SuperAPI AI Directory

pubmed-quick-search

General & other

View repository

Updated 10-16 09:30

About this skill

Quick literature search

Description

Use unified_search for a first round of literature exploration. It automatically analyzes the query, selects sources, merges and deduplicates, and is suitable for quickly confirming whether a topic has research, which representative articles exist, and whether it is worth doing a more systematic search.

Trigger conditions

  • "Help me find papers on..."
  • "Are there any articles about..."
  • "Let me quickly search for..."
  • 「search papers about...」
  • 「find articles on...」

Core tools

unified_search(
    query="remimazolam ICU sedation",
    limit=10,
    ranking="balanced",
    output_format="markdown"
)

Key parameters

  • query: Natural language, Boolean queries, or queries containing ICD codes are all fine
  • limit: Maximum number of items to retrieve from each source
  • sources: Can specify pubmed,openalex,semantic_scholar,europe_pmc,crossref
  • ranking: balanced, impact, recency, quality
  • filters: Comma-separated filter conditions
  • options: Additional flags, such as preprints and whether to deep search

For quick searches, prioritize natural language + filters. Only switch to the systematic search skill when you need to carefully assemble a query.


Most common examples

1. General topic search

unified_search(query="remimazolam ICU sedation", limit=10)

2. Clinical research from recent years

unified_search(
    query="diabetes treatment",
    limit=20,
    ranking="quality",
    filters="year:2020-2025, species:humans, clinical:therapy"
)

3. See the latest research

unified_search(
    query="lung cancer liquid biopsy",
    limit=15,
    ranking="recency"
)

4. See high-impact literature

unified_search(
    query="CAR-T therapy lymphoma",
    limit=15,
    ranking="impact"
)

5. Restrict sources

unified_search(
    query="CRISPR gene therapy",
    sources="pubmed,openalex,europe_pmc",
    limit=20
)

6. Include preprints

unified_search(
    query="COVID-19 vaccine efficacy",
    limit=20,
    options="preprints"
)

filters Quick reference

filters="year:2020-2025, age:aged, sex:female, species:humans, lang:english, clinical:therapy"

Available fields

  • year: 2020-2025, 2020-, -2025, 2024
  • age: newborn, infant, preschool, child, adolescent, young_adult, adult, middle_aged, aged, aged_80
  • sex: male, female
  • species: humans, animals
  • lang: english, chinese etc.
  • clinical: therapy, therapy_narrow, diagnosis, diagnosis_narrow, prognosis, etiology etc.

options Quick reference

options="preprints, shallow"

Common flags

  • preprints: Include arXiv, medRxiv, bioRxiv
  • all_types: Relax the peer-reviewed restriction
  • no_oa: Do not supplement OA links
  • no_analysis: Do not show query analysis
  • no_scores: Do not show scores
  • no_relax: Disable automatic relaxation on zero results
  • shallow: Disable deep search for faster speed

When the query is uncertain

Analyze first, do not blindly search directly:

analyze_search_query(query="remimazolam vs propofol for ICU sedation")

If you need MeSH, synonyms, or suggested queries, then use:

generate_search_queries(topic="remimazolam ICU sedation")

Common next steps after searching

  1. Read details:fetch_article_details(pmids="12345678")
  2. Find similar research:find_related_articles(pmid="12345678")
  3. See who cited it:find_citing_articles(pmid="12345678")
  4. Get full text:get_fulltext(pmid="12345678")
  5. Export citation:prepare_export(pmids="last", format="ris")

When not to stop at this skill

  • Need full MeSH and synonym expansion: switch to pubmed-systematic-search
  • It is a clinical comparison question: switch to pubmed-pico-search
  • Need full text, figures, PDF: switch to pubmed-fulltext-access