Skills¶

Each skill is a directory under skills/ with a required SKILL.md. The frontmatter gives the name and short description; the body gives instructions, inputs, outputs, quality gates, and references. Optional subdirectories hold examples, scripts, tool notes, literature summaries, and reusable assets.

The table below is a reader-facing map. The source SKILL.md files are still canonical.

Bioinformatics and Scientific Analysis¶

Skill	Use when	Main result
`bioinformatics-project`	Starting, reorganizing, or reproducibility-hardening a bioinformatics project.	Project layout, pinned environment, task records, hypothesis register, restartable drivers, provenance, and sharing metadata.
`bio-foundation-housekeeping`	Defining metadata schemas or adding a queryable data catalog.	Generated LinkML/Pydantic models, cross-record validation, normalized Parquet tables, fixtures, and a DuckDB catalog.
`exploratory-data-analysis`	Inspecting an unfamiliar scientific data file before choosing a workflow.	A Markdown report covering file type, structure, quality issues, and downstream analysis options.
`bio-reads-qc-mapping`	Ingesting raw reads, trimming or filtering them, and mapping reads to references or assemblies.	QC reports, cleaned reads, alignments, and coverage summaries.
`bio-assembly-qc`	Building genome, metagenome, or long-read assemblies.	Assemblies with QC metrics and assembly-level interpretation.
`tracking-taxonomy-updates`	Reconciling taxonomy across NCBI, GTDB, ICTV, eukaryote frameworks, and QuickClade-first domain triage.	Versioned taxonomy evidence and domain-specific routing decisions.
`bio-binning-qc`	Recovering and refining metagenomic bins.	MAG/bin sets with completeness, contamination, and chimerism checks.
`bio-gene-calling`	Calling genes and basic features in prokaryotic, viral, or eukaryotic sequence sets.	Predicted CDS, protein FASTA files, GFF annotations, and feature summaries.
`bio-annotation`	Assigning function and taxonomy from sequence homology.	Functional annotations, taxonomy calls, and evidence-ranked hit tables.
`bio-fasta-database-curator`	Preparing sequence databases for BLAST, DIAMOND, MMseqs2, HMMER, pyhmmer, or custom reference searches.	Curated FASTA/FAA files, stable headers, deduplicated records, mapping tables, and database statistics.
`bio-phylogenomics`	Building marker-gene or protein alignments and trees.	Alignments, phylogenetic trees, topology checks, and interpretation.
`bio-interdomain-hgt`	Testing interdomain horizontal gene transfer and donor direction.	Homology, context, contamination, and per-gene phylogenetic evidence for candidate transfers.
`bio-protein-clustering-pangenome`	Clustering proteins into orthogroups or building pangenome matrices.	Orthogroups, copy-number matrices, and core/accessory gene summaries.
`bio-structure-annotation`	Adding structure-based evidence to protein interpretation.	Predicted or searched structures, fold-level annotations, and confidence notes.
`bio-viromics`	Detecting, classifying, and QCing viral contigs.	Viral calls, quality summaries, taxonomy evidence, and candidate discovery tables.
`bio-stats-ml-reporting`	Aggregating results, training models, or preparing final analysis reports.	Validated statistics, model outputs, plots, and a reproducible report.
`bio-prefect-dask-nextflow`	Designing executable bioinformatics workflows.	Prefect+Dask or Nextflow scaffolds with clear execution boundaries.
`bio-workflow-methods-docwriter`	Turning workflow artifacts into a Methods section.	Reproducible Methods text with commands, versions, parameters, QC gates, and outputs.
`bio-logic`	Auditing scientific reasoning, study design, bias, or strength of evidence.	A structured critique with uncertainty, alternative explanations, and follow-up checks.
`jgi-lakehouse`	Querying JGI Lakehouse, GOLD, IMG, Mycocosm, or Phytozome data.	SQL-backed metadata pulls and, when allowed, downloaded IMG genome files.

Literature, Metadata, and APIs¶

Skill	Use when	Main result
`polars-dovmed`	Searching PMC Open Access and bioRxiv with structured literature queries.	Source-backed literature summaries when users provide API access or local parquet corpora prepared with upstream `polars-dovmed`.
`arxiv-search`	Searching recent or specific arXiv preprints.	Local Markdown summaries from arXiv metadata and IDs.
`biorxiv-search`	Searching bioRxiv preprints by keyword, date range, DOI, or author.	Filtered preprint shortlists with bioRxiv-native metadata.
`crossref-lookup`	Validating DOIs or matching titles to citation metadata.	Crossref records, DOI matches, and bibliography cleanup evidence.
`scientific-impact-assessment`	Comparing papers, journals, or literature shortlists by influence.	OpenAlex citation counts, optional Altmetric context, and impact summaries.
`pdf-to-md`	Converting papers, PDFs, or office documents into analysis-ready Markdown.	Clean Markdown and, for papers, structured article and section-audit artifacts.

Writing, Review, and Evaluation¶

Skill	Use when	Main result
`scientific-writing`	Drafting, revising, or reviewing manuscripts, response letters, and grounded scientific prose.	A claim-safe draft or revision plan that stays tied to supplied evidence.
`csag-extraction`	Converting manuscripts into structured Conditional Scientific Argumentation Graphs.	Schema-valid claim/evidence/inference graphs with TextSpan grounding, validation reports, and paper-grounded Q&A items.
`manuscript-review-council`	Running a multi-angle manuscript review.	Parallel specialist reviews, disagreement checks, and an editor synthesis.
`proposal-review`	Reviewing grant, project, or funding proposals.	A decision-ready critique of strengths, risks, missing evidence, and fundability.
`ai-scientist-evaluator`	Evaluating finished outputs from one or more AI scientists.	A scored audit of rigor, reproducibility, novelty, task completion, and publication readiness.

Visualization and Notebooks¶

Skill	Use when	Main result
`notebooks`	Building or converting reproducible marimo or Jupyter notebooks.	A fully executed notebook with embedded figures and clear analysis flow.
`beautiful-data-viz`	Producing publication-quality matplotlib or seaborn figures.	High-data-ink static figures with readable axes, direct labels where feasible, tight layout, and appropriate palettes.
`plotly-dashboard-skill`	Building interactive Plotly Dash dashboards.	A themed Dash layout with callback performance guidance.

Skill Source Links¶

The source directory for every skill is available under skills/. When editing a skill, keep the SKILL.md frontmatter name equal to the directory name and rebuild the catalog afterward.