
variomeanalytics/bioinformatics-agent-skills
Developer toolsこのページは AI により自動翻訳されています。技術的な詳細は原文をご確認ください。 原文を見る
MCP server for bioinformatics agent skills — query a knowledge graph of 78 bioinformatics workflows directly from Claude Code
概要
By — Agentic bioinformatics for wet-lab biologists. Important: The Skill Graph front end server and this repo are maintained by Claude Opus 4.7 (somewhat autonomously). Query pipelines mid-conversation. No install. No data download. Just point your AI coding agent at a URL. 🔗 skillgraph.pipette.bio 🔗 pipette.bio MCP servers only load at startup. Fully exit and reopen your session. Claude will call the MCP tools automatically and use the results in its answer. - — scanpy, seurat, variant-calling, chipseq, atacseq, gwas-analysis, admet-prediction, molecular-docking, metagenome-binning, and more - — each annotated with the data types that flow between steps (BAM, VCF, H5AD, FASTQ, etc.
README
以下は原文 README の抜粋です。原文を見る。
Bioinformatics Agent Skills
By Pipette.bio — Agentic bioinformatics for wet-lab biologists.
Important: The Skill Graph front end server and this repo are maintained by Claude Opus 4.7 (somewhat autonomously).
A hosted MCP server that gives Claude Code (and any MCP client) direct access to a knowledge graph of 78 bioinformatics workflows — single-cell, variant calling, RNA-seq, drug discovery, metagenomics, and more.
Query pipelines mid-conversation. No install. No data download. Just point your AI coding agent at a URL.
🔗 Live graph explorer: skillgraph.pipette.bio 🔗 Main site: pipette.bio
Quick start
1. Add the MCP server to Claude Code
claude mcp add --transport http --scope user skillgraph https://skillgraph.pipette.bio/mcp
Or edit ~/.claude.json manually:
{
"mcpServers": {
"skillgraph": {
"type": "http",
"url": "https://skillgraph.pipette.bio/mcp"
}
}
}
2. Restart Claude Code
MCP servers only load at startup. Fully exit and reopen your session.
3. Verify it’s connected
claude mcp list
You should see:
skillgraph: https://skillgraph.pipette.bio/mcp (HTTP) - ✓ Connected
4. Ask Claude anything
> What tools does the scanpy skill use?
> Find the shortest path from wgs-alignment to pathway-enrichment
> List all single-cell skills
> What comes after variant-calling in a typical pipeline?
> Search for CRISPR-related skills
Claude will call the MCP tools automatically and use the results in its answer.
What’s in the graph
- 78 bioinformatics skills — scanpy, seurat, variant-calling, chipseq, atacseq, gwas-analysis, admet-prediction, molecular-docking, metagenome-binning, and more
- 483 pipeline transitions — each annotated with the data types that flow between steps (BAM, VCF, H5AD, FASTQ, etc.), the number of supporting papers, and whether the edge was confirmed in a curated ground-truth graph
- ~300 tools across skills — Scanpy, Seurat, bwa, samtools, DESeq2, GATK, fastp, MultiQC, STAR, Salmon, Cell Ranger, AlphaFold, AutoDock, and many more
- Full SKILL.md documentation for each skill — triggers, required tools, input/output formats, workflow steps, critical rules, and output artifacts
Tools exposed by the MCP server
| Tool | What it does |
|---|---|
get_skill |
Returns full SKILL.md documentation for a skill, plus overview, tools, I/O, and connections |
list_skills |
Lists all available skills, optionally filtered by domain keyword |
search_skills |
Keyword search across skill IDs, trigger keywords, and tool names |
get_transitions |
Upstream and/or downstream skills with edge evidence (paper counts, data types) |
find_path |
Shortest pipeline path between two skills, annotated with data flow |
get_graph_stats |
Summary statistics: total skills, edges, domain breakdown |
推奨ツール
別のキーワードを試すか、フィルタを外してください。
インストール
npx skillfish add variomeanalytics/bioinformatics-agent-skills