Skills
96,770 packaged capabilities and playbooks for agents. Install any of them with one command.
Data & databases
Molecular ML with diverse featurizers and pre-built datasets. Use for property prediction (ADMET, toxicity) with traditional ML or GNNs when you want extensive featurization options and MoleculeNet benchmarks. Best for quick experiments with pre-trained models, diverse molecular representations. For graph-first PyTorch workflows use torchdrug; for benchmark datasets use pytdc.
Design privacy-preserving analytics systems using differential privacy, k-anonymity, l-diversity, and t-closeness. Covers privacy budget allocation with epsilon tracking, references Google DP library, OpenDP, and Apple PPML. Includes Python differential privacy implementation for GDPR-compliant statistical analysis.
Professional diagram generation skill: automatically selects the appropriate chart type based on the requirement and generates PNG charts that conform to the design spec. Covers 29 chart types: structural diagrams (flowchart, swimlane diagram, sequence diagram, architecture diagram, state diagram, ER diagram, class diagram, mind map, Gantt chart, Kanban, Git Graph, etc.) and statistical charts (bar chart, line chart, pie chart, radar chart, heatmap, scatter plot, Sankey diagram, funnel chart, waterfall chart, treemap, bar-line combo chart, etc.). Unified toolset: HTML/SVG + inline JS layout computation. ER/class diagrams use the ELKjs layout engine. All charts follow a unified design spec (color/font/component/spacing) with a modern, clean style. Supports three outputs: PNG (default), HTML (rich document embedding), DSL (Mermaid text, embedded in MD documents). Trigger words: draw a diagram, draw one, generate a chart, flowchart, architecture diagram, sequence diagram, bar chart, comparison chart, relationship diagram. Even if the user does not say "draw a diagram", this skill should trigger whenever the request involves visual presentation (process, architecture, data comparison, relationships). When called by other skills such as deep-research, it likewise follows this spec to generate diagrams.
Comprehensive patterns for AI-powered document understanding including PDF parsing, OCR, invoice/receipt extraction, table extraction, multimodal RAG with vision models, and structured data output. Use when "document parsing, PDF extraction, OCR, invoice processing, receipt extraction, document understanding, LlamaParse, Unstructured, vision document, table extraction, structured output from PDF, " mentioned.
Orchestrator for the DynamoDB single-table design pipeline. Use when the user wants to design a DynamoDB table end-to-end or is unsure which step to start with. Routes to the appropriate specialized skill: dynamodb-access-patterns, dynamodb-table-design, or dynamodb-query-interfaces.
Analyze inventory health using turnover ratios, ABC classification, safety stock calculations, and stockout vs overstock diagnostics. Use this skill when the user needs to optimize inventory levels, reduce carrying costs, prevent stockouts, or classify products by inventory priority — even if they say 'we have too much stock', 'we keep running out of bestsellers', 'how much safety stock do we need', or 'which products should we focus on'.
Calculate and analyze promotional ROI including incremental sales lift, margin impact, and promo type comparison. Use this skill when the user needs to evaluate whether a promotion was profitable, compare promotion types, simulate promo scenarios, or optimize promotional spending — even if they say 'was this sale worth it', 'which promo type works best', 'are we giving away too many discounts', or 'plan our next promotion'.
Perform comprehensive exploratory data analysis on scientific data files across 200+ file formats. This skill should be used when analyzing any scientific data file to understand its structure, content, quality, and characteristics. Automatically detects file type and generates detailed markdown reports with format-specific analysis, quality metrics, and downstream analysis recommendations. Covers chemistry, bioinformatics, microscopy, spectroscopy, proteomics, metabolomics, and general scientific data formats.
This skill provides guidance for extracting text commands, moves, or typed input from video recordings using OCR. It applies when extracting gameplay commands (e.g., Zork), terminal sessions, or any text-based interactions captured in video format. Use this skill when processing videos of text-based games, terminal recordings, or any scenario requiring OCR-based command extraction from screen recordings.
Build three-statement financial models (Income Statement, Balance Sheet, Cash Flow) with revenue forecasting, assumption management, and scenario analysis. Use this skill when the user needs to project financials, build a fundraising model, create financial projections for a business plan, or evaluate M&A targets — even if they say 'build a financial model', 'project our revenue', 'how much money will we make next year', or 'model this acquisition'.
Implements financial records retention requirements across EU directives (5-7 years), SOX Section 802 (7 years), MiFID II (5-7 years), tax records, payment data, and AML obligations under AMLD. Maps financial data categories to statutory retention periods with cross-jurisdictional reconciliation. Activate for financial retention, SOX records, MiFID retention, AML retention, tax record keeping queries.
Guides FinOps analysis on AWS, GCP, and Azure—cost visibility and allocation, tagging and showback/chargeback models, rightsizing and waste removal, RI/Savings Plan/CUD recommendations, budgets and forecasts, anomaly detection, unit economics (cost per service/customer), and FinOps cadence with engineering accountability. Use when optimizing cloud spend, analyzing CUR/billing exports, building cost dashboards, explaining bill spikes, or improving allocation—not for GL mapping, capex, depreciation, or month-end ledger close (compute-accounting-manager), enterprise EA negotiation (enterprise-cloud-architect), hands-on resource provisioning (cloud-engineer), or hardware supply efficiency (data-center-compute-supply-efficiency).
Python library for working with geospatial vector data including shapefiles, GeoJSON, and GeoPackage files. Use when working with geographic data for spatial analysis, geometric operations, coordinate transformations, spatial joins, overlay operations, choropleth mapping, or any task involving reading/writing/analyzing vector geographic data. Supports PostGIS databases, interactive maps, and integration with matplotlib/folium/cartopy. Use for tasks like buffer analysis, spatial joins between datasets, dissolving boundaries, clipping data, calculating areas/distances, reprojecting coordinate systems, creating maps, or converting between spatial file formats.
Fast CLI/Python queries to 20+ bioinformatics databases. Use for quick lookups: gene info, BLAST searches, AlphaFold structures, enrichment analysis. Best for interactive exploration, simple queries. For batch processing or advanced BLAST use biopython; for multi-database Python workflows use bioservices.
Covers the Google Gemini embeddings API (gemini-embedding-001) for RAG, semantic search, document clustering, and similarity matching: @google/genai SDK usage, fetch-based calls from Cloudflare Workers, batch embedding, the 8 task types (RETRIEVAL_QUERY, RETRIEVAL_DOCUMENT, SEMANTIC_SIMILARITY and others), dimension selection from 128 to 3072, cosine similarity, chunking, caching, and Cloudflare Vectorize integration. Prevents 8+ embedding errors including dimension mismatches, wrong task type, free-tier rate limiting (100 RPM), vector normalization mistakes, 2,048 token text truncation, and model version confusion. Use when generating text embeddings with Gemini, building a RAG or vector search pipeline, storing vectors in Vectorize, or debugging embedding dimension, rate limit, or normalization problems.
Manages RoPA for complex multi-entity corporate groups including entity-level versus group-level records, intra-group transfer documentation, and shared processing coordination. Activate for group RoPA, multi-entity, corporate group, intra-group transfers, subsidiary records, holding company.
Lightweight WSI tile extraction and preprocessing. Use for basic slide processing tissue detection, tile extraction, stain normalization for H&E images. Best for simple pipelines, dataset preparation, quick tile-based analysis. For advanced spatial proteomics, multiplexed imaging, or deep learning pipelines use pathml.
Direct REST API access to KEGG (academic use only). Pathway analysis, gene-pathway mapping, metabolic pathways, drug interactions, ID conversion. For Python workflows with multiple databases, prefer bioservices. Use this for direct HTTP/REST work or KEGG-specific control.
Use this skill for ANY question about creating test or evaluation datasets for LangChain agents. Covers generating datasets from traces (final_response, single_step, trajectory, RAG types), uploading to LangSmith, and managing evaluation data.
MATLAB and GNU Octave numerical computing for matrix operations, data analysis, visualization, and scientific computing. Use when writing MATLAB/Octave scripts for linear algebra, signal processing, image processing, differential equations, optimization, statistics, or creating scientific visualizations. Also use when the user needs help with MATLAB syntax, functions, or wants to convert between MATLAB and Python code. Scripts can be executed with MATLAB or the open-source GNU Octave interpreter.
Based on model_route.json, the data plan, and cleaned data, generate the results evidence contract and q1/q2/q3 modeling code scaffolding for the mathematical modeling paper. Invoke when model outputs, evaluation metrics, structured conclusions, paper tables, and problem-specific modeling code need to be consolidated into paper_output/results/, paper_output/tables/, and paper_output/code/modeling/ for QA and body-text generation to read.
Use CodexBar CLI local cost usage to summarize per-model usage for Codex or Claude, including the current (most recent) model or a full model breakdown. Trigger when asked for model-level usage/cost data from codexbar, or when you need a scriptable per-model summary from codexbar cost JSON.
Run and analyze molecular dynamics simulations with OpenMM and MDAnalysis. Set up protein/small molecule systems, define force fields, run energy minimization and production MD, analyze trajectories (RMSD, RMSF, contact maps, free energy surfaces). For structural biology, drug binding, and biophysics.
Queries and manages Microsoft SQL Server databases via the mssql CLI. Use when running SQL queries, inspecting table schemas, or managing server connections. Triggers on phrases like "SQL query", "database", "table schema", "run query", "SQL Server", "mssql", "check the database", "look up in the database".
Implicit use requires an explicit or already-established Nature Portfolio context. Design, revise, audit, and export submission-grade quantitative figures and multi-panel layouts using Python or R, and route manuscript schematics or graphical abstracts through an appropriate visual workflow. Within that scope, use for 论文配图、科研绘图、Figure 1 or journal-ready SVG/PDF/TIFF. Do not use for statistics-only inference, dashboards, slide decks, photo editing, or general illustration
News site content extraction. Supports WeChat official accounts, Toutiao, NetEase News, Sohu News, and Tencent News. Activate when users need to extract news content, scrape official account articles, crawl news, or obtain news as JSON/Markdown.
Comprehensive SaaS architecture planner for Next.js and Supabase applications. Converts ideas into production-ready technical plans with database schemas, file structures, feature breakdowns, and week-by-week development roadmaps. Use when planning subscription-based applications, multi-tenant SaaS products, or building from idea to launch with Next.js and Supabase stack.
General ontological analysis skill. Use when: analysing a domain to identify what things exist, how they relate, and what makes them the same thing over time; producing a domain ontology model from requirements or existing systems; reviewing an existing model for ontological coherence. Produces ontology models that inform architects (for solution design) and engineers (for implementation). Does NOT produce architecture designs or code — those are downstream concerns.
GPU-accelerate Python code using CuPy, Numba CUDA, Warp, cuDF, cuML, cuGraph, KvikIO, cuCIM, cuxfilter, cuVS, cuSpatial, and RAFT. Use whenever the user mentions GPU/CUDA/NVIDIA acceleration, or wants to speed up NumPy, pandas, scikit-learn, scikit-image, NetworkX, GeoPandas, or Faiss workloads. Covers physics simulation, differentiable rendering, mesh ray casting, particle systems (DEM/SPH/fluids), vector/similarity search, GPUDirect Storage file IO, interactive dashboards, geospatial analysis, medical imaging, and sparse eigensolvers. Also use when you see CPU-bound Python code (loops, large arrays, ML pipelines, graph analytics, image processing) that would benefit from GPU acceleration, even if not explicitly requested.
PostgreSQL procedural language (PL/pgSQL). Covers stored procedures, functions, triggers, exception handling, and control structures. Use for PostgreSQL server-side programming. USE WHEN: user mentions "plpgsql", "PostgreSQL functions", "PostgreSQL procedures", "PostgreSQL triggers", "RETURNS TABLE", "RETURNS SETOF", "RAISE NOTICE" DO NOT USE FOR: basic PostgreSQL SQL - use `postgresql` instead, PL/SQL (Oracle) - use `plsql` instead, T-SQL - use `tsql` instead
Fast in-memory DataFrame library for datasets that fit in RAM. Use when pandas is too slow but data still fits in memory. Lazy evaluation, parallel execution, Apache Arrow backend. Best for 1-100GB datasets, ETL pipelines, faster pandas replacement. For larger-than-RAM data use dask or vaex.
Build clinical/healthcare deep-learning pipelines with PyHealth — loading EHR/signal/imaging datasets (MIMIC-III/IV, eICU, OMOP, SleepEDF, ChestXray14, EHRShot), defining tasks (mortality, readmission, length-of-stay, drug recommendation, sleep staging, ICD coding, EEG events), instantiating models (Transformer, RETAIN, GAMENet, SafeDrug, MICRON, StageNet, AdaCare, CNN/RNN/MLP), training with the PyHealth Trainer, computing clinical metrics, and using medical code utilities (ICD/ATC/NDC/RxNorm lookup and cross-mapping). Use this skill whenever the user mentions PyHealth, MIMIC, eICU, OMOP, EHR modeling, clinical prediction, drug recommendation, sleep staging, medical code mapping, ICD/ATC codes, or any healthcare ML pipeline that fits the dataset → task → model → trainer → metrics pattern, even if "PyHealth" isn't named explicitly.
If the user asks to explore, analyze, query, chart, dashboard, or govern (master items, business glossary, data products) a Qlik Sense app or Qlik Cloud analytics content via the Qlik MCP server — THEN invoke this skill. This covers requests like "what's in this Qlik app", "chart X by Y in [app]", "build a sheet for...", "create a reusable measure for...", "what does [glossary term] mean", "is this dataset trustworthy/fresh", or any question that requires calling qlik_* MCP tools. DO NOT invoke for SQL warehouse questions, DW/BI-tool-agnostic data modelling, or requests with no Qlik MCP tool involved.
Design, build, and refine a QR code–based tracking system. This skill helps generate QR codes, define what happens when they are scanned, and track scan events (who, where, when, how) for marketing, operations, product flows, or internal tools.
Run a comprehensive quarterly CRM audit covering list health, bounce monitoring, data quality, scoring calibration, engagement metrics, and property cleanup. Produces a health report with quarter-over-quarter trend comparison.
Use when results are valid but surprising, negative, unstable, or ambiguous — to decide debug/rerun/ablate/revise/park. Not for engineering failures like NaN/OOM (use experiment-debugger). Not for confound or claim-drift audit before locking results into the paper (use research-results-auditor).
Analyze user retention and churn using survival analysis, cohort analysis, and machine learning. Calculate retention rates, build survival curves, predict churn risk, and generate retention optimization strategies. Use when working with user subscription data, membership information, or when user mentions retention, churn, survival analysis, or customer lifetime value.
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.