High-level research methodology skills for AI coding agents.
Not another tool list — each skill encodes a research workflow that takes months to develop.
The Six Skills
Tool-integrated requires external APIs · Methodology pure workflow, no dependencies
literature-search
Multi-engine academic paper search with adaptive engine selection across Semantic Scholar, arXiv, Tavily, Exa, Gemini, AMiner.
paper-reading
Three-level reading workflow (skim 2 min → read 10 min → deep 30 min) producing structured digests with hidden assumptions and reproducibility notes.
social-media-paper-triage
Extract paper recommendations from 小红书 / 微信公众号 / Twitter / Bilibili posts, find the original arXiv source, then triage for relevance before adding to the library.
related-work-survey
Systematic survey: define dimensions → search each axis → build a taxonomy → identify the gap → produce a positioning narrative for your Related Work section.
zotero-management
Structured library management via local + Web API: Inbox / Active Projects / Background / Reading Queue / Archive collections with project, status, priority tags.
academic-figure-generation
Publication-quality figures from method text via a multi-agent pipeline (PaperBanana). Camera-ready PNG/PDF for ACL, NeurIPS, ICML, EMNLP.
How the Skills Compose
A typical end-to-end research session uses four to six of these skills in sequence.
Install
Each skill is a self-contained folder with a SKILL.md following the open agent skills standard. Drop into any skills directory — auto-discovered, no restart.
# clone
git clone https://github.com/jxtse/scientific-research-skills.git
cd scientific-research-skills
# install all (user-level)
cp -r skills/* ~/.openclaw/skills/
# or pick specific
cp -r skills/literature-search ~/.openclaw/skills/
cp -r skills/paper-reading ~/.openclaw/skills/# clone
git clone https://github.com/jxtse/scientific-research-skills.git
cd scientific-research-skills
# user-level (all projects)
cp -r skills/* ~/.claude/skills/
# or project-level (single repo)
cp -r skills/* .claude/skills/# clone
git clone https://github.com/jxtse/scientific-research-skills.git
cd scientific-research-skills
# user-level (all repos)
cp -r skills/* ~/.agents/skills/
# or project-level
cp -r skills/* .agents/skills/Dependencies at a Glance
Methodology skills work out of the box. Tool-integrated skills require setup before they run.
| Skill | What to configure | Where to get it | Required? |
|---|---|---|---|
literature-search |
One of: TAVILY_API_KEY, EXA_API_KEY, GEMINI_API_KEY, AMINER_API_KEY |
Semantic Scholar + arXiv work with no key | Recommended |
paper-reading |
None | — | None |
social-media-paper-triage |
Jina Reader (no key) covers most URLs. For X / 小红书 / 微信: Agent Reach | Jina Reader / xreach / Agent Reach | Optional |
related-work-survey |
Inherits literature-search config | — | Configure search first |
zotero-management |
ZOTERO_API_KEY + ZOTERO_USER_ID |
zotero.org/settings/keys | Required |
academic-figure-generation |
Local PaperBanana deployment | See skill SKILL.md | Required |
export lines to your shell profile (not just the current session), and verify each key with a smoke test before moving on. Never write secrets to a git-tracked file.