Skills pubmed-database
📦

pubmed-database

Safe 🌐 Network access

Search PubMed biomedical literature

Also available from: davila7

Researchers need efficient access to biomedical literature for systematic reviews and evidence-based research. This skill provides comprehensive documentation for PubMed queries including Boolean operators, MeSH terms, and E-utilities API access.

Supports: Claude Codex Code(CC)
📊 70 Adequate
1

Download the skill ZIP

2

Upload in Claude

Go to Settings → Capabilities → Skills → Upload skill

3

Toggle on and start using

Test it

Using "pubmed-database". Search for systematic reviews on machine learning in medical imaging published 2023-2024

Expected outcome:

  • Query: (machine learning[mh] OR deep learning[tiab] OR artificial intelligence[tiab]) AND (medical imaging[mh] OR radiology[tiab] OR diagnostic imaging[tiab]) AND systematic review[pt] AND 2023:2024[dp]
  • Result: 247 articles found
  • Top matches include:
  • - Deep learning in radiology: a systematic review (PMID: 38012345)
  • - Machine learning for medical image analysis: systematic review (PMID: 37987654)

Using "pubmed-database". Find clinical trials on metformin for type 2 diabetes with cardiovascular outcomes

Expected outcome:

  • Query: (metformin[nm] OR biguanides[mh]) AND diabetes mellitus, type 2[mh] AND cardiovascular diseases[mh] AND randomized controlled trial[pt] AND hasabstract[text]
  • Result: 89 articles found
  • Key studies:
  • - Metformin and cardiovascular outcomes in type 2 diabetes (PMID: 36543210)
  • - Long-term cardiovascular effects of metformin (PMID: 36123456)

Security Audit

Safe
v5 • 1/21/2026

All static findings are false positives. This is a documentation-only skill providing reference materials for PubMed queries. The detected patterns (backticks, API key placeholders, URLs) are markdown documentation elements, not executable code. The skill contains no executable code, no network calls, and no credential access.

5
Files scanned
5,088
Lines analyzed
1
findings
5
Total audits

Risk Factors

🌐 Network access (5)
references/api_reference.md:8 references/api_reference.md:295 SKILL.md:128 SKILL.md:453 SKILL.md:454
Audited by: claude View Audit History →

Quality Score

45
Architecture
90
Maintainability
87
Content
22
Community
100
Security
91
Spec Compliance

What You Can Build

Systematic Literature Review

Construct comprehensive search strategies using Boolean operators, MeSH terms, and filters to find all relevant studies for systematic reviews and meta-analyses.

Biomedical Research Queries

Search for articles on specific diseases, treatments, or drug interactions using precise field tags and controlled vocabulary.

Programmatic Data Extraction

Build automated workflows to search PubMed and extract article metadata using the E-utilities API for large-scale research projects.

Try These Prompts

Basic Article Search 
Find recent articles on {topic} published in 2024. Use appropriate MeSH terms and field tags to limit results to English language research papers.
Clinical Trial Search 
Search for randomized controlled trials comparing {treatment_a} versus {treatment_b} for {condition} from 2020 to 2024. Include only free full-text articles.
Systematic Review Query 
Use the PICO framework to construct a comprehensive search strategy for {clinical_question}. Include all relevant synonyms and MeSH terms.
API-Based Batch Search 
Write a Python script using the E-utilities API to search for articles matching {criteria}, retrieve up to 500 PMIDs using the history server, and export results to a CSV file with PMID, title, and abstract.

Best Practices

  • Start with broad searches using MeSH terms, then narrow with field tags and filters for precision
  • Include synonyms and alternative terms to capture all relevant literature
  • Use the history server (usehistory=y) for large result sets exceeding 500 articles
  • Obtain an API key from NCBI for higher rate limits (10 requests/second vs 3)

Avoid

  • Using only free-text terms without MeSH terms misses indexed articles
  • Running many small API requests instead of batch operations wastes rate limits
  • Not caching results leads to redundant API calls and slower workflows
  • Ignoring automatic term mapping can produce unexpected search results

Frequently Asked Questions

What is the difference between MeSH terms and text word searches?
MeSH terms search the controlled vocabulary index where all articles are manually tagged. Text word searches only match terms appearing in titles and abstracts. Combining both provides comprehensive coverage.
How do I get an API key for PubMed E-utilities?
Register for a free NCBI account at ncbi.nlm.nih.gov, then visit your account settings to generate an API key. Include it in requests with the api_key parameter for higher rate limits.
What are the PubMed rate limits?
Without an API key: 3 requests per second. With an API key: 10 requests per second. Always include a descriptive User-Agent header in your requests.
How do I handle large result sets with thousands of articles?
Use the history server: first call ESearch with usehistory=y to store results, then use the WebEnv and QueryKey parameters with ESummary or EFetch to retrieve data in batches of up to 500 records per request.
What export formats are available for PubMed results?
Common formats include MEDLINE (text), XML (structured data), RIS (reference managers), and NBIB (EndNote). Use the rettype and retmode parameters to specify your desired format.
How do I find articles that cite a specific paper?
Use the ELink utility with dbfrom=pubmed, db=pubmed, and cmd=cite. This returns PMIDs of articles that cite the input article. Note that citation data may not be complete for very recent publications.

Developer Details

Author

K-Dense-AI

License

Unknown

Repository

https://github.com/K-Dense-AI/claude-scientific-skills/tree/main/scientific-skills/pubmed-database

Ref

main

File structure

📁 references/

📄 api_reference.md

📄 common_queries.md

📄 search_syntax.md

📄 SKILL.md