Skills
97,501 packaged capabilities and playbooks for agents. Install any of them with one command.
Data & databases
Builds local vector search, RAG, embeddings, and semantic search on the Azure SQL Database container using the native VECTOR type and VECTOR_DISTANCE. Use when you need to store embeddings, do similarity search, top-k nearest neighbor, cosine distance, retrieval-augmented generation, "find similar documents", chatbot memory, or semantic lookup against a local SQL database. Use this instead of pgvector, FAISS, Chroma, Pinecone, or a separate vector store when the data already lives in (or can live in) Azure SQL. Covers the VECTOR(n) column type, inserting embeddings with CAST(CAST(? AS NVARCHAR(MAX)) AS VECTOR(n)) where the dimension is a literal, a pluggable embed() so only the endpoint changes for cloud, and a working CREATE VECTOR INDEX with the two errors that block it. Provisions appdb on master first so every script runs on a fresh container.
Generate client-facing performance reports for Chinese wealth management clients. Adapted from the original client-report skill for A-share portfolios, Chinese regulatory requirements, and local reporting conventions. Triggers on "A股客户报告", "客户持仓报告", "client report China", "wealth report A-share", "客户业绩报告", or "portfolio report for [client]".
Comparable company analysis for A-share equities. Adapts the original comps-analysis skill for Chinese market data sources, A-share trading multiples, and domestic peer selection. Triggers on "A股可比公司", "可比分析", "comps China", "peer comparison A-share", "可比公司分析", or "trading comps [industry]".
Discounted cash flow (DCF) valuation model for A-share equities. This alias skill redirects to china-dcf, which provides comprehensive DCF modeling adapted for Chinese market conventions. Triggers on "A股DCF模型", "DCF估值", "DCF model China", "现金流折现", "discounted cash flow A-share", or "DCF [company]".
Provides comprehensive guidance for Alibaba Cloud RDS including database creation, configuration, backups, and RDS management. Use when the user asks about Alibaba Cloud RDS, needs to create RDS databases, configure RDS, or manage Alibaba Cloud database services.
Detects and prevents visual misleads, cognitive biases, and data integrity violations in visualizations, dashboards, reports, and presentations. Audits charts for honesty, diagnoses misinterpretation causes, and provides specific fixes. Invoke when user mentions chartjunk, misleading chart, truncated axis, data integrity, visual deception, 3D chart problems, cherry-picking data, or needs to audit visualizations for accuracy. For general design evaluation, use `design-evaluation-audit`. For cognitive foundations, use `cognitive-design`.
Fetch, rank, and enrich recent arXiv and Hugging Face papers. English triggers: “fetch papers”, “fetch papers from the last three days”. This is pipeline stage 1. Paper fetching (step 1 of a 3-step pipeline). Fetch the latest arXiv + HuggingFace papers, score and filter them, enrich their information, and output daily_papers_enriched.json to a shared temporary directory for subsequent skills. Triggers: “论文抓取”, “跑一下论文抓取”. Supports multi-day mode: “过去3天论文推荐”, “过去一周论文推荐”, “过去一周的论文”, “抓 3 天的论文”, “最近5天”
Provides PyTorch-native distributed LLM pretraining using torchtitan with 4D parallelism (FSDP2, TP, PP, CP). Use when pretraining Llama 3.1, DeepSeek V3, or custom models at scale from 8 to 512+ GPUs with Float8, torch.compile, and distributed checkpointing.
Optimizing vector embeddings for RAG systems through model selection, chunking strategies, caching, and performance tuning. Use when building semantic search, RAG pipelines, or document retrieval systems that require cost-effective, high-quality embeddings.
Event sourcing patterns for functional TypeScript — persist state as an append-only log of past events and rebuild it by folding them. Use when implementing a Decider write model, an event store, projections and read models, event versioning, or snapshots. Builds on the Decider from domain-driven-design and the ports/adapters from hexagonal-architecture. Do NOT reach for it on CRUD, or when an audit log or the outbox pattern already suffices.
Fine-tune LLMs using reinforcement learning with TRL - SFT for instruction tuning, DPO for preference alignment, PPO/GRPO for reward optimization, and reward model training. Use when need RLHF, align model with preferences, or train from human feedback. Works with HuggingFace Transformers.
Geospatial vector analysis extending pandas. Read/write spatial formats (Shapefile, GeoJSON, GeoPackage, Parquet, PostGIS), CRS handling, geometric ops (buffer, simplify, centroid, affine), spatial analysis (joins, overlays, dissolve, clipping, distance), visualization (choropleth, interactive maps, basemaps). Use for spatial joins, overlays, CRS transforms, area/distance, maps.
MATLAB/GNU Octave numerical computing: matrices, linear algebra, ODEs, signal processing, optimization, statistics, scientific visualization. MATLAB-syntax examples run on both. For Python use numpy/scipy; for statistical modeling use statsmodels.
Guidance for Microsoft Priva — privacy risk management and subject rights requests. Covers Priva Privacy Risk Management and Priva Subject Rights Requests to find privacy risks and fulfill data subject requests (GDPR/CCPA). WHEN: Microsoft Priva, privacy risk management, subject rights request, DSAR, data subject request, GDPR fulfillment, privacy risk policies, data minimization, overexposed personal data.
Build end-to-end MLOps pipelines from data preparation through model training, validation, and production deployment. Use when creating ML pipelines, implementing MLOps practices, or automating model training and deployment workflows.
Interactive viewer for microscopy. Displays 2D/3D/4D arrays as Image, Labels, Points, Shapes, Tracks layers; supports annotation, plugin analysis, headless screenshots. Core visualization for Python bioimage workflows. Use ImageJ/FIJI for macro processing; napari for Python-native interactive visualization and DL segmentation review.
(NS) PostgreSQL retrieval — pgvector, hybrid FTS+vector, GraphRAG tables, multi-hop, 1MM+ scale. Use for RAG or GraphRAG on Postgres, pgvector, hybrid search, document links, entity merge, multi-hop, or 1MM+ retrieval — even if they say search, embeddings, or knowledge graph in SQL. Do NOT use for ontology/cite-or-refuse (`ns-graphrag`), app code, competing vector DBs, CrewAI, or generic web apps.
Use when the user wants to verify paper claims against code or data, audit numerical accuracy, check formula-code alignment, or validate citation accuracy. Triggers on phrases like "verify claims", "check numbers", "do the numbers match", "formula vs code", "audit the paper", or "cross-check results".
Programmatic PubMed access via NCBI E-utilities REST API. Covers Boolean/MeSH queries, field-tagged search, endpoints (ESearch, EFetch, ESummary, EPost, ELink), history server for batches, citation matching, systematic review strategies. Use for biomedical literature search or automated pipelines.
Guidance for Microsoft Purview data classification and sensitivity labels - sensitive information types (SITs), trainable classifiers, exact data match, and sensitivity label taxonomy with auto-labelling. Covers building a label taxonomy, picking detection methods, and rolling out auto-labelling in simulation. WHEN: data classification, sensitivity labels, sensitive information types, SIT, trainable classifier, exact data match EDM, auto-labelling, label taxonomy, classify data, information protection labels, MIP labels.
Guidance for Microsoft Purview Data Lifecycle Management and Records Management—retention labels and policies to retain and delete content across Microsoft 365, plus declaring records. Covers retention strategy, adaptive vs. static scopes, principle of precedence, disposition review, and regulatory records. WHEN: data lifecycle management, retention policy, retention label, records management, declare record, disposition review, adaptive scopes, retain and delete content, file plan, regulatory retention, immutable records.
Guidance for the Microsoft Purview Data Map—the foundation that scans and maps data sources across multicloud and on-premises estates to power cataloging and governance. Covers source registration, integration runtimes (managed vs. self-hosted), scan rule sets, classifications, collections, and cost control. WHEN: Purview Data Map, scan data sources, register data source, data discovery, integration runtime, map enterprise data, multicloud data scanning, classification scan, collections, glossary.
Orientation skill for the Microsoft Purview data security, governance, and compliance suite. Helps choose the right Purview solution for a goal and understand the unified portal, roles, and licensing model. WHEN: Microsoft Purview overview, which Purview solution, Purview portal, data governance vs. data security vs. compliance, Purview roles and permissions, where to start with Purview, Purview licensing, I do not know which Purview feature to use, what does Purview cover, getting started with Purview, Purview product overview, which compliance tool do I need. Use this skill first for Purview orientation, then follow up with the specific skill for your use case.
Scalable data processing for ML workloads. Streaming execution across CPU/GPU, supports Parquet/CSV/JSON/images. Integrates with Ray Train, PyTorch, TensorFlow. Scales from single machine to 100s of nodes. Use for batch inference, data preprocessing, multi-modal data loading, or distributed ETL pipelines.
This skill MUST be used for product-review salience scoring (0–7), review × attribute Excel matrices, or PCA and K-means analysis of those review scores (逐則評論評分、提及程度、顯著度矩陣、評論×屬性 Excel、評論 PCA、評論分群). It MUST support requests for individual stages and MUST NOT be used for unrelated PCA or clustering tasks.
Meta-skill for publication-ready figures. Use when creating journal submission figures requiring multi-panel layouts, significance annotations, error bars, colorblind-safe palettes, and specific journal formatting (Nature, Science, Cell). Orchestrates matplotlib/seaborn/plotly with publication styles. For quick exploration use seaborn or plotly directly.
Classical ML in Python: classification, regression, clustering, dim reduction, evaluation, tuning, preprocessing pipelines. Linear models, tree ensembles, SVMs, K-Means, PCA, t-SNE. Use PyTorch/TF for deep learning; XGBoost/LightGBM for scale.
Use when the user wants Instagram research or workflow guidance for lead generation, influencer discovery, brand monitoring, competitor analysis, content analytics, trend research, or audience analysis, including profile analysis, feed collection, post or reel inspection, transcript extraction, comment analysis, reel discovery, highlight retrieval, or embed generation.
Professional sub-skill for scikit-learn focused on robust pipeline architecture, custom estimator development, advanced feature engineering, and rigorous model validation. Covers Target Encoding, Nested Cross-Validation, and Production Deployment.
Master SQL query optimization, indexing strategies, and EXPLAIN analysis to dramatically improve database performance and eliminate slow queries. Use when debugging slow queries, designing database schemas, or optimizing application performance.
Use when doing ANY task involving Supabase. Triggers: Supabase products (Database, Auth, Edge Functions, Realtime, Storage, Vectors, Cron, Queues); client libraries and SSR integrations (supabase-js, @supabase/ssr) in Next.js, React, SvelteKit, Astro, Remix; auth issues (login, logout, sessions, JWT, cookies, getSession, getUser, getClaims, RLS); Supabase CLI or MCP server; schema changes, migrations, declarative schemas, security audits, Postgres extensions (pg_graphql, pg_cron, pg_vector); debugging and troubleshooting errors or unexpected behavior on Supabase projects (HTTP errors, Postgres errors, RLS surprises, permission denied, schema cache issues, timeouts, Edge Function crashes, Realtime drops, Storage failures) and reading or querying logs (Logs Explorer, ClickHouse).
Guide for implementing Syncfusion Gauge controls in Windows Forms applications. Use when creating data visualization gauges such as RadialGauge for circular displays (speedometers, temperature dials), LinearGauge for horizontal/vertical bars and progress indicators, or DigitalGauge for LED-style alphanumeric displays. Covers dashboard gauges, instrument panels, real-time monitoring, and KPI displays with needles, ranges, and scales.
Implement Syncfusion Windows Forms Smith Chart (SfSmithChart) control for visualizing impedance and admittance of transmission lines. Use this when working with high-frequency circuit applications, RF engineering, or transmission line analysis. Covers normalized resistance/reactance plotting, normalized conductance/susceptance visualization, series configuration, axes customization, markers, legends, and tooltips.
Implement Syncfusion Windows Forms TreeMap control for hierarchical data visualization with nested rectangles. Use this when working with TreeMap controls, hierarchical data display, space-efficient charts, or color-coded visualizations. Covers TreeMap configuration, layout algorithms, color mapping, and customization for stock market visualizations and data categorization with rectangular layouts.
Plan, run and report systematic reviews and meta-analyses to PRISMA 2020 standards. Use when writing a review protocol (PROSPERO/OSF), building reproducible database search strings (PubMed, Embase, Scopus, Web of Science), deduplicating exports, organizing title/abstract and full-text screening, assessing risk of bias, drawing the PRISMA flow diagram, or completing the PRISMA 2020 checklist. Includes a dedup tool and a checked PRISMA diagram generator.
Track and analyze US government shutdown liquidity impacts by monitoring TGA (Treasury General Account), bank reserves, EFFR, and SOFR data from FRED API. Use when user wants to (1) analyze current or past government shutdown effects on financial markets, (2) track liquidity conditions during fiscal policy disruptions, (3) assess "stealth tightening" effects, (4) compare shutdown episodes across different monetary policy regimes (QE vs QT), or (5) generate liquidity stress reports with historical context. Recommended usage frequency is weekly on Wednesdays after TGA/reserve data releases.
Use when the user wants a VALID experiment instead of a guess — hypothesis, one variable, sample size and runtime math, statistical significance, primary versus secondary metrics, multi-arm designs, and a results template, across Optimizely, VWO, and native Meta and Google tests. Trigger on 'A/B test', 'split test', 'how long should I run the test', 'is this result significant', 'test two versions', 'which creative is actually better'. Also use when a winner was declared after two days on tiny numbers. Not for — scaling the proven winner, see `55-scaling-ads-global`; analyzing data already collected, see `13-data-analysis-global`; auditing the account, see `21-ads-audit-global`.
Audit and exactly repair a Yandex KIT catalog. Use for Russian requests such as «Проверь каталог», «Проведи глубокий аудит», «Проверь группировку, карточки, медиа или коллекции», «Поставь цену 4 990 для SKU-42», «Исправь остатки по этому файлу» and exact single or bulk catalog fixes. Audits are read-only; writes require an explicit command with an exact target, operation and owner value or named authoritative source. Use a1-yandex-kit-operator instead for a fast current store signal.
Add Temps analytics to React applications with comprehensive tracking capabilities including page views, custom events, scroll tracking, engagement monitoring, session recording, and Web Vitals performance metrics. Use when the user wants to: (1) Add analytics to a React app (Next.js App Router, Next.js Pages Router, Vite, Create React App, or Remix), (2) Track user events or interactions, (3) Monitor scroll depth or element visibility, (4) Add session recording/replay, (5) Track Web Vitals or performance metrics, (6) Measure user engagement or time on page, (7) Set up product analytics or telemetry. Triggers: "add analytics", "track events", "session recording", "web vitals", "user tracking", "temps analytics", "react analytics".
Migrates a local SQL Server setup to the Azure SQL Database container for Azure-faithful local development. Use when a project already uses mcr.microsoft.com/mssql/server, mssql/server, an sqlcmd plus SA password docker setup, a "SQL Server in docker" or "local mssql container", or a docker-compose with the mssql/server image; and use when the user asks for "SQL Server locally", "run mssql in Docker", "spin up a local SQL database", or "test against SQL Server" but actually wants the Azure SQL Database engine (EngineEdition 5). Detects the SQL Server image, rewrites it to the Azure SQL Database container, lets the runtime select the image architecture, keeps the SA login, flags SQL Server-only features (SQL Agent, FILESTREAM, full Service Broker, cross-server distributed transactions, Windows Auth), and re-points connection strings from master to a provisioned user database.
Imports an existing Azure SQL Database or SQL Server schema and data INTO the local Azure SQL Database container using SqlPackage. Use when asked to "import a bacpac", "load my existing database locally", "restore a dacpac into the container", "bring my prod schema into the dev container", "run my .bacpac/.dacpac against the local Azure SQL engine", or migrate an existing database into the preview container. Handles provisioning the target database on master first, then running SqlPackage /Action:Import against the provisioned user database. Use this for any "get my real database running in the container" request instead of hand-writing SqlPackage flags.
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.