Claude-skill-registry bio-workflow-methods-docwriter
Generate reproducible bioinformatics/data-science Methods + run documentation from workflow run artifacts (Nextflow/Snakemake/CWL), including a top-of-doc workflow summary, exact executed commands, tool versions, parameters, QC, and outputs.
install
source · Clone the upstream repo
git clone https://github.com/majiayu000/claude-skill-registry
Claude Code · Install into ~/.claude/skills/
T=$(mktemp -d) && git clone --depth=1 https://github.com/majiayu000/claude-skill-registry "$T" && mkdir -p ~/.claude/skills && cp -r "$T/skills/data/bio-workflow-methods-docwriter" ~/.claude/skills/majiayu000-claude-skill-registry-bio-workflow-methods-docwriter && rm -rf "$T"
manifest:
skills/data/bio-workflow-methods-docwriter/SKILL.mdsource content
Bio Workflow Methods Docwriter
Create publication-ready Methods + internal run documentation for a specific executed bioinformatics or data science workflow run.
When to use
Use this Skill when you need to document:
- exact workflow flow + steps that actually ran
- tool + reference versions (and/or container image digests)
- key parameters + QC gates
- outputs and how to reproduce the run
Hard rules (reliability)
- Never invent commands, tool versions, reference versions, or dataset accessions.
- If an item is missing from evidence, write
and add a “How to capture next time” note.NOT CAPTURED - Prefer verbatim command scripts from the workflow engine (e.g., Nextflow
) over paraphrases..command.sh - Keep user data private: redact tokens, credentials, and PHI.
Inputs (evidence package)
Ask for (or locate) an evidence package folder containing:
- Workflow engine artifacts (Nextflow/Snakemake/CWL)
- Pipeline config + params
- Software version artifacts (e.g., nf-core
, conda env export, container digest)software_versions.yml - Run logs + QC reports
See:
reference/evidence-checklist.md
Output artifacts (always produce all 3)
- Workflow Summary (top of doc): 5–12 lines, plain language + step bullets.
- Methods & Run Documentation (
): detailed, step-by-step.METHODS.md - Machine-readable Run Manifest (
): exact run metadata & steps (validate).run_manifest.yaml
Schema + validator:
- Schema:
schemas/run-manifest.schema.json - Validate:
python scripts/validate_run_manifest.py run_manifest.yaml
Optional structured literature record (only if asked):
record usingPaperSummaryschemas/bio-paper-schema.yaml
Workflow
Copy this checklist into your working notes and check off as you go:
- 1) Inventory evidence files (log what you have / what is missing)
- 2) Build
from evidence (no guessing)run_manifest.yaml - 3) Validate manifest (fix schema errors)
- 4) Draft
(summary first, then detailed steps)METHODS.md - 5) Run quality gates (versions, parameters, QC, outputs, reproducibility command)
Quality gates (must pass)
- Every major step has: purpose • inputs • outputs • command (or NOT CAPTURED) • tool+version (or NOT CAPTURED)
- Reproduction section includes: pinned pipeline revision • container/conda info • full run command • params/config paths
- “Workflow Summary” appears at the very top
Quick examples (triggers)
- “Write the Methods section for this Nextflow run directory, include exact commands and versions.”
- “Document this Snakemake pipeline run with a workflow summary at the top and a reproducibility appendix.”
- “Summarize this paper’s protocol into the PaperSummary schema.”