What does this skill do, and when should you use it?
nature-academic-search is a skill for academic workflows that coordinates multi-source literature search, citation verification, strict other-citation auditing, article-level citation metric tables, high-impact citer profiling (academy members, presidents, talent-award holders, fellows), MeSH search strategies, and citation file management (.nbib/.ris/.bib conversion) via MCP tools (PubMed, CrossRef, arXiv, Scopus, ScienceDirect). It uses a router architecture that detects the user's workflow and loads the matching workflow fragment, preferring to load fragments from disk rather than relying on memory. The skill emphasizes strict independent other-citation determination and supports both English and Chinese triggers.
Calls MCP tools to search literature across PubMed, CrossRef, arXiv, Scopus, and ScienceDirect. Verifies citations extracted from a document, performs strict independent other-citation audits, excludes self-citations, and builds article-level citation metric tables. Profiles high-impact citers (academy members, presidents/deans, talent-award holders, fellows, field leaders) and extracts the context in which they cited the target paper. Constructs MeSH/PubMed search strategies. Converts and manages .nbib/.ris/.bib citation files. Manages references, including BibTeX generation, related-article discovery, and ID conversion. Provides a no-MCP fallback discovery search script (academic_search.py) and a format converter script (format-converter.py).
- Graduate students need to find literature across multiple sources on a topic and build a MeSH search strategy for a systematic review in PubMed.
- Authors want to verify citations in their manuscript, ensuring DOIs are valid and that volume/page details match, and to flag citation errors.
- Researchers want to count citations and strict other-citations for their paper and check whether any citers are academic elite such as academy members or fellows.
- Scholars need to convert .nbib files to .ris or .bib to import into different reference managers.
- Editors reviewing a manuscript verify the authenticity and format of references and generate a summary table of citation metrics.
How do you install this skill?
- Install script directly modifies ~/.claude/.mcp.json and settings.json; review and backup these files before running.
- Core functions depend on overseas services (PubMed, CrossRef, etc.); access from mainland China may fail or be slow; verify availability and consider fallbacks.
- Publisher not enterprise-verified; rely on actual execution results, not brand trust.
- Shell / CLI
- Local filesystem
- MCP Server
PythonMCP server (PubMed, CrossRef, arXiv, Scopus, ScienceDirect)
Use npx skills to install the whole collection: npx skills add Yuan1z0825/nature-skills --global --agent codex --skill '*' --yes --copy. Or for Claude Code, clone the repository to a stable path and create a subagent or slash command wrapper pointing to skills/nature-academic-search/SKILL.md. Install Python dependencies: python -m pip install -r skills/nature-academic-search/mcp-server/requirements.txt.
tmp="$(mktemp -d)"
git clone --depth 1 https://github.com/Yuan1z0825/nature-skills.git "$tmp"
mkdir -p ~/.claude/skills
cp -R "$tmp/skills/nature-academic-search" ~/.claude/skills/
rm -rf "$tmp"Generated from the source repository and skill path; it copies only this skill's folder. If the author's install steps above differ, follow those first. To scope it to one project, replace ~/.claude/skills with that project's .claude/skills.
How do you use this skill?
Once installed, send your agent any of these to trigger it:
- Compile citation counts, strict other-citation counts, and DOI for this paper, and check if any citers are academy members, fellows, or field leaders.
After installation, open a new session and describe your need in English or Chinese, e.g., 'Compile citation counts, strict other-citation counts, and DOI for this paper, and check if any citers are academy members, fellows, or field leaders.' The skill will detect the workflow and load the relevant instructions. Ensure MCP services are configured with PUBMED_EMAIL and credentials for Scopus/ScienceDirect if needed.
What are this skill's strengths and limitations?
- Multi-source coverage across PubMed, CrossRef, arXiv, Scopus, ScienceDirect enhances literature recall.
- Strict independent other-citation audit excludes self-citations, improving reliability.
- High-impact citer profiling helps identify academic influence.
- Router architecture loads only needed workflows, saving context.
- Supports Chinese triggers, friendly to Chinese-speaking users.
- Provides script fallback for basic search without MCP.
- Relies on MCP services and requires configuring multiple API credentials (PubMed email, Scopus, ScienceDirect), which raises setup overhead.
- Scopus/ScienceDirect providers are not detailed in SKILL.md, requiring self-exploration.
- No mention of automated tests or real-case validation coverage, so reliability is unverified.
- Some features (like ID conversion) depend on external tools; fallback to manual generation can be time-consuming.
How does this skill compare with similar options?
Side by side with related skills; every score comes from the same FSRS standard.
| Skill | FS score | Stars | Last updated | License |
|---|---|---|---|---|
| Academic Search & Citation Audit this page | 58 · Recommended | ★ 47k | 3d ago | Apache-2.0 |
| Nature Literature Pipeline | 49 · Use with care | ★ 47k | 3d ago | Apache-2.0 |
| II-Commons Retrieval Skill | 56 · Use with care | ★ 16 | 4mo ago | Apache-2.0 |
| Paper Search | 51 · Use with care | ★ 61 | 6mo ago | MIT |
| Probability Analysis — Event Probability Analyst | 60 · Recommended | ★ 75 | 12d ago | MIT |
How did FollowSkills review this skill?
Evidence: Install script modifies user's .mcp.json and settings.json with clear output, requires email; depends on external MCP services (PubMed etc.), no malicious behavior found. Deductions: writes config files without user confirmation; data flow (external network calls) not fully disclosed; publisher unverified, limited trust evidence.
Evidence: CI workflow runs pytest for mcp-server, script tests exist. Static review cannot verify actual execution. Deductions: no test details or coverage; error handling descriptions generic; not up to reliability cap.
Evidence: README/SKILL.md clearly define use cases (multi-source search, strict other-citation audit), target users (researchers), inputs; triggers are rich and semantically clear. Deductions: no explicit non-fit scenarios; environment fit evidence limited; core depends on overseas services (PubMed, CrossRef) with reachability concerns from mainland China, though fallback mentioned.
Evidence: Complete bilingual READMEs with clear structure, boundaries, dependencies; Apache-2.0 license; version number in manifest. Deductions: no changelog or version history; maintenance responsibility unclear (publisher unverified); known limitations partially mentioned but may be incomplete.
Evidence: Outputs described (deduplicated table, citation files, audit table), typical requests and inputs clear. Static review cannot verify direct usability. Deductions: no execution evidence for result correctness; comparative benefit over manual not quantified.
Evidence: CI workflow and script tests exist, but no test results or coverage; claims MCP support but no independent validation. Deductions: static review cannot run tests; only author-provided docs, no cross-source corroboration.
Open a dimension to read why it scored that way
Evidence confidence：Low — Mostly static review, author material or a limited demo; useful for discovery, not high-risk decisions.
See the full review method →