What problem does it solve? Drafting ML conference papers is slow and error-prone: structuring a compelling narrative, meeting venue-specific formatting rules, and avoiding hallucinated citations all demand significant effort. This Skill guides the full workflow from a research repository to a submission-ready paper for NeurIPS, ICML, ICLR, ACL, AAAI, and COLM. ## Core Features & Use Cases - Structured Paper Drafting: Step-by-step workflows for abstract, introduction, methods, experiments, related work, and limitations, grounded in writing guidance from researchers like Neel Nanda, Sebastian Farquhar, and Zachary Lipton. - Citation Verification: Programmatic citation workflow using Semantic Scholar, CrossRef, and arXiv APIs to fetch verified BibTeX via DOI, preventing the ~40% error rate of AI-generated references. - Conference Templates & Checklists: Official LaTeX templates for six ML venues plus mandatory checklist references (NeurIPS 16-item checklist, ICML Broader Impact, ICLR LLM disclosure, ACL Limitations). - Use Case: You have a research repo with experimental results and need an ICLR submission. The Skill explores the repo, drafts each section iteratively, verifies every citation via API, and formats the paper with the iclr2026 template. ## Quick Start Use the ml-paper-writing skill to draft a NeurIPS paper from my research repository, starting with the abstract and introduction.
