Academic Research Skills
Imbad0202/academic-research-skillsReviewed by AI
Checked by RepoPicks Radar on
What is Academic Research Skills?
Academic Research Skills (ARS) is a pack of skills, slash commands and prompts you install inside Claude Code. It guides the whole path of an academic paper: searching the literature, screening studies, planning the argument, drafting, checking every citation, running a simulated review panel and handling reviewer comments. It stops for your confirmation at each stage and the author stays you. It is text and prompts, not a program: you need Claude Code (or a paid claude.ai plan) and you pay Anthropic for the tokens, roughly US$3-7 for a 15,000-word paper by the project's own estimate.
How easy is it to install Academic Research Skills?
B · installer plus one simple step · Web
- Windows C
- macOS C
- Linux C
- Web B
On claude.ai (paid plans) you only add the repository under Customize → Plugins and install it, so the web route is B — but the README warns that hooks and plugin agents do not run in plain chat there. On Windows, macOS and Linux it is C: you first install the Claude Code command-line tool, set an ANTHROPIC_API_KEY and then run `/plugin marketplace add` and `/plugin install`; Pandoc, tectonic and Python are extra installs only if you want DOCX or PDF output.
Academic Research Skills on GitHub
- Stars
- 56,350on Nov 10, 2026
- Downloads
- —all releases
- Growth
- +1,132stars in 7 days
Stars over time
Facts
- Runs on
- Windows, macOS, Linux, Web
- You need to install
- Nothing extra
- To build from source
- Python, Node.js
- Disk space
- 200 MB–800 MB · estimate
- Accounts or keys
- Anthropic (an API key, or a paid Claude plan)
- Cost
- Free, API costs extra · The skills themselves are free but non-commercial only (CC BY-NC 4.0). docs/PERFORMANCE.md estimates about US$3-7 of Anthropic tokens for a 15,000-word paper at 2026-09 list prices, before cache discounts.
- License
- NOASSERTION (Other)
- Last change on GitHub
- Nov 10, 2026
Academic Research Skills risk signals (7)
- Attentionnoncommercial license only
LICENSE is Creative Commons Attribution-NonCommercial 4.0 and package.json says "license": "CC-BY-NC-4.0": "NonCommercial — You may not use the material for commercial purposes." Consultancy work, company research or anything you are paid for is outside the licence, and this is not an approved open-source licence.
AI review
- Attentionai citations still wrong after checks
README.md showcase table: the Stage 2.5 integrity report "flagged 15 problem references (8 with bibliographic errors, 6–8 likely fabricated)", and the Post-Publication Audit found that "21 of the 68 final references still had problems after 3 rounds of integrity checks". Every reference still has to be checked by hand.
AI review
- Attentionacademic integrity responsibility
README.md: "It can draft text, including a whole paper in full mode" and "You remain the author, and you answer for every claim you submit." The pack includes a "disclosure" mode for AI-use statements. Many journals and universities restrict or require disclosure of this kind of help.
AI review
- Infointegrity checks do not verify the research
README.md, "Integrity and verification boundary": "ARS does not establish that procedures were actually performed, raw data are authentic, or results reproduce; a consistently reported fabrication can pass these checks." Some checks are described as sampled or LLM-mediated.
AI review
- Inforeview panel not calibrated
README.md, Academic Paper Reviewer: "Current live reviews and Schema 6 packages remain NOT_CALIBRATED", "No numerical total is mapped to Accept, Minor Revision, Major Revision, or Reject", and the claim-audit calibration test "drives the runner with a stub judge that returns the gold labels, so it checks the tooling, not a live judge".
AI review
- Infosends your text to outside services
README.md points to docs/DATA_FLOWS.md for "what leaves your machine (bibliographic resolvers, optional consent-gated cross-model calls, the plugin update check)". Besides Anthropic, parts of your work can go to citation databases such as Semantic Scholar, and optionally to a second AI model for cross-checking.
AI review
- Infodeclared alternative false match
The automatic check read "invideo-ai" out of a README paragraph that is actually about AI-written peer reviews ("the AI review won 62% of decided pairs without web retrieval"). This project has nothing to do with video, so that paid product is not listed as an alternative here.
AI review
Evidence is shown in its original language.
What Radar could not check about Academic Research Skills
- Whether the drafts and reviews are good enough for a real journal submission
- How much the full pipeline really costs in tokens for your own paper
- Which parts work in each install channel (the project itself ships docs/CONTROL_AVAILABILITY.md because availability varies)
- Whether the optional Python scripts, Pandoc and tectonic steps work on Windows
- What exactly leaves your machine during a run