Focuses on structure and claim-evidence alignment specifically — sentence-level grammar and word-choice polishing is a separate, complementary editing pass.

$Related Playbooks

Academic Research

Scientific Manuscript Editor

Five focused editing passes for scientific manuscripts: verb tense conventions by section, topic-sentence structure, line-by-line grammar, substantive clarity editing, and a pre-submission compliance check.

10 minutes
Intermediate
Academic Research

Scientific Matchms

Spectral similarity and compound identification for metabolomics. Use for comparing mass spectra, computing similarity scores (cosine, modified cosine), and identifying unknown compounds from spectral libraries. Best for metabolite identification,...

10 minutes
Intermediate
Academic Research

Scientific Medchem

Medicinal chemistry filters. Apply drug-likeness rules (Lipinski, Veber), PAINS filters, structural alerts, complexity metrics, for compound prioritization and library filtering.

10 minutes
Intermediate
Academic Research

Scientific Metabolomics Workbench Database

Access NIH Metabolomics Workbench via REST API (4,200+ studies). Query metabolites, RefMet nomenclature, MS/NMR data, m/z searches, study metadata, for metabolomics and biomarker discovery.

5 minutes
Beginner
Academic Research

Scientific Molecular Dynamics

Run and analyze molecular dynamics simulations with OpenMM and MDAnalysis. Set up protein/small molecule systems, define force fields, run energy minimization and production MD, analyze trajectories (RMSD, RMSF, contact maps, free energy surfaces)...

15 minutes
Advanced
Academic Research

Scientific Molfeat

Molecular featurization for ML (100+ featurizers). ECFP, MACCS, descriptors, pretrained models (ChemBERTa), convert SMILES to features, for QSAR and molecular ML.

10 minutes
Intermediate
Academic Research

Scientific Monarch Database

Query the Monarch Initiative knowledge graph for disease-gene-phenotype associations across species. Integrates OMIM, ORPHANET, HPO, ClinVar, and model organism databases. Use for rare disease gene discovery, phenotype-to-gene mapping, cross-speci...

5 minutes
Beginner
Academic Research

Scientific Neurokit2

Comprehensive biosignal processing toolkit for analyzing physiological data including ECG, EEG, EDA, RSP, PPG, EMG, and EOG signals. Use this skill when processing cardiovascular signals, brain activity, electrodermal responses, respiratory patter...

10 minutes
Intermediate
Academic Research

Scientific Neuropixels Analysis

Neuropixels neural recording analysis. Load SpikeGLX/OpenEphys data, preprocess, motion correction, Kilosort4 spike sorting, quality metrics, Allen/IBL curation, AI-assisted visual analysis, for Neuropixels 1.0/2.0 extracellular electrophysiology....

15 minutes
Advanced
Academic Research

Scientific Omero Integration

Microscopy data management platform. Access images via Python, retrieve datasets, analyze pixels, manage ROIs/annotations, batch processing, for high-content screening and microscopy workflows.

10 minutes
Intermediate
Academic Research

Scientific Openalex Database

Query and analyze scholarly literature using the OpenAlex database. This skill should be used when searching for academic papers, analyzing research trends, finding works by authors or institutions, tracking citations, discovering open access publ...

5 minutes
Beginner
Academic Research

Scientific Opentargets Database

Query Open Targets Platform for target-disease associations, drug target discovery, tractability/safety data, genetics/omics evidence, known drugs, for therapeutic target identification.

5 minutes
Beginner

Browse all Academic Research playbooks →