biomcp-research

Conducts biomedical literature and variant research with the BioMCP CLI and files tool gaps as repo issues.

622|113|Updated Apr 1, 2025
One-click install
npx skills add https://github.com/genomoncology/biomcp --skill biomcp-research
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: biomcp-research
Source: https://github.com/genomoncology/biomcp/tree/main/sdlc/skills/biomcp-research
Command: npx skills add https://github.com/genomoncology/biomcp --skill biomcp-research

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Biomedical research requires pulling evidence from many disconnected sources (PubMed, ClinVar, registries, guidelines), and gaps in research tooling are discovered during sessions but rarely recorded. This Skill runs a disciplined research loop with the BioMCP CLI while capturing bugs and feature gaps as structured issues in the biomcp repository.

Core Features & Use Cases

  • Guided research loop: Start with broad web search, confirm records with biomcp get article and citation walking, then verify against primary sources like ClinGen and Europe PMC.
  • Context discipline: Teaches bounded reading of cached full texts and filtered --json output so large papers do not flood the session context.
  • Evidence verification: Cross-checks joined cross-entity data against upstream sources (MyVariant, Europe PMC, NCBI) before trusting it, and reports negative results honestly.
  • Structured feedback filing: Writes bugs and feature requests as kebab-case issue files with severity levels, reproduction evidence, and root-cause analysis for later triage.
  • Use Case: While investigating a PTEN variant classification, use the Skill to pull the paper record via BioMCP, verify assertions in the ClinGen Evidence Repository, and file a feature issue documenting that filtering allele frequency is not reported by any existing command.

Quick Start

Ask the assistant to research a gene, variant, or paper using BioMCP and to file any tool limitations it encounters as issues in the biomcp repo.

Frequently Asked Questions about biomcp-research

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I research a biomedical paper with the BioMCP CLI?

Start with a web search to find what exists, then use `biomcp get article <pmid>` for citable metadata and the abstract. Walk outward with `search article`, `article citations`, and `article references` to build an evidence map from the seed paper.

How do I report a bug or feature request for BioMCP?

Write one markdown file per issue in `repos/biomcp/sdlc/issues/` with a kebab-case filename, an H1 title, and a Severity line. Bugs need reproduction, observed output, and root cause; feature requests use a `feature-` prefix and lead with the question you could not answer.

Does BioMCP full text output print the article content?

No, `get article <id> fulltext` prints a path to a cached file, not the text itself. Read the abstract first, then grep the cached file or read a bounded range only when you genuinely need most of it.

What open biomedical sources does BioMCP not cover?

Key gaps include the ClinGen CSpec registry for gene-specific classification rules, the ClinGen Evidence Repository for expert panel assertions, Europe PMC for open-access checks, and gnomAD documentation via its GitHub source markdown. All are keyless JSON APIs.

Why can BioMCP cross-entity data be wrong?

Cross-entity results are joined from several upstreams, so a protein change and cDNA change on one line may come from different transcripts. Verify facts against a second source or query upstreams like MyVariant and NCBI directly before trusting them.