Academic Research Skills (ARS)

Academic Research Skills (ARS) is an open-source suite of academic research Skills by Cheng-I Wu (GitHub user Imbad0202), built mainly for Claude Code. It has 5 Skills: deep research, paper writing, multi-perspective peer review, a pipeline that runs from research to publication, and study screening for systematic reviews. After installing, you can start with /ars-plan, which walks you through your paper's structure in a question-and-answer dialogue. The README opens by calling AI the copilot: the pipeline stops for the researcher's confirmation at every stage, you remain the author, and you answer for every claim. It also says it isn't a tool for hiding that you used AI. What sets it apart is that citation checks and research-integrity checks are gates the pipeline can't skip, and the README is open about which checks haven't been calibrated against real results yet. It works in English and Traditional Chinese and can write bilingual abstracts.

Who it's for:graduate students and researchers writing papers, and teams running systematic reviews. You'll need to be comfortable with Claude Code.

Before you start: install it from the Claude Code plugin marketplace, or copy it into your Skill folder; for Word and PDF output you can add Pandoc and tectonic. The plugin installs two hooks, scripts that run automatically at session start and before tool calls: one announces that it's loaded and checks for updates, the other checks where sub-agents may write files and lets things through when its environment is broken, so it isn't a security sandbox. Citation checks send reference details to scholarly indexes such as Semantic Scholar, OpenAlex and Crossref; sending parts of your paper to another company's model for cross-review is optional and needs your consent each session. The license is CC BY-NC 4.0, for non-commercial use only. It doesn't guarantee that citations or conclusions are true, and it doesn't run experiments on its own. Model costs are separate; cost figures in the README are the author's estimates.

Based on the project's public documentation. Not tested by AI Inspiration World.