skills-mcp

Write GitHub Actions CI/CD workflows - automated testing, linting, building, Docker image publishing, deployment, and release management. Covers workflow syntax, triggers, jobs, steps, matrix builds, caching dependencies, secrets management, environment protection rules, and reusable workflows. Use when setting up CI/CD pipelines, automating tests on pull requests, deploying applications, publishing packages, or configuring GitHub Actions workflows.

Master reference for the Skill MCP server - a self-hosted agent skills registry that delivers curated, expert instructions for common AI tasks via 6 MCP tools. Covers the full 3-tier progressive disclosure workflow: Tier 1 discovery (skills_find_relevant), Tier 2 loading (skills_get_body, skills_get_options), and Tier 3 supplementary resources (skills_get_reference, skills_run_script, skills_get_asset). Use this skill whenever you need to understand, operate, or get the most out of the Skill MCP server.

Build web applications with Django - models, views, URL routing, middleware, Django ORM, authentication, testing, deployment, and best practices

Build interactive user interfaces with Vue.js - components, state management, templates, composition API, lifecycle hooks, and reactive data binding

Integrate with REST APIs, handle authentication (API keys, OAuth 2.0, JWT), manage pagination, rate limiting, retries, and error handling. Use when the user needs to call an external API, integrate a third-party service, handle webhooks, or build an API client.

Build, debug, and optimize applications using the Anthropic Claude API and official SDKs. Covers single API calls, tool use, streaming, prompt caching, vision/multimodal, extended thinking, batch processing, and managed agents. Use when writing code that imports the anthropic SDK, working with Claude models (Opus, Sonnet, Haiku), building agentic pipelines, or migrating between Claude model versions.

Build and deploy applications on the Cloudflare developer platform. Covers Workers (serverless edge compute), Pages (full-stack web), Durable Objects (stateful coordination), KV (key-value storage), D1 (SQLite edge database), R2 (object storage), Workers AI (LLM inference), Vectorize (vector database), Queues, and Wrangler CLI. Use when building Cloudflare Workers, deploying to Cloudflare Pages, or integrating any Cloudflare storage, AI, or networking product.

Review code for correctness, bugs, security vulnerabilities, performance issues, and best practices. Produces structured findings with severity ratings and concrete fixes. Use when the user asks to review code, check for bugs, find security issues, audit code quality, or get a second opinion on an implementation.

Analyze CSV and tabular data to generate insights, summaries, statistics, and visualizations. Handle data cleaning, aggregation, grouping, pivot tables, and outlier detection. Use when the user needs to analyze data, understand a dataset, find patterns, clean data, or generate a data summary.

Containerize applications with Docker and orchestrate with Docker Compose. Covers writing production-ready Dockerfiles for Python, Node.js, and Go applications, multi-stage builds, security hardening (non-root users, minimal base images), Docker Compose for local development and multi-service stacks, health checks, volumes, networking, and optimization for layer caching. Use when containerizing an app, writing a Dockerfile, setting up Docker Compose, or debugging container issues.

Create, read, and edit Microsoft Word (.docx) documents. Generate reports, contracts, letters, and structured documents with headings, tables, images, lists, and formatting. Use when the user needs to create a Word document, edit a .docx file, generate a report, or export content to Word format.

Build production-grade REST APIs with FastAPI and Python. Covers request/response models with Pydantic v2, path/query/body parameters, dependency injection, authentication (JWT, OAuth2), background tasks, async database access with SQLAlchemy 2.0, middleware, error handling, OpenAPI documentation, and testing with pytest and httpx. Use when building a Python REST API, adding endpoints to a FastAPI app, writing FastAPI middleware, or setting up FastAPI with a database.

Design and implement distinctive, production-grade web interfaces that avoid generic AI aesthetics. Covers establishing an aesthetic direction, typography selection, color system design, layout composition, micro-animations, responsive design, and design-to-code workflow. Use when designing a website, landing page, dashboard, or UI component from scratch, or when existing UI feels generic and needs a distinctive visual identity.

Build applications with the Google Gemini API and Google GenAI SDKs. Covers model selection (Gemini 2.5 Pro/Flash and newer), multimodal content (text, images, audio, video), function calling, structured outputs, streaming, and embeddings across Python, JavaScript/TypeScript, Go, and Java SDKs. Use when integrating Gemini models, working with Google AI Studio, or migrating from legacy google-generativeai / @google/generative-ai SDKs.

Generate standardised git commit messages following the Conventional Commits specification. Analyzes staged changes or diffs to produce atomic, descriptive commits. Use when the user wants to write a commit message, commit changes, follow conventional commits format, or generate git history.

Design and implement GraphQL APIs - schema definition, resolvers, mutations, subscriptions, authentication, DataLoader for N+1 prevention, pagination (cursor and offset), error handling, and testing. Covers both server-side (GraphQL Yoga, Apollo Server, Strawberry for Python) and client-side (Apollo Client, urql, TanStack Query). Use when building a GraphQL API, adding GraphQL to an existing service, writing GraphQL queries/mutations/subscriptions, or optimizing GraphQL performance.

Design effective prompts for large language models - chain-of-thought reasoning, few-shot examples, role assignment, XML structure, output format control, temperature tuning, and prompt chaining for multi-step tasks. Covers both general LLM prompting and model-specific best practices for Claude, GPT-4o, and Gemini. Use when writing system prompts, designing AI agent instructions, improving LLM output quality, reducing hallucinations, or engineering prompts for production AI applications.

Build high-quality Model Context Protocol (MCP) servers that expose tools, resources, and prompts to AI agents. Covers MCP architecture, tool/resource/prompt definitions, stdio and HTTP/SSE transports, authentication, error handling, and deployment patterns using FastMCP (Python) or the TypeScript MCP SDK. Use when creating an MCP server, adding tools to an existing MCP server, integrating an API or service as MCP tools, or debugging MCP connectivity issues.

Apply Next.js App Router best practices - file conventions, React Server Components, async APIs (Next.js 15+), data fetching patterns, metadata, image/font optimization, error handling, route handlers, bundling, and self-hosting. Use when writing or reviewing Next.js code, building full-stack React apps, or migrating from the Pages Router.

Build applications using the OpenAI API - chat completions, function calling, streaming, embeddings, image generation (DALL-E), vision (GPT-4 Vision), audio (Whisper, TTS), assistants, and structured outputs. Covers Python and TypeScript OpenAI SDK, model selection (GPT-4o, o3, o4-mini), cost optimization with prompt caching, and batch processing. Use when integrating OpenAI models, building GPT-powered features, working with the openai Python package or openai npm package.

Extract text, tables, and metadata from PDF files. Fill PDF forms, merge multiple PDFs, split documents, and convert PDF content to structured formats. Use when the user asks to read a PDF, extract data from a PDF, fill a form, merge PDFs, or work with PDF documents.

Create, edit, and analyze PowerPoint presentations (.pptx files). Covers building slides from scratch with pptxgenjs, reading and extracting text with markitdown, editing existing presentations by unpacking and modifying XML, and converting slides to images for review. Use when the user wants to create a presentation, add slides, modify an existing PPTX file, or generate a slide deck programmatically.

Apply React best practices - hooks patterns, state management, component composition, performance optimization, error boundaries, accessibility, and testing. Covers React 18+, Suspense, concurrent features, custom hooks, Context API, and TypeScript integration. Use when writing React components, reviewing React code, optimizing renders, or designing component architecture.

Create professional, comprehensive README.md files for software projects. Generates badges, installation instructions, usage examples, API references, and contributing guides tailored to the project type. Use when the user needs a README, wants to improve project documentation, or needs to write docs for a new repo.

Write optimized SQL queries for complex data requirements. Handle multi-table joins, aggregations, window functions, CTEs, subqueries, and performance optimization. Use when the user needs to write a SQL query, optimize a slow query, understand a query plan, design a schema, or work with database data.

Build Stripe payment integrations - one-time payments, subscriptions, marketplaces, and Connect platforms. Covers API selection (Checkout Sessions vs PaymentIntents vs Setup Intents), webhooks, restricted API keys, Stripe Connect (Accounts v2), billing, and security best practices. Use when accepting payments, integrating Stripe, building subscription billing, creating connected accounts, or reviewing Stripe integration code.

Integrate Supabase into applications - PostgreSQL database queries, Row Level Security (RLS) policies, authentication (email/password, OAuth, magic link), real-time subscriptions, storage (file uploads), and Edge Functions. Covers the Supabase JavaScript/TypeScript SDK, Python client, type-safe queries with generated types, and local development with the Supabase CLI. Use when building with Supabase, setting up auth, writing RLS policies, querying a Supabase database, or managing file storage.

Write and manage infrastructure as code with HashiCorp Terraform. Covers resource definitions, variables, outputs, modules, state management, workspaces, provider configuration (AWS, GCP, Azure, Cloudflare), remote backends (S3, Terraform Cloud), data sources, locals, lifecycle rules, and CI/CD integration. Use when provisioning cloud infrastructure, writing Terraform modules, managing Terraform state, or importing existing resources.

Write unit tests, integration tests, and end-to-end tests following best practices. Covers pytest, Jest, Go testing, and JUnit. Generates tests for functions, classes, REST endpoints, and database operations. Use when the user wants to write tests, increase test coverage, practice TDD, or add a test suite to existing code.

Apply advanced TypeScript patterns - strict type safety, utility types, generics, discriminated unions, branded types, conditional types, template literal types, and declaration merging. Covers tsconfig best practices, type-safe API design, narrowing, and eliminating any/unknown. Use when writing TypeScript, reviewing TypeScript types, designing type-safe APIs, debugging type errors, or improving type coverage in a codebase.

Build complex, self-contained interactive web artifacts - rich HTML/CSS/JavaScript applications, React components with Tailwind CSS and shadcn/ui, data visualizations with D3.js or Chart.js, interactive tools, and games - all bundled into a single portable HTML file. Use when the user wants to create an interactive demo, a data visualization, a UI prototype, a mini web app, or any rich browser-rendered artifact that can be embedded or shared as a single file.

Scrape and extract structured content from websites. Handle pagination, dynamic JavaScript-rendered pages, rate limiting, and anti-bot measures. Use when the user needs to extract data from a website, scrape product listings, collect article text, or automate web data collection.

Create, edit, and analyze Excel spreadsheets (.xlsx, .xlsm, .csv, .tsv files). Covers building spreadsheets with openpyxl and pandas, writing Excel formulas, applying professional formatting and color conventions, creating charts, and validating formula correctness. Use when the user wants to create an Excel file, analyze spreadsheet data, build a financial model, generate a report as XLSX, or edit an existing spreadsheet.