Research Paper Writing Skills
Section-focused ML/CV/NLP paper revision with reverse outlining and explicit claim–evidence maps.
Last reviewed by the paperbanana team on Oct 20, 2026
Install
git clone https://github.com/Master-cai/Research-Paper-Writing-Skills.git
mkdir -p "$HOME/.claude/skills"
cp -R Research-Paper-Writing-Skills/research-paper-writing "$HOME/.claude/skills/"What is Research Paper Writing Skills?
Research Paper Writing Skills is a focused writing package for ML, computer-vision, and NLP papers. Its workflow starts with the paper's story and a compact outline, then loads only the relevant section guide for the Abstract, Introduction, Related Work, Method, Experiments, or Conclusion. It checks each paragraph for a single message and uses reverse outlining to connect topic sentences and supporting evidence to the section's thesis.
The output contract includes revised paragraphs with explicit roles, a short self-review checklist, and a Claim / Evidence / Status map for major claims. Unsupported claims should be weakened or removed rather than made more persuasive. This is instruction-driven editing and self-review, not an independent experiment or citation-verification engine.
The README credits Prof. Peng Sida's public research notes as the main source of its writing methodology and describes Master-cai's contribution as organization and adaptation into skills. This directory lists that adaptation, not an official endorsement by Prof. Peng. The repository declares MIT; check the original source terms before redistributing third-party notes or examples.
Core capabilities
Section-specific writing guidance
Load the guide and examples for the section being edited instead of injecting every reference file into context.
Reverse outlining
Map paragraph topic sentences and supporting points back to the main claim to diagnose missing links and redundant passages.
Claim–evidence maps
Label each major revised claim as supported or needing evidence, especially in the Abstract and Introduction.
Five-dimension self-review
Examine contribution, writing clarity, experimental strength, evaluation completeness, and method soundness before submission.
What you can use it for
Revise an Introduction
Supply the current Introduction and actual findings; ask for a compact outline, clearer paragraph roles, and a claim–evidence map.
Explain a method module
Use the Method guide to state the module's motivation, design, and technical advantage with consistent terminology.
Audit claims before submission
Compare Abstract and Introduction claims with reported experiments and retain a list of items needing revision or new evidence.
How to get started
- 1
Clone and install the package
Run the commands above to clone the repository and copy research-paper-writing into ~/.claude/skills/. The README also documents Codex and Gemini skill-directory installs.
- 2
Provide a concrete target
Supply a section, the research question, and the results supporting its claims. Name the research-paper-writing skill explicitly.
- 3
Review the structured rewrite
Check the outline, paragraph roles, revised text, and evidence status against your own research records.
- 4
Resolve missing evidence
Add verified evidence or weaken unsupported claims, then rerun the relevant section or paper-review checklist.
How it compares to similar skills
Choose this package for section-level ML/CV/NLP writing when your research and results already exist. Its value is a small, explicit editing workflow rather than a full research platform.