mskcc-omics-workflows skills

Shareable Agent Skills for working with the mskcc-omics-workflows organization. Each skill is a directory with a SKILL.md entry point and optional reference/ files loaded on demand (progressive disclosure).

Available skills

Skill	Purpose
contributing-msk-workflows	Contribute modules, subworkflows, containers, and test data to the org — or debug an existing msk module/subworkflow.

Installing a skill

Skills are loaded from a directory your agent can read. For Claude Code, copy a skill into your skills directory:

git clone https://github.com/mskcc-omics-workflows/skills.git
cp -r skills/contributing-msk-workflows ~/.claude/skills/

The skill then activates automatically when its description matches your task.

Structure

contributing-msk-workflows/
├── SKILL.md              # overview, repo map, end-to-end flow, pitfalls
└── reference/
    ├── containers.md     # Dockerfiles, OCI labels, JFrog/GHCR CI
    ├── modules.md        # main.nf, meta.yml, environment.yml, tags.yml
    ├── subworkflows.md   # chaining modules, components, channel shapes
    ├── test-data.md      # branch-per-tool model, registering URLs
    ├── nf-test.md        # module/subworkflow tests, stubs, snapshots
    ├── nf-core-tools.md  # install/test/scaffold/lint via nf-core
    └── bug-fixes.md      # patch an existing msk module upstream, then sync into a pipeline

SKILL.md stays under 500 lines; each reference file is self-contained and linked one level deep, so an agent loads only the phase it's actually working on.