crazymsn/academic-shcolar-skills

Academic Scholar Skills is a suite of AI agent workflows for the full research-to-paper process. It is designed to run inside AI coding assistants: Claude Code, Codex, and OpenCode. Once installed, you invoke it through slash commands like /ars-plan or $academic-research-suite to move through stages of a research project, from finding literature to writing, revising, and formatting a paper. The suite is organized into four modules. The Deep Research module uses a 13-agent team to conduct literature reviews, systematic reviews, and meta-analyses. It includes guided Socratic dialogue, support for the PRISMA systematic review standard, and verification against the Semantic Scholar API. The Academic Paper module uses 12 agents to handle writing tasks: outlining, drafting, abstract writing, citation formatting, figure verification, disclosure statements, and revision coaching. The Academic Paper Reviewer module simulates peer review with 7 agents, including an editor-in-chief role and three dynamic reviewers, scoring submissions on a 0 to 100 rubric and producing a revision and resubmission traceability matrix. The Academic Pipeline module ties all of these together into a 10-stage orchestrator with checkpoints, claim verification, and a "Material Passport" that carries structured metadata about data access levels and reproducibility claims between stages. Installation in Claude Code is done through a plugin marketplace command. For Codex, a shell script installs a router skill that maps plain aliases like ars-plan and ars-lit-review to the suite. For OpenCode, you clone the repository and start the tool from that directory, it reads a configuration file that sets the default agent. This repository is described as a cleaned redistribution mirror of an upstream project. It removes binary showcase PDFs, GitHub Actions, and sponsorship metadata. The documentation is available in English, Simplified Chinese, Traditional Chinese, and Japanese. The license is CC BY-NC 4.0, which permits reuse for non-commercial purposes with attribution.