What is research-paper-pipeline?

This is a skill plugin for Claude Code that automates the process of writing academic research papers from start to finish. It tackles the long and complex workflow involved in getting a paper accepted at a top computer science conference, breaking that process into seven coordinated AI agents that each handle a distinct phase. The pipeline begins with an Idea Agent that searches for trends in a research area, identifies gaps in existing work, generates candidate ideas, and checks whether they are genuinely new. A Literature Agent then retrieves and categorizes related papers from multiple sources, deduplicates them, and drafts a related work section with verified citations. A Method Agent designs the technical approach and produces pseudocode. An Experiment Agent plans datasets, evaluation metrics, ablation studies, and result tables. A Writing Agent assembles all sections into a complete draft. A Figures Agent scans the draft for placeholder markers and generates SVG diagrams, then collects them into a PowerPoint file. Finally, a Review Agent simulates multiple peer reviewers and produces a meta-review, and a Rebuttal Agent drafts a formal rebuttal letter responding to reviewer comments point by point. Progress is saved in a state file so work can be paused and resumed at any phase. The tool supports LaTeX output formatted for NeurIPS, ICML, ICLR, and CVPR conference templates, and includes scripts to convert Markdown to LaTeX and compile to PDF. You would use it if you are a researcher who wants AI assistance navigating the full arc from research idea to submission-ready paper. It runs on Python 3.9 or later and requires LaTeX to be installed.