>
Skills（SKILL.md）は、AIエージェント（Claude Code、Cursor、Codexなど）に特定の能力を追加するための設定ファイルです。
詳しく見る →>
Guide to annotating cell types in single-cell RNA-seq data. Covers manual marker-based, automated (CellTypist, scAnnotate), and reference-based (scArches, Azimuth, SingleR) approaches with decision framework for tool selection. Includes key marker genes for blood/immune, epithelial, and stromal lineages, validation checklist, and common pitfalls. Cross-references: scanpy-scrna-seq for preprocessing, celltypist-cell-annotation for automated classification, scvi-tools-single-cell for deep-learning-based label transfer.
Query PharmGKB (Clinical Pharmacogenomics) database via REST API for drug-gene interactions, clinical annotations, dosing guidelines (CPIC, DPWG), variant-drug associations, and pharmacogenomic pathways. Search by gene, drug, rsID, or pathway. No authentication required. For somatic cancer pharmacogenomics use cosmic-database or opentargets-database; for drug structures use chembl-database-bioactivity.
Query COSMIC (Catalogue Of Somatic Mutations In Cancer) for cancer somatic mutations, gene census data, mutational signatures, drug resistance variants, and cancer gene annotations. REST API v3.1 supports gene/sample/variant queries. Free registration required. For germline clinical variants use clinvar-database; for drug-target data use opentargets-database or chembl-database-bioactivity.
Query NCBI dbSNP for SNP records by rsID, gene, or genomic region via E-utilities (esearch, efetch, epost) and NCBI Variation Services REST API. Retrieve allele data, minor allele frequency, variant class (SNV, indel, MNV), clinical significance links, and cross-database IDs (ClinVar, dbVar, 1000G). Free access; 3 req/sec without API key, 10 req/sec with key. For clinical pathogenicity classifications use clinvar-database; for population frequencies use gnomad-database.
Counts aligned RNA-seq reads overlapping gene features in a GTF annotation. Takes sorted BAM files from STAR alignment and a GTF file; outputs a tab-delimited count matrix per gene across all samples. Handles strandedness (0=unstranded, 1=stranded, 2=reverse-stranded), paired-end, and multi-sample batch counting in a single command. Use Salmon instead for alignment-free quantification; use featureCounts when STAR BAMs already exist and a gene-level count matrix is needed.
Query NCBI Gene Expression Omnibus (GEO) for gene expression datasets and sample metadata via GEOparse Python library and E-utilities. Search datasets by keyword/organism/platform, download GSE series matrices, parse GPL platform annotations, extract GSM sample metadata, and load expression matrices into pandas. For single-cell data use cellxgene-census; for programmatic multi-DB access use gget-genomic-databases.
De novo motif discovery and known TF motif enrichment in ChIP-seq and ATAC-seq peak sets using HOMER. findMotifsGenome.pl searches for over-represented sequence patterns against a background; annotatePeaks.pl assigns genomic context (TSS distance, gene, repeat). Use after MACS3 peak calling to identify which transcription factors are enriched in your peaks, annotate peaks with nearest genes, and validate ChIP-seq quality by checking the target TF's own motif.
Calls narrow and broad peaks from ChIP-seq and ATAC-seq BAM files using a Poisson model. MACS3 callpeak identifies enriched genomic regions (transcription factor binding sites or histone marks) against an input/IgG control; outputs BED narrowPeak/broadPeak files for downstream motif analysis, annotation, and differential binding. Use narrow peaks for TF ChIP-seq and ATAC-seq; use broad peaks for H3K27me3, H3K9me3, and other broad histone marks.
Retrieve quantitative phenotype measurements across inbred mouse strains from the Mouse Phenome Database (MPD) for metabolic, behavioral, and physiological traits. Query strain means and raw individual measurements for body weight, glucose, blood pressure, behavioral assays, and 40+ additional procedures. Use for QTL analysis support, cross-strain phenotype comparison, and identifying mouse models for metabolic or behavioral traits. For mouse gene-disease-phenotype associations use monarch-database; for mouse genome annotations use ensembl-database.
Consensus cell type annotation by running 10+ algorithms (KNN-Harmony, KNN-BBKNN, KNN-Scanorama, KNN-scVI, CellTypist, ONCLASS, Random Forest, SCANVI, SVM, XGBoost) on a labeled reference and transferring labels to a query dataset via majority voting. popV produces per-method labels, an overall consensus prediction, and an agreement score quantifying confidence across methods. Use when single-method annotation is insufficient or when you need ensemble uncertainty estimates for novel cell states.
Query RegulomeDB v2 REST API to score genetic variants for regulatory function and retrieve overlapping regulatory evidence (TF binding sites, histone marks, DNase-seq peaks, eQTLs, motifs). Score a single rsID or genomic position, batch-score variant lists, search regions for all regulatory variants, and retrieve full annotation details. Scores range from 1a (strongest: eQTL + TF + DNase + motif) to 7 (no known regulatory function). Use for GWAS hit prioritization, regulatory variant annotation, and cis-regulatory element discovery. For clinical pathogenicity use clinvar-database; for GWAS associations use gwas-database.
Deep generative models for single-cell omics. Probabilistic batch correction (scVI), semi-supervised annotation (scANVI), CITE-seq RNA+protein modeling (totalVI), transfer learning to new datasets (scARCHES), and differential expression with uncertainty quantification. Unified setup→train→extract API on AnnData. Use harmony-batch-correction for fast linear batch correction without deep learning; use muon for multi-modal MuData workflows; use scVI for probabilistic, deep learning-based integration with uncertainty quantification.
Decision framework for choosing between manual marker-based, automated (CellTypist), and reference-based (popV) cell type annotation strategies in single-cell RNA-seq analysis. Covers a three-tier strategy: Tier 1 (manual canonical markers), Tier 2 (CellTypist pre-trained models), and Tier 3 (popV ensemble label transfer). Use when planning or troubleshooting cell type annotation in any scRNA-seq project.
Annotate and filter genetic variants in VCF files using SnpEff and SnpSift. SnpEff predicts functional effects (HIGH/MODERATE/LOW/MODIFIER impact), gene names, transcript IDs, amino acid changes, and HGVS notation; SnpSift filters annotated VCFs and adds ClinVar/dbSNP annotations. Java-based command-line tools with Python subprocess integration. Use ANNOVAR for alternative multi-database annotation; use Ensembl VEP for web-based or REST API annotation; use SnpEff for fast command-line annotation with pre-built genome databases.
Automatically annotate plasmid sequences with functional features (promoters, terminators, resistance genes, origins of replication, tags, fluorescent proteins) using BLAST-based detection against curated databases (Addgene, fpbase, SnapGene). Accepts FASTA or raw sequence; outputs annotated GenBank files, interactive HTML maps, and CSV feature tables. Handles circular topology correctly. Use for verifying synthetic plasmid construction, preparing Addgene submissions, sharing plasmid maps, or batch-annotating a cloning library.
Query the InterPro REST API for protein domain architecture, family classification, and member database integration. Search InterPro entries by name or accession, retrieve all domains in a protein (domain architecture), list proteins in a family, get taxonomic distribution, and link to PDB structures. Integrates Pfam, PANTHER, PIRSF, PRINTS, PROSITE, SMART, CDD, and NCBIfam into a unified hierarchy. For protein sequence and Swiss-Prot annotations use uniprot-protein-database; for experimental 3D structures use pdb-database.
Query UniProt protein database via REST API. Search by gene/protein name, retrieve FASTA sequences, map IDs across databases (Ensembl, PDB, RefSeq), access Swiss-Prot annotations. For unified multi-database access use bioservices; for protein structure use alphafold-database.
Query drug-drug interaction (DDI) data from DDInter via REST API. Search interactions by drug name or ID, retrieve severity levels (major/moderate/minor), interaction mechanisms, and clinical recommendations for drug pairs. Covers 1.7M+ interactions across 2,400+ drugs. No authentication required. For FDA drug labeling use dailymed-database; for pharmacogenomics use clinpgx-database.
Create a new genre README.md for the bitwize-music genre library at `${CLAUDE_PLUGIN_ROOT}/genres/`.
Use when running claudikins-kernel:ship, preparing PRs, writing changelogs, deciding merge strategy, or handling CI failures — enforces GRFP-style iterative approval, code integrity validation, and human-gated merges
Zero-shot zone detection: text prompts → patch-grid cosine heatmaps → bboxes.
>-
>
>
>
>
>
Understand this repository quickly before making changes. Use for architecture discovery, ownership mapping, command selection, and initial implementation planning.
Onboard quickly to this repository. Use when asked to understand architecture, locate ownership, choose the right module, or identify the correct commands and verification strategy before coding.
Apply 2026's top graphic design trends to any creative brief. Based on Kittl × Savee's 2026 Design Trends Report (10 trends + 2 honorable mentions), backed by Adobe, Figma, and Pinterest data. Use when: **Designing a brand identity** — pick the right aesthetic for your audience; **Creating social media assets** — use trending visual languages that perform; **Briefing a designer or AI image tool** — give precise style direction with vocabulary and references; **Refreshing a visual identity** — know what's rising vs saturating; **Building mood boards** — combine trends intentionally with data-backed rationale.
Scale your brand voice across multiple languages using AI voice synthesis, maintaining consistent character and quality for global content. Use when: Expanding video content to new language markets; Creating multilingual courses or training; Localizing ads and marketing videos; Dubbing existing content for international audiences; Building consistent global brand voice
Detect semantic-level bugs by analyzing whether code behavior matches its intended purpose inferred from function/variable names, comments, docstrings, and documentation. Use when users need to: (1) Find logic errors where implementation contradicts stated intent, (2) Identify off-by-one errors and boundary mismatches, (3) Detect inverted logic or wrong operators, (4) Catch missing edge case handling, (5) Verify code matches its documentation. Highlights mismatches between intent and implementation across multiple programming languages.
Analyzes a feature or sprint for scope creep by comparing current scope against the original plan, flagging additions and recommending cuts. Use when a sprint feels bloated or when the user mentions scope creep, scope check, or feature bloat.
>
Use when generating Google Stitch DESIGN.md systems for premium typography, color, layout, motion intent, and anti-generic UI rules.
Query InterPro for protein family, domain, and functional site annotations. Integrates Pfam, PANTHER, PRINTS, SMART, SUPERFAMILY, and 11 other member databases. Use for protein function prediction, domain architecture analysis, evolutionary classification, and GO term mapping.
Automate project management workflows in Shortcut -- create stories, manage tasks, track epics, and organize workflows through natural language commands.
Observational drawing and visual perception techniques for art education. Covers contour drawing, gesture drawing, negative space, proportion and measurement, value mapping, spatial depth cues, and the cognitive shift from symbolic to perceptual seeing. Use when teaching drawing fundamentals, analyzing observational accuracy, or developing visual literacy in any medium.
Grammar of graphics, chart type selection, color theory, interactive visualization, dashboards, and the principles of honest, effective data display. Covers Tufte's data-ink ratio, Bertin's visual variables, perceptual principles, accessibility, small multiples, annotation, and the full workflow from exploratory plots to publication-quality graphics. Use when creating charts, designing dashboards, critiquing visualizations, or choosing how to display data.
Translation and interpretation as language meta-skills -- equivalence theory (Nida's formal vs. dynamic equivalence), translation strategies (borrowing, calque, transposition, modulation, adaptation), interpretation modes (simultaneous, consecutive, sight), source text analysis, register preservation, cultural adaptation, untranslatability and compensatory strategies, machine translation literacy, and back-translation for verification. Covers literary, technical, legal, and community interpreting contexts. Use when translating between languages, evaluating translation quality, teaching translation skills, or understanding the limits of cross-linguistic transfer.
Root cause analysis techniques for modern distributed systems, microservices, and cloud infrastructure. Covers trace-based causality (OpenTelemetry, Jaeger, Tempo), service dependency graph analysis, DynaCausal-style dynamic causality-aware RCA, anomaly localization with metrics/logs/traces correlation, the Microsoft AgentRx framework for AI agent failures, chaos-engineering-as-RCA, and the 2025 industry shift toward causal-graph-based AIOps. Use when investigating a production incident in a microservice mesh, a Kubernetes cluster, a distributed database, or any system where the fault could originate in any of dozens of services and the causal chain runs through network hops, retries, timeouts, queues, and asynchronous message flows.
Block-based world-building paradigms — voxel grids, modular prefabs, LEGO/Minecraft-style assembly, parametric voxel art, and the engineering mindset of building from discrete units. Covers load-bearing patterns, spanning gaps, blueprint-to-build translation, material selection, pixel-art scaling into block space, and iterative build cycles. Use when teaching or applying block-based construction in voxel games, modular prefab CAD, or any system where the design space is quantized into discrete primitives.
Aminet search and browse: full-text search, category tree navigation, architecture filtering, package detail, and curated collections. Use when searching, browsing, or managing package collections.
Manages Aminet INDEX infrastructure: fetch, parse, cache, and incremental update of ~84,000-entry package database. Use when managing INDEX data, checking freshness, or parsing .readme files.
Aminet package installation: LhA/LZX extraction, Amiga filesystem mapping, dependency detection, install tracking, and scan gate enforcement. Use when installing, extracting, or uninstalling Amiga packages.
Read before updating changelogs — update CHANGELOG.md with notable changes since last release
agiled
akismet
announcekit