Claude Skills
The open format is called Agent Skills and works in Claude Code, Codex, Cursor and other agents — most people know it as Claude Skills.
Every Agent Skill we could find on GitHub, deduplicated by content. 79 870 files from 1 769 authors, of which 62 217 are unique — the rest is the same skill repackaged into someone else's repository. For each one: what it weighs in tokens, whether it ships runnable scripts, and which MCP servers it needs.
Prepares statistical reports for clinical trials following CONSORT 2025, SPIRIT 2025, ICH E9(R1) estimands, and FDA 2023 covariate adjustment guidance. Covers Table 1 generation, analysis populations (ITT/FAS/PP/Safety), the 5 ICH E9(R1) intercurrent-event strategies, MMRM under MAR (mmrm), reference-based MI (rbmi J2R/CR/CIR), Permutt tipping-point sensitivity, and Rubin's-rules vs frequentist variance debate. Use when preparing regulatory submissions, defining estimands, or implementing missing-data sensitivity analyses.
Estimates circulating tumor DNA fraction from shallow whole-genome sequencing using ichorCNA. Detects copy number alterations via HMM segmentation and calculates ctDNA percentage. Requires 0.1-1x sWGS coverage. Use when quantifying tumor burden from liquid biopsy or monitoring treatment response.
Calculates tumor mutational burden from WES/WGS/panel data with Friends of Cancer Research harmonization equations, per-assay calibration (FDA 10/Mb = 7.8 TSO500 = 8.4 OncomineTML), synonymous/indel/germline filtering, hypermutator tiering, blood TMB, and integration with HLA-LOH and neoantigen quality (Luksza 2017 fitness). Use when assessing ICI eligibility under tumor-specific cutoffs (McGrail 2021), comparing tissue vs bTMB, or auditing TMB-H reporting against ESMO 2024 and FDA pembrolizumab pan-tumor 2020.
Import transcript-level quantifications from Salmon/kallisto into R for gene-level analysis with DESeq2/edgeR using tximport or tximeta. Use when importing transcript counts into R for DESeq2/edgeR.
Extract, process, and deduplicate reads using Unique Molecular Identifiers (UMIs) with umi_tools. Use when library prep includes UMIs and accurate molecule counting is needed, such as in single-cell RNA-seq, low-input RNA-seq, or targeted sequencing to distinguish PCR from biological duplicates.
Query UniProt's REST API (post-2022 endpoint at rest.uniprot.org) for protein sequences, annotations, GO terms, cross-references, ID mappings, and proteomes. Use when fetching UniProtKB entries, navigating the JSON schema, choosing between UniProtKB/UniRef/UniParc/Proteomes resources, deciding stream vs search endpoint for batch retrieval, running ID-mapping jobs with the async pattern, handling isoform suffixes, or filtering reviewed Swiss-Prot vs auto-annotated TrEMBL. Encodes the legacy URL migration (2022), the new JSON schema layout, and bulk-pull patterns.
Build UpSet plots to visualize set intersections beyond 4 sets (where Venn fails) using ComplexUpset (modern, ggplot2-grammar) or the unmaintained UpSetR, with explicit cardinality vs degree sorting, attribute panels, and query highlighting. Use when comparing overlap across many gene sets, peak sets, variant lists, or any set membership matrix where Venn diagrams become illegible.
Comprehensive variant annotation using bcftools annotate/csq, VEP, SnpEff, and ANNOVAR. Add database annotations, predict functional consequences, and assess clinical significance with MANE transcript selection and pathogenicity scoring. Use when annotating variants with functional and clinical information.
Call SNPs and indels from aligned reads using bcftools mpileup and call. Use when detecting variants from BAM files or generating VCF from alignments.
Normalize indel representation, decompose MNPs, and split multiallelic variants using bcftools norm. Use when comparing variants from different callers, preparing VCF for database annotation, or merging VCFs from multiple sources.
Prioritizes rare-disease variants from trio/quad WES/WGS with de novo (DeNovoGear, Triodenovo), compound-heterozygous phasing (WhatsHap), mosaic VAF tiering, phenotype-driven ranking (Exomiser, Phen2Gene, AMELIE), ClinGen gene-disease validity gating, and ACMG SF v3.2 secondary findings reporting. Use when running diagnostic exome / genome pipelines, identifying candidate Mendelian disease genes, screening for incidental findings, or auditing VUS reclassification cycles. The ACMG/AMP classification framework (PVS1 decision tree, Pejaver PP3/BP4 calibration, Tavtigian point system) is in clinical-databases/acmg-classification.
Assign pathogen lineages and track variants using Nextclade and pangolin for viral surveillance. Monitor variant prevalence and identify emerging variants of concern. Use when classifying viral sequences, tracking lineage dynamics, or monitoring for variants of concern.
View, query, and understand VCF/BCF variant files using bcftools and cyvcf2. Use when inspecting variants, extracting specific fields, or understanding VCF format structure.
Merge, concatenate, sort, intersect, and subset VCF files using bcftools. Use when combining variant files, comparing call sets, or restructuring VCF data.
Generate variant statistics, sample concordance, and quality metrics using bcftools stats and gtcheck. Use when evaluating variant quality, comparing samples, or summarizing VCF contents.
Calculate immune repertoire diversity metrics, compare samples, and track clonal dynamics using VDJtools. Use when analyzing repertoire diversity, finding shared clonotypes, or comparing immune profiles between conditions.
Performs structure-based virtual screening using AutoDock Vina, SMINA, GNINA (CNN scoring), and DiffDock-L hybrid workflows with explicit choice rules across rigid vs flexible docking, cross-docking vs self-docking, binding-site detection (P2Rank, fpocket), receptor preparation (PDB2PQR, PROPKA), ligand preparation (meeko, OpenBabel), and ultralarge-library screening (ZINC22, Enamine REAL). Use when screening chemical libraries against a protein target to find candidate binders, ranking docking poses, or selecting a docking workflow for a specific scenario.
Build volcano and MA plots from differential-expression / association results with LFC shrinkage, FDR-adjusted thresholds, sensible label placement, and axis-truncation conventions. Covers EnhancedVolcano, ggplot2, matplotlib, and the apeglm/ashr/normal shrinkage decision. Use when visualizing differential-expression results (RNA-seq, ChIP-seq, ATAC-seq, proteomics) or any per-feature effect-size + p-value table.
Create portable bioinformatics pipelines with Workflow Description Language (WDL) using Cromwell or miniwdl execution engines. Use when running GATK best practices pipelines, working with Terra/AnVIL platforms, or building workflows for cloud execution on Google Cloud or AWS.
Build whole-genome alignments using Progressive Cactus (Armstrong 2020 reference-free clade-level WGA), Minigraph-Cactus (Hickey 2024 pangenome-aware), LASTZ chain/net (UCSC pipeline), MUMmer4 (Marçais 2018 pairwise), minimap2 -x asm5/10/20 (Li 2018 fast pairwise), AnchorWave (Song 2022 WGD-aware), and Mauve / progressiveMauve (bacterial). Operates the HAL toolkit (Hickey 2013) for downstream extraction including halSynteny, halLiftover, halBranchMutations, and hal2maf. Use when constructing multi-species alignments for comparative-annotation projection (TOGA), synteny detection, conservation analyses (phyloP / PhastCons), or pangenome graph construction; selecting between reference-free (Cactus) and reference-anchored (LASTZ chains/nets) approaches; tuning sensitivity for closely vs distantly related genomes; or producing HAL files for genome-wide downstream tools.
Detect, date, and contextualize whole-genome duplication (WGD / paleopolyploidy) events using wgd v2 (Chen & Zwaenepoel 2024), KsRates (Sensalari 2022 substitution-rate-corrected Ks dating), DupGen_finder (Qiao 2019), MAPS (Li 2018 phylogenomic), POInT (Conant 2008 ordered-block), SLEDGe (2024 ML-based), Whale.jl (Bayesian DTL+WGD), and synteny-anchored paranome construction. Use when identifying ancient polyploidy from Ks distributions and synteny block analysis, positioning WGD events relative to speciation, distinguishing tandem from segmental from WGD duplications, dating the 2R/3R vertebrate / fish / salmonid WGDs, building paranome and Ks-age mixture models, applying KsRates substitution-rate correction across lineages, or testing alternative biased-fractionation / dosage-balance models post-WGD.
WikiPathways enrichment using clusterProfiler and rWikiPathways. Use when analyzing gene lists against community-curated open-source pathways. Performs over-representation analysis and GSEA for 30+ species.
Write biological sequences to files (FASTA, FASTQ, GenBank, EMBL) using Biopython Bio.SeqIO. Use when saving sequences, creating new sequence files, or outputting modified records.
XCMS3 workflow for LC-MS/MS metabolomics preprocessing. Covers peak detection, retention time alignment, correspondence (grouping), and gap filling. Use when processing raw LC-MS data into a feature table for untargeted metabolomics.
> Predict protein structure for monomers and multimers with AlphaFold2 via the ColabFold runner (Mirdita et al. 2022, github.com/sokrypton/ColabFold; AlphaFold2 Jumper et al. 2021). Reach for this skill to fold a sequence or complex with the AF2/AF2-Multimer evoformer, to validate designed sequences by self-consistency pLDDT, ipTM, and RMSD, or to run a quick MSA-backed prediction using the public MMseqs2 server.
> Structure prediction for protein, nucleic-acid, and small-molecule complexes with Boltz-2 (Passaro & Wohlwend et al. 2025, github.com/jwohlwend/boltz). Reach for this skill to validate designed binders against a target, to co-fold a protein with a SMILES or CCD ligand, or to get an open-source AlphaFold3 alternative with optional binding-affinity prediction.
> Predict genome-wide functional tracks (RNA-seq, CAGE, DNase, ChIP) from DNA (1) Scoring the regulatory effect of a variant on expression/accessibility, (2) Generating predicted coverage tracks for a locus, (3) Prioritising non-coding variants by predicted track delta.
> Structure prediction for protein, nucleic-acid, and small-molecule complexes with the Chai-1 foundation model (Chai Discovery 2024, github.com/chaidiscovery/chai-lab). Reach for this skill to predict an antibody-antigen or protein-ligand complex from a single FASTA, to re-fold designed binders as an AlphaFold-multimer alternative, or to drive co-folding from Python for batched campaigns on a GPU.
Set up a compute environment on a remote provider so Claude Science jobs can run there. Covers direct SSH/conda hosts, Slurm clusters, container-via-bridge runners, and managed-API providers (Modal, GCP, RunPod). Use when standing up a new provider, porting an env to a different backend, adding a tool that needs its own software stack, or wiring weight caches. Triggers on "new compute provider", "set up env on", "port env to", "build GPU image", "weight cache", "compute_details", "conda env on the box", "apptainer on slurm".
Create, configure, and maintain custom agent profiles and author new skills via the `repl` tool. Use when the user wants to create an agent profile, build a custom agent, modify agent capabilities, attach or detach skills/connectors on a profile, author a skill, or inspect which connectors and tools are available. Also use whenever you need the `host.agents.*` or `host.skills.*` Python SDK.
> Predict small-molecule binding poses with DiffDock-L (Corso et al. 2023/2024, github.com/gcorso/DiffDock) — blind diffusion docking that places a ligand into a protein pocket without a predefined search box and ranks the samples with a learned confidence model. Reach for this skill to dock a SMILES or SDF against a PDB, to generate ranked 3D poses for a small fragment library, or to get a starting pose for downstream rescoring. DiffDock predicts geometry, not affinity.
> Biohub ESMFold2 / ESMFold2-Fast all-atom co-folding (Candido et al. 2026, github.com/Biohub/esm). Single-sequence and MSA modes; protein, DNA, RNA, ligand (CCD/SMILES), modified residues. FoldBench Ab-Ag 50-55%, PPI 70-77% DockQ-pass. Also covers the ESMC-{300M,600M,6B} protein language models from prediction, and the SAE interpretability head. MIT-licensed weights on structures with single-sequence input, (2) Validating designed binders with ESMFold2-Fast, (3) Running ESMFold2 with MSA input, (4) Getting ESMC embeddings or per-residue mutation scores, (5) Choosing kernel backend and sampling-step settings for paper-faithful throughput.
> Score, embed, and generate DNA sequences with Evo 2, a long-context genomic (1) Computing per-nucleotide or per-sequence likelihoods for variant effect scoring, (2) Embedding genomic windows for downstream classification, (3) Generating DNA conditioned on a prefix, (4) Scoring regulatory or coding regions across species.
> Embed proteins with Meta AI's ESM-2 (`fair-esm` package). Use this skill ML, (2) Masked-LM likelihood / mutation effect scoring, (3) Contact prediction from a sequence.
Compose one publication-grade multi-panel figure. Entry from a one-line claim + data refs, OR from an existing figure via `derive_outline(png)`. Runs a per-figure loop: outline (12-col grid, per-panel ask + label_budget) → fan-out one sub-agent per panel (each loads `figure-style`) → tile + stamp letters → adversarial composite review with two-tier feedback (Tier-1 outline_revisions / Tier-2 per-panel violations) → regen affected panels, ≤3 rounds. Loads panel_task / compose_figure / compose_crops / composite_review_task / derive_outline into the kernel. For one standalone plot use `figure-style`; for whole-paper figure ordering use `paper-narrative`.
Publication-grade figure correctness and legibility rules. Load before drawing any plot and call `apply_figure_style()` — sets a role-mapped font-size ladder, outward ticks, frameless legends, and 300-dpi output. The skill is a checklist, not a house look: data fidelity (claim-titles tested against every row, excluded data never enters summaries), label economy (floor and ceiling), colour threading, chart-choice-by-data-shape, layout, and a render-then-verify QA loop (bbox collision + per-panel perceptual check). Ships helpers: focal_palette, bar_with_points, strip_with_median, end_of_line_labels, panel_letter, set_frame, panel_crops. For multi-panel figures load `figure-composer`; for whole-paper figure arc load `paper-narrative`.
> Generate a therapeutic indication dossier. Covers the patient population, epidemiology, disease biology, standard of care, regulatory precedent, and landmark clinical trials.
> Inverse-fold a backbone with ligand, nucleic-acid, and metal context using LigandMPNN (Dauparas et al. 2023, github.com/dauparas/LigandMPNN). Reach for this skill to redesign the residues lining a binding pocket around a bound small molecule or cofactor, to design metal-coordinating sites where the geometry must be respected, or to get threaded designed-sequence PDBs out of any MPNN run.
Find, verify, and synthesize scientific literature — from "what's the seminal paper for X" through full multi-source reviews. Covers grounding claims in real retrieved sources, avoiding fabricated citations, handling retractions, and calibrating confidence to evidence strength.
Register a model service in the managed family — a local model server container the daemon starts/stops on demand, or a remote upstream model API (https). Read the runbook, allocate a port (local only), compose idempotent start/stop scripts (local only), register once. Load when the user wants a model service available for inference, or when list_compute shows managed endpoints.
> Structure prediction using OpenFold3, an open-weights PyTorch reproduction of AlphaFold3 from the AlQuraishi Lab. Use this skill when predicting protein/nucleic-acid/ligand complex structures with an Apache-2.0-licensed AF3 reimplementation.
Judge and reshape the STORY a paper's figures tell. Input is the work itself — manuscript (or abstract) + figure deck — no hand-written brief. `derive_paper_brief(abstract, captions)` extracts pitch/vision/per-figure-claims; a handling-editor reviewer on the full deck returns hook_verdict (would Fig 1 make me send this for review?), arc (hook→mechanism→evidence→application), figure_moves (panels in the wrong figure), missing_panels (concrete analyses to RUN), kill_list, and boldest_defensible_fig1. Hands per-figure claims to `figure-composer`. Load when writing or revising a paper.
Use this skill when the user has attached a PDF, paper, report, or other document and the answer needs content from more than one place in it: summarize the methods or any other section, compare sections, find where a topic is discussed, read a value or label off a figure or chart, or find/list/extract every instance of something across the whole document (datasets, benchmarks, citations, figures, table rows, accession numbers — including appendices). Skip it only for a single lookup of 1–4 pages quoted in your very next response — `read_file(pages=[...])` attaches pages as images that are dropped from context after one turn, so multi-section answers end up re-reading the same ranges repeatedly. Parses the PDF once in the Python kernel: `pdf_pages` (pages as persistent text), `pdf_outline` (TOC), `pdf_scan` (rank pages by relevance), `pdf_map`/`pdf_extract` (per-page summary / structured fields via parallel haiku calls). For PDF creation/manipulation, use reportlab/pypdf directly.
> Inverse-fold a protein backbone (PDB structure) into amino-acid sequence with ProteinMPNN (Dauparas et al. 2022, github.com/dauparas/ProteinMPNN). Reach for this skill to run sequence design on RFdiffusion backbones, to redesign one chain of a PDB while holding interface residues fixed, or to generate a temperature-swept set of sequences for downstream folding.
Run GPU jobs on the user's own Modal account via host.compute.create('byoc:modal', ...). Covers the create→submit→wait_for_notification flow, the compute_provider kernel for env setup, image/volume resolution, and the two approval cards. Load once you've decided to dispatch to Modal.
Submit→wait_for_notification→harvest workflow for the user's SSH/SLURM hosts. Load once you've decided to dispatch remote.
> Embed and annotate single-cell expression data with scGPT, a foundation model (1) Producing cell embeddings from an AnnData for clustering/integration, (2) Zero-shot or fine-tuned cell-type annotation, (3) Gene-level representation for perturbation/GRN tasks. For probabilistic single-cell models (scVI etc.), use the scvi-tools library.
> Probabilistic single-cell RNA-seq with scvi-tools — scVI for a batch-corrected latent space, scANVI for semi-supervised label transfer, and Bayesian differential expression. Reach for this skill to integrate scRNA-seq batches, embed cells for clustering, transfer annotations from a reference onto a query, or score differentially expressed genes per cluster. For spatial deconvolution / mapping use the cell2location, DestVI, or Tangram methods instead.
Claude Science's own session database schema and SDK surface for introspection via host.query(). Load this when you need to query your own conversation history, token usage, cost accounting, execution log, or artifact metadata beyond what host.frames()/host.artifacts() provide — e.g. "how many tokens has this session used", "what was my last tool call", "list every file I've written", "where are messages stored", "what tables can I query", "inspect frames.context_data", or any time you're about to PRAGMA-probe the Claude Science metadata DB to discover its schema.
Create new skills, modify and improve existing skills, and measure skill performance. Use when users want to create a skill from scratch, edit, or optimize an existing skill, run evals to test a skill, benchmark skill performance with variance analysis, or optimize a skill's description for better triggering accuracy.
Call a registered model endpoint over its native HTTP API from the endpoint's scoped inference kernel (BASE_URL preloaded). Load once a task needs predictions from a registered model endpoint.
Unified analysis pipeline for affinity-based proteomics platforms — Olink (PEA, NPX) and SomaLogic SomaScan (SOMAmer, RFU). Platform-aware QC, normalisation, differential abundance, volcano plots, heatmaps, and PCA.
Analyze a single FASTA file (nucleotide or protein), compute sequence-level metrics (GC, ORFs, MW, pI, GRAVY, secondary-structure fractions) with Biopython, and write a Markdown report plus structured JSON for downstream chaining.
Detect Neanderthal and Denisovan introgression segments from modern human genomes
Search scientific papers via the BGPT MCP server and retrieve structured experimental data — methods, results, conclusions, quality scores, and 25+ metadata fields per paper.
Meta-agent that routes bioinformatics requests to specialised sub-skills. Handles file type detection, analysis planning, report generation, and reproducibility export.
Bioconductor package discovery, workflow recommendation, setup inspection, and starter code generation grounded in official Bioconductor containers and BiocManager.
Cell segmentation in fluorescence microscopy images. Supports Cellpose/cpsam (Cellpose 4.0) with additional backends planned. Produces segmentation masks, per-cell morphology metrics (area, diameter, centroid, eccentricity), overlay figures, and a report.md.
Ancestry decomposition PCA against the Simons Genome Diversity Project
Shotgun metagenomics profiling — taxonomy, resistome, and functional pathways
Claude Skills — questions
Answers built from the skills we actually parsed.