Awesome Journal Skills (AJS)

What it is: A curated collection of AI "skills" (instruction packs) that help AI agents — specifically Claude Code — follow the writing and review conventions of specific academic journals. Think of it as a set of style guides, but machine-readable and designed to be used by an AI coding assistant.

Scale: The banner states 4,166 agent skills, 300 packs, and 744 journals/conference venues covered. A "pack" groups multiple skills for one journal; a "skill" is a single instruction set.

Coverage: Organized by 11 academic disciplines:

• Comprehensive/Interdisciplinary
• Economics & Business
• Social Sciences
• Humanities
• Math & Physical Sciences
• Life Sciences
• Medicine & Health
• Engineering & Technology
• Computer Science & AI
• Agriculture, Environment & Earth Sciences
• Sports Science

What's inside:

• Chinese top social-science journals — e.g., Economic Research (《经济研究》), Management World (《管理世界》), Social Sciences in China (《中国社会科学》), China Industrial Economics, Journal of World Economy, China Economic Quarterly, Accounting Research, Journal of Financial Research, and more. Each has its own dedicated skill pack.
• English top economics/management journals — e.g., American Economic Review, Quarterly Journal of Economics, Journal of Political Economy, Econometrica, Journal of Finance, Management Science, MIS Quarterly, and others, collected in a "breadth" pack covering 100 journals with journal-selection routing. Labels like Top-5, Top-3, FT50, UTD24 indicate tiers.
• International natural-science & medical flagship journals — a wall of 154 journals including Science, Nature, Cell, PNAS, NEJM, JAMA, The Lancet, Physical Review Letters, JACS, Environmental Science & Technology, plus many Nature sub-journals. Twelve of these (Science, Nature, Cell, Cancer Cell, PNAS, PNAS Nexus, NEJM, JAMA, The Lancet, PRL, JACS, ES&T) have "deep packs" with 12 skills each; the rest link to a breadth collection.
• Venue index — a machine-readable TSV file listing all 744 venues.
• Showcase — a folder with "automatic empirical Showcase" containing real execution evidence (i.e., proof the skills actually run).

How it's organized: The README is a navigation hub. Clicking a journal cover jumps to that journal's skill pack folder. A "coverage overview" section lists representative subfields per discipline. There's also a layout guide, a skill-pack list, a usage section, and a roadmap.

Who made it: Maintained by CoPaper.AI from Stanford REAP (Stanford's empirical methodology team). Powered by StatsPAI. Licensed under MIT. Available in English and Simplified Chinese.

How to use it: The README points to a "How to use" section (⚡ 如何使用) and links to copaper.ai, but the README excerpt doesn't detail the exact usage steps — it's implied you'd load the skill packs into Claude Code and point it at a journal to get journal-specific formatting/writing guidance.