We welcome skills from anyone working in bioinformatics, computational biology, or related fields.

Join the contributors community: t.me/ClawBioContributors

mkdir -p skills/your-skill-name/tests skills/your-skill-name/examples
cp templates/SKILL-TEMPLATE.md skills/your-skill-name/SKILL.md

Use templates/SKILL-TEMPLATE.md as the structural source of truth.

Edit SKILL.md with:

• Top-level YAML fields: name, description, and license
• metadata fields: version, author, domain, tags, inputs, outputs, dependencies, demo_data, and endpoints
• metadata.openclaw fields: runtime and routing metadata such as requires, always, emoji, homepage, os, install, and trigger_keywords
• Markdown body: follow the template sections and keep them aligned with the PR audit expectations in CLAUDE.md

If there is any discrepancy between documents, follow the template for structure and CLAUDE.md for audit requirements.

If your skill needs Python/R scripts, add them alongside the SKILL.md:

skills/your-skill-name/
├── SKILL.md           # Required
├── your_script.py     # Optional
├── tests/             # Encouraged
│   └── test_script.py
└── examples/          # Encouraged
    ├── input.csv
    └── expected_output.md

If your skill requires a conda environment, add an environment.yml using conda-forge as the sole channel and nodefaults to prevent fallback to the Anaconda defaults channel (which has commercial licensing restrictions):

channels:
  - conda-forge
  - nodefaults

# Confirm the runner is working
python clawbio.py list

# Test a registered skill through the real CLI
python clawbio.py run <registered-alias> --demo

# Test a new skill directly before registration
python skills/your-skill-name/your_skill.py --demo --output /tmp/your-skill-demo

Use the direct script path until the skill is registered in clawbio.py. Once you add the alias to clawbio.py, confirm it appears in python clawbio.py list and then validate it through python clawbio.py run <registered-alias> --demo.

If the skill includes tests, run:

python -m pytest skills/your-skill-name/tests/ -v

If you changed SKILL.md YAML frontmatter, regenerate the catalog:

python scripts/generate_catalog.py

Option A: Pull request to this repo

git checkout -b add-your-skill-name
git add skills/your-skill-name/
git commit -m "Add your-skill-name skill"
git push -u origin add-your-skill-name
# Open PR on GitHub

Option B: Submit to ClawHub Follow the ClawHub submission guide.

1. Local-first: No mandatory cloud data uploads. Network calls only for public databases (PubMed, PDB, UniProt).
2. Reproducible: Generate audit logs and reproducibility bundles.
3. One job well: Each skill does one thing. Compose via the Bio Orchestrator.
4. Documented: Include example queries, expected outputs, and dependency lists.
5. Safe: Minimal permissions. Warn before destructive actions. No hardcoded credentials.

• Skill folder: lowercase, hyphens (vcf-annotator, not VCF_Annotator)
• Python files: lowercase, underscores (equity_scorer.py)
• Skill name in YAML: matches folder name exactly

• Python 3.10+
• Type hints encouraged
• pathlib for all file paths
• No hardcoded absolute paths
• Tests with pytest

AI coding agents (Codex, Devin, Claude Code, Cursor, etc.) should follow the same workflow as human contributors, plus:

1. Read AGENTS.md for setup, commands, code style, and project structure
2. Read the target skill's SKILL.md before modifying any code
3. Use python clawbio.py list to verify skills still load after changes
4. Run python -m pytest -v to confirm all tests pass
5. Regenerate skills/catalog.json if you changed any SKILL.md YAML frontmatter: python scripts/generate_catalog.py

Treat this as a contributor summary, not as a second source of truth.

• templates/SKILL-TEMPLATE.md defines the canonical SKILL.md structure.
• CLAUDE.md defines the formal PR audit and conformance checklist.
• Every skill should cover Trigger, Scope, Workflow, Example Output, Gotchas, Safety, and Agent Boundary.
• Include synthetic demo data and support --demo whenever the skill has executable automation.
• Add tests for demo mode and the main expected path when the skill includes code.
• Keep dependencies, installation steps, inputs, outputs, and safety boundaries accurate and specific.
• Regenerate skills/catalog.json whenever you change YAML frontmatter.

If you are unsure whether a SKILL.md is PR-ready, defer to the checklist in CLAUDE.md.

If you are looking for something to build:

• GWAS Pipeline: PLINK/REGENIE automation
• Metagenomics Classifier: Kraken2/MetaPhlAn wrapper
• Pathway Enricher: GO/KEGG enrichment analysis
• Clinical Variant Reporter: ACMG classification
• Phylogenetics Builder: IQ-TREE/RAxML automation
• Proteomics Analyser: MaxQuant/DIA-NN wrapper
• Spatial Transcriptomics: Visium/MERFISH analysis

ClawBio skills are independently audited by clawbio_bench, a standalone safety, correctness, and honesty benchmark maintained by Sergey Kornilov (Biostochastics, LLC). The benchmark runs in CI on every PR and produces tamper-evident verdict artifacts.

If you find a scientific error in ClawBio, you can:

• Open an issue on this repo with the finding
• Add a test case to clawbio_bench (the external standard)
• Both (preferred)

We commit to triaging all externally reported findings and referencing the finding ID in fix commits.

• Sergey Kornilov (Biostochastics) for building the independent clawbio_bench audit suite and identifying critical correctness and honesty findings across multiple skills.

Open an issue or reach out via the repo discussions.