OpenParable
Skip to content
← Back to skills

Bio Workflow Methods Docwriter

Documents
bio-workflow-methods-docwriter

Bio Workflow Methods Docwriter is an Agent Skill for documents published in the majiayu000/claude-skill-registry repository on GitHub: a SKILL.md instruction file an AI agent can load to do this task.

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.

View on GitHub ↗
QUICK START

How to use this skill

Bring this guide into your coding agent with a prompt tailored to the tool you use.

  1. Open your project in Codex.
  2. Copy and paste the prompt below.
  3. Review files and risks before approving any action.
Prompt to paste
Inspect the Agent Skill at https://github.com/majiayu000/claude-skill-registry/blob/HEAD/skills/data/bio-workflow-methods-docwriter/SKILL.md and all supporting files. Report provenance, permissions, scripts, dependencies, and security risks. Propose an exact installation into .agents/skills/bio-workflow-methods-docwriter, but wait for my approval before writing files or running scripts.

Copying this prompt does not install or run the skill. Review third-party files before use.

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 NOT CAPTURED and add a “How to capture next time” note.
  • Prefer verbatim command scripts from the workflow engine (e.g., Nextflow .command.sh) over paraphrases.
  • 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 software_versions.yml, conda env export, container digest)
  • Run logs + QC reports

See: reference/evidence-checklist.md

Output artifacts (always produce all 3)

  1. Workflow Summary (top of doc): 5–12 lines, plain language + step bullets.
  2. Methods & Run Documentation (METHODS.md): detailed, step-by-step.
  3. Machine-readable Run Manifest (run_manifest.yaml): exact run metadata & steps (validate).

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):

  • PaperSummary record using schemas/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 run_manifest.yaml from evidence (no guessing)
  • 3) Validate manifest (fix schema errors)
  • 4) Draft METHODS.md (summary first, then detailed steps)
  • 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.”