# Search PubMed and Build Reproducible Queries

Biomedical researchers need precise, reproducible searches across a large literature database. This skill builds PubMed queries and guides E-utilities retrieval workflows.

## Install

```bash
npx skillstore add davila7/pubmed-database
```

## Metadata

- Status: approved
- Slug: davila7-pubmed-database
- Skillstore revision: r2
- Version status: missing
- Tree hash: a3236c83b3f5583f4e3a47785d7711df5f10d26c94a75851694d2315b0339006
- Author: davila7
- GitHub username: davila7
- License: MIT
- Repository: https://github.com/davila7/claude-code-templates/tree/main/cli-tool/components/skills/scientific/pubmed-database
- Ref: c43861a65bb95efcae259cd161c9d6f4dc7eec6f
- Supported tools: Claude, Codex, Claude Code
- Audit status: complete
- Agent install advisory: allowed
- Manual install advisory: allowed
- Artifact signature: available
- Audit attestation: unavailable
- Human verification: not\_verified
- Risk factors: network, env\_access, external\_commands
- Quality score: 76
- Public page: https://skillstore.pages.dev/skills/davila7-pubmed-database
- Manifest: https://skillstore.pages.dev/api/skills/davila7-pubmed-database/manifest

## Capabilities

- Constructs PubMed queries with Boolean operators, MeSH terms, field tags, dates, languages, and publication types.
- Provides request patterns for ESearch, EFetch, ESummary, EPost, ELink, EInfo, EGQuery, ESpell, and ECitMatch.
- Guides batched record retrieval with the NCBI history server and EPost.
- Finds articles using PMIDs, DOIs, authors, journals, dates, or partial citation details.
- Structures systematic review searches with PICO concepts, synonyms, MeSH terms, and study filters.
- Explains citation export formats, saved searches, related articles, and literature monitoring.

## Use Cases

- Design a systematic search: Translate a clinical question into documented MeSH, synonym, and study-design search blocks.
- Retrieve research records: Search PubMed and fetch abstracts or structured citations through an E-utilities workflow.
- Automate literature monitoring: Plan repeatable searches, batching, caching, exports, and alerts for an evidence pipeline.

## Prompt Templates

### Create a basic PubMed query

```
Create a PubMed query for [topic]. Include common synonyms, one relevant MeSH term, and a date range of [years].
```

### Refine results by study type

```
Refine this PubMed query for [study type], [population], [language], and [date range]: [current query]. Explain each added field tag.
```

### Build a PICO search strategy

```
Build a reproducible PubMed strategy for this PICO question: [question]. Separate concepts, synonyms, MeSH terms, and final Boolean combinations.
```

### Plan a batched E-utilities workflow

```
Design an E-utilities workflow for [query] and approximately [record count] records. Cover ESearch, history, EFetch batching, rate limits, caching, and errors.
```

## Limitations

- Executing searches requires network access to PubMed or the NCBI E-utilities service.
- API examples contain placeholders and require valid contact details, rate limiting, and optional NCBI credentials.
- PubMed often provides citations and abstracts, while full text depends on external access rights.
- Search strategies still require domain review before use in clinical or systematic review decisions.

## Best Practices

- Combine MeSH terms with free-text synonyms, then verify PubMed query translation before relying on results.
- Document the exact query, search date, filters, result count, and export format for reproducibility.
- Respect NCBI rate limits and use history, batching, caching, retries, and a descriptive User-Agent.

## Anti Patterns

- Do not treat one narrow keyword query as a complete systematic search.
- Do not place a real API key directly into shared prompts, files, logs, or examples.
- Do not fetch large result sets in one request or ignore NCBI rate-limit responses.

## Security Audit

- Audited at: 2026-07-23T15:43:36.706\+00:00
- Summary: The official NCBI URLs and requests are expected for a PubMed API skill, and every API key occurrence is a non-secret placeholder. The command-execution and reconnaissance alerts are Markdown formatting or ordinary guidance; no prompt injection, credential theft, or harmful intent was found.

## Stats

- Views: 501
- Downloads: 13
- Favorites: 0
- Popularity score: 0