Create OmicsClaw-native skill scaffolds for new reusable workflows that are not yet represented in the current skill catalog.
Scanned 9/7/2026
Install to Claude Code
npx -y skills add mdbabumiamssm/LLMs-Universal-Life-Science-and-Clinical-Skills- --skill omics-skill-builder --agent claude-codeInstalls into .claude/skills of the current project.
Are you the author of Omics Skill Builder?
Add the live security badge to your README — it updates automatically with every re-scan.
[](https://www.skillsdirectory.com/skills/mdbabumiamssm-omics-skill-builder-llms-universal-life-science-an)More formats (shields.io, HTML) on the badges page.
---
name: omics-skill-builder
description: >-
Create OmicsClaw-native skill scaffolds for new reusable workflows that are
not yet represented in the current skill catalog.
version: 0.1.0
author: OmicsClaw
license: MIT
tags: [orchestrator, meta-skill, skill-scaffold, automation]
metadata:
omicsclaw:
domain: orchestrator
script: omics_skill_builder.py
requires:
bins:
- python3
env: []
config: []
emoji: "🛠"
homepage: https://github.com/TianGzlab/OmicsClaw
os: [macos, linux]
install:
- kind: pip
package: pyyaml
bins: []
trigger_keywords:
- create omicsclaw skill
- add new skill
- scaffold skill
- create reusable workflow
- 新增 skill
- 创建 skill
- 封装成 skill
allowed_extra_flags:
- "--request"
- "--skill-name"
- "--domain"
- "--summary"
- "--trigger-keyword"
- "--method"
- "--input-format"
- "--output-item"
- "--no-tests"
legacy_aliases: []
saves_h5ad: false
requires_preprocessed: false
---
# 🛠 Omics Skill Builder
You are **Omics Skill Builder**, the OmicsClaw meta-skill for turning a missing
analysis workflow into a reusable, repo-native skill scaffold.
## Why This Exists
- **Without it**: one-off analyses remain trapped in ad hoc prompts or notebooks.
- **With it**: OmicsClaw can create a new `skills/<domain>/<skill-name>/` folder with a valid `SKILL.md`, runnable entrypoint, test stub, and scaffold spec.
- **Why OmicsClaw**: the scaffold follows the project registry, `templates/SKILL-TEMPLATE.md`, and the CLI conventions already used by `oc run`.
## When To Use
- The user explicitly asks to **add**, **create**, **package**, or **persist** a new OmicsClaw skill.
- The user names a workflow that is not in the current skill catalog and wants it to become a reusable skill.
- A team member wants a starting scaffold before implementing the real science code.
- A previously successful `custom_analysis_execute` notebook should be promoted into a repo-native skill draft.
Do **not** use this for one-off analyses. For temporary analyses, use the
`web_method_search` + `custom_analysis_execute` fallback instead.
## What It Creates
1. `SKILL.md` generated from the OmicsClaw template structure.
2. A runnable Python entrypoint with `--input`, `--output`, `--demo`, `--method`, and `--species`.
3. A minimal `tests/` stub so the new skill has an executable validation hook.
4. `scaffold_spec.json` capturing the original creation intent.
## Required Decisions
Before finalizing a new scaffold, capture:
- **Skill name**: short, lowercase, hyphenated alias
- **Domain**: one of `spatial`, `singlecell`, `genomics`, `proteomics`, `metabolomics`, `bulkrna`, `orchestrator`
- **Summary**: one sentence describing the reusable workflow
- **Methods**: the major backends or algorithm names
- **Trigger keywords**: 3-6 routing phrases users might say naturally
## Workflow
1. Resolve whether the user wants a reusable skill or a one-off analysis.
2. Normalize the requested skill name and target domain.
3. Generate the scaffold under `skills/<domain>/<skill-name>/`.
4. If a successful autonomous notebook is supplied, reuse its Python code and copy the source notebook into `references/`.
4. Refresh the registry so the new skill is discoverable immediately.
5. Return the created file paths and the next implementation steps.
## CLI Reference
```bash
oc run omics-skill-builder \
--output output/skill_builder \
--request "Create a CellCharter spatial domains skill" \
--skill-name spatial-cellcharter-domains \
--domain spatial \
--summary "Spatial domain identification scaffold for CellCharter-based workflows." \
--method cellcharter \
--trigger-keyword "cellcharter domains"
# Promote the most recent successful custom notebook into a skill draft
oc run omics-skill-builder \
--output output/promoted_skill \
--request "Package the last successful autonomous analysis into a reusable OmicsClaw skill" \
--skill-name peak-detection-skill \
--promote-from-latest
```
## Example Queries
- "Add a new OmicsClaw skill for CellCharter-based spatial domains."
- "Please scaffold a reusable phosphoproteomics kinase activity skill."
- "把这个 workflow 封装成一个新的 OmicsClaw skill。"
## Output Contract
The scaffolded skill folder will contain:
```text
skills/<domain>/<skill-name>/
├── SKILL.md
├── <skill_name>.py
├── scaffold_spec.json
└── tests/
├── __init__.py
└── test_<skill_name>.py
```
## Guardrails
- Create a new skill only when the user explicitly wants a reusable artifact.
- Do not overwrite an existing skill directory silently.
- Keep the scaffold OmicsClaw-native: valid frontmatter, stable CLI flags, standard output artifacts.
Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.
No comments yet. Be the first to comment!