Agent Skills for Healthcare and Life Sciences
43 skills for healthcare and life sciences: genomics, clinical data, drug discovery, CDISC and claims analytics
- 76 ★ on the repository
- Last commit
- MIT-0
Minimal access
Why this level
- The skills describe methodology and commands and run nothing themselves, but an agent follows them to run bioinformatics tools and handle medical data
Install
- Works in
Manual install
npx skills add awslabs/hcls-agent-skillsFor Claude Code, Cursor, Codex and other agents. Kiro and Claude Code also have `./install.sh --target <platform>` from a local copy.
Access and checks
How we reviewMinimal accessThe tool runs nothing and makes no network calls.
- The skills describe methodology and commands and run nothing themselves, but an agent follows them to run bioinformatics tools and handle medical data
This is third-party code. Review the repository files before installing.
What it does
A collection of 43 skills for healthcare and life sciences workflows, written to the open Agent Skills standard. Skills come in two kinds. Reasoning skills encode methodology and decision frameworks: ACMG variant classification, target trial emulation, HEDIS measure specification. Pipeline skills hold commands and parameter tables for runnable artifacts: GATK4 variant calling, RDKit cheminformatics, HL7v2 parsing. The domains cover genomics, single-cell analysis, medical imaging, protein structure, translational research, real-world-data pharmacoepidemiology, clinical data, drug discovery, proteomics, clinical data review, multi-omics, healthcare operations, regulatory and AWS architecture. Among the skills are `variant-calling`, `rna-seq-analysis`, `scrna-seq-pipeline`, `molecular-docking`, `dicom-processing`, `cdisc-compliance`, `risk-adjustment` and `claims-billing-rules`. In a 410-prompt comparison judged by Claude Opus 4.7, skills win 70–86 percent of head-to-head matches depending on the harness. Install with `npx skills add awslabs/hcls-agent-skills` or with the `./install.sh --target <platform>` script for Kiro and Claude Code.
Who it is for. Bioinformaticians, epidemiologists, clinical data and claims analytics specialists who work with an AI agent and want vetted methodology instead of general model knowledge.
Good fit when
- A variant needs classifying by ACMG criteria with ClinVar evidence
- A variant calling or RNA-seq pipeline must be built with BWA-MEM2, GATK4 and DESeq2
- Clinical data is brought to CDISC standards: SDTM, ADaM and define.xml
- RAF must be computed from claims and HCC coding gaps found
- DICOM files are converted to NIfTI, de-identified and arranged into BIDS
Not a fit when
- You need a clinical decision about a patient: the set helps with methodology and the decision stays with a specialist
- The task is outside the listed domains: the skills are narrowly specialized
- You need a single runtime: skills describe procedures and commands and run nothing themselves
Example request
Classify this variant by ACMG criteria, gather ClinVar evidence and write up a conclusion with confidence levels.Limitations
The skills describe methodology and commands, while running them and verifying results stay with the user and their environment. Pipeline skills need the matching tools: GATK4, STAR, Scanpy, RDKit, AutoDock Vina, TIAToolbox and others. Skills with codes and payment rules should be checked against current versions, since HCC coefficients and criteria change. In regulated domains a specialist verifies the conclusions. The `install.sh` script targets Kiro, Claude Code and Quick Desktop, and other platforms use `npx skills add`.
How to disable. Remove the installed skills with `npx skills remove` or delete the folders from `.claude/skills/` or `~/.kiro/skills/`. The agent config in `~/.kiro/agents/` is removed by hand.
FAQ
How do I install the skills?
With `npx skills add awslabs/hcls-agent-skills` for most agents or `./install.sh --target claude-code` and `./install.sh --target kiro` from a local copy.
How do reasoning skills differ from pipeline skills?
Reasoning skills set methodology and decision criteria, for example ACMG and target trial emulation. Pipeline skills give commands and parameters, for example GATK4 and RDKit.
Which domains are covered?
Genomics, single-cell, imaging, proteins, pharmacoepidemiology, clinical data, drugs, proteomics, multi-omics, operations, regulatory and AWS.
Do the skills replace a specialist?
No. They give an agent methodology and commands, and conclusions in regulated domains are checked by a specialist.
Which platforms does it work with?
Kiro, Amazon Quick, Amazon Bedrock AgentCore, AWS Strands SDK, Claude Code, Codex, CrewAI, LangChain and more than twenty others.
Discussion
No questions or reviews yet.
Related
SkillsAround 200 skills for Claude Code on bioinformatics, chemistry and biostatistics, from RNA-seq analysis to molecular docking
AWS hcls-agents plugin gives Claude Code and Codex skills for genomics, clinical trials and biomarkers plus MCP servers.
17 skills for computational biology: RNA-seq, single-cell, protein structure, docking, literature and figures
166 skills for scientific work: bioinformatics, cheminformatics, clinical data, geospatial analysis and 100+ databases
37 skills for computational drug discovery: from target selection and docking to ADMET and the clinical record.
A template for data scientists: AGENTS.md, CLAUDE.md and skills for Python, SQL, Polars and model diagnostics.
README in short
Covers the two kinds of skills, domains with skill tables, results of the 410-prompt comparison, install by script and through the skills CLI and a multi-agent mode for Kiro.