This project offers a comprehensive suite of 16 Claude Code Skills designed to automate and enhance academic paper writing, polishing, and submission workflows. Targeting researchers and academics, it aims to significantly reduce the manual effort involved in preparing manuscripts for publication, from initial drafting and translation to final review simulation.

How It Works

The workflow is built around modular Claude Code Skills, each prefixed with ppw:, callable within a Claude Code session. Core functionalities include translation, text polishing, AI-generated content detection and rewriting, and peer review simulation. Advanced features leverage external MCP (Model-centric Processing) servers for tasks like Semantic Scholar API integration for literature searches and Chrome DevTools for Google Scholar automation. The system supports structured, multi-step processes and a "Team Mode" for parallel section processing.

Quick Start & Requirements

Installation involves executing two commands within a Claude Code session: /plugin marketplace add Lylll9436/Paper-Polish-Workflow-skill /plugin install paper-polish-workflow@paper-polish-workflow All 16 skills become immediately available. Optional MCP server configurations (Semantic Scholar, Chrome DevTools) require setup within Claude Code's settings and may involve launching browsers with specific debugging flags.

Highlighted Details

• A robust set of 16 skills covers the entire academic writing lifecycle, including translation, polishing, AI detection, reviewer simulation, abstract generation, literature search, and figure captioning.
• The "Repo-to-Paper" skill automates the generation of academic papers directly from Python Machine Learning experiment repositories.
• "Team Mode" enables parallel processing of paper sections (e.g., polish, translation, de-AI) via subagents, supporting both LaTeX and Markdown formats.
• Features include geography-aware captioning and visualization recommendations, and a two-phase AI detection and rewriting workflow.
• Optional integrations with Semantic Scholar and Google Scholar via MCP servers streamline literature discovery and citation management.

Maintenance & Community

Bug reports and feature requests are managed through GitHub Issues, with detailed contribution guidelines provided in CONTRIBUTING.md and CONTRIBUTING_CN.md. The project adapts prompt templates from awesome-ai-research-writing and utilizes the get-shit-done development framework.

Licensing & Compatibility

The provided documentation does not specify a software license. Compatibility is primarily within the Claude Code environment; specific external dependencies for optional MCP servers may apply.

Limitations & Caveats

Certain advanced skills, such as literature search and paper retrieval from Google Scholar, are dependent on the optional configuration and setup of external MCP servers. The "Team Mode" currently employs a proof-of-concept quality gate for subagent processing, with full parallel dispatch across all sections planned for future versions.