{
  "markdown": "# Claude MPM Skills\n\n> Production-ready Claude Code skills for intelligent project development\n\n## Overview\n\nThis repository contains a comprehensive collection of **174 Claude Code skills** designed for the Claude Multi-Agent Project Manager (MPM) ecosystem. Skills cover modern development workflows with **95%+ coverage** across Python, TypeScript, JavaScript, Golang, PHP, Rust, Elixir, AI, and universal tooling.\n\n## What is Claude MPM?\n\n**Claude MPM (Multi-Agent Project Manager)** is an advanced orchestration framework that runs within Claude Code (Anthropic's official CLI). It enables:\n\n- **Multi-Agent Coordination**: Specialized agents for different tasks (research, engineering, QA, ops)\n- **Intelligent Delegation**: PM agent coordinates work across specialist agents\n- **Context Management**: Efficient token usage with progressive disclosure\n- **Skill System**: Modular, reusable knowledge bases (this repository)\n\n**Key Components:**\n- **Claude Code**: Anthropic's official CLI environment\n- **Claude MPM**: Multi-agent framework running in Claude Code\n- **Skills**: Domain-specific knowledge modules (this repo contains 174 skills)\n\n**How They Work Together:**\n```\nClaude Code (CLI)\n    ↓\nClaude MPM (Multi-Agent Framework)\n    ↓\nSkills (Knowledge Modules) ← You are here\n```\n\n## Features\n\n- **Progressive Loading**: Skills load on-demand with compact entry points, expanding to full references when needed\n- **Token Efficiency**: ~87% token savings during discovery phase\n- **Toolchain Detection**: Automatically deploy relevant skills based on project type\n- **Production-Ready**: All skills include real-world examples, best practices, and troubleshooting\n- **Research-Backed**: Built on latest 2025 techniques and industry patterns\n\n## Quick Stats\n\n- **Total Skills**: 174 production-ready skills\n- **Coverage**: 95%+ of modern development workflows\n- **Token Efficiency**: ~66.7k entry tokens vs ~512.4k full tokens (~87% savings)\n- **Categories**: Python, TypeScript, JavaScript, Golang, PHP, Rust, Elixir, Next.js, UI, AI, Platforms, Universal\n- **Complete Stacks**: Full-stack TypeScript, Python Web, React Frontend, AI Workflows\n\n## Repository Structure\n\n```\nclaude-mpm-skills/\n├── toolchains/          # Language/framework-specific skills (130 skills)\n│   ├── python/         # 11 skills\n│   │   ├── frameworks/     # Django, FastAPI, Flask\n│   │   ├── testing/        # pytest\n│   │   ├── data/           # SQLAlchemy\n│   │   ├── async/          # asyncio, Celery\n│   │   ├── tooling/        # mypy, pyright\n│   │   └── validation/     # Pydantic\n│   ├── typescript/     # 14 skills\n│   │   ├── frameworks/     # React, Vue, Node.js backend, Fastify\n│   │   ├── testing/        # Vitest, Jest\n│   │   ├── data/           # Drizzle, Kysely, Prisma\n│   │   ├── validation/     # Zod\n│   │   ├── state/          # Zustand, TanStack Query\n│   │   ├── api/            # tRPC\n│   │   └── build/          # Turborepo\n│   ├── javascript/     # 22 skills\n│   │   ├── frameworks/     # React, Vue, Svelte, SvelteKit\n│   │   ├── testing/        # Playwright, Cypress\n│   │   ├── build/          # Vite\n│   │   └── tooling/        # Biome\n│   ├── php/            # 6 skills\n│   │   ├── frameworks/     # WordPress, EspoCRM\n│   │   └── testing/        # PHPUnit, PHPCS\n│   ├── golang/         # 7 skills\n│   │   ├── web/            # net/http, Chi, Gin, Echo, Fiber\n│   │   ├── testing/        # Go testing, testify, httptest\n│   │   ├── data/           # SQL, migrations, ORMs/query builders\n│   │   ├── cli/            # CLI tooling patterns\n│   │   └── observability/  # Logging and telemetry\n│   │   ├── grpc/           # Protobuf APIs, interceptors, streaming\n│   │   └── concurrency/    # errgroup, worker pools, bounded fan-out\n│   ├── rust/           # 7 skills\n│   │   ├── frameworks/     # Tauri, Axum\n│   │   ├── cli/            # Clap\n│   │   └── desktop-applications/ # Desktop app patterns\n│   ├── elixir/         # 4 skills\n│   │   ├── frameworks/     # Phoenix + LiveView (BEAM), Phoenix API + Channels\n│   │   ├── data/           # Ecto patterns\n│   │   └── ops/            # Phoenix operations & releases\n│   ├── nextjs/         # 3 skills\n│   │   ├── nextjs-core/    # Next.js fundamentals\n│   │   └── nextjs-v16/     # Next.js 16 (Turbopack, cache components)\n│   ├── ui/             # 4 skills\n│   │   ├── styling/        # Tailwind CSS\n│   │   └── components/     # shadcn/ui, DaisyUI, Headless UI\n│   ├── ai/             # 11 skills\n│   │   ├── sdks/           # Anthropic SDK\n│   │   ├── frameworks/     # LangChain, DSPy, LangGraph\n│   │   ├── services/       # OpenRouter\n│   │   ├── protocols/      # MCP\n│   │   └── techniques/     # Session Compression\n│   ├── platforms/      # 26 skills\n│   │   ├── deployment/     # Vercel (9), Netlify, DigitalOcean (10)\n│   │   ├── database/       # Neon\n│   │   ├── backend/        # Supabase\n│   │   ├── observability/  # Datadog\n│   │   └── auth/           # Better Auth (4)\n│   ├── databases/      # 1 skill\n│   │   └── mongodb/        # MongoDB\n│   ├── visualbasic/    # 4 skills\n│   └── universal/      # 9 skills\n│       └── (cross-language infrastructure patterns)\n└── universal/           # 42 skills\n    ├── infrastructure/     # Docker, GitHub Actions\n    ├── data/              # GraphQL\n    ├── architecture/      # Software patterns\n    └── testing/           # TDD, systematic debugging\n```\n\n## Complete Skill Catalog\n\n### Python (11 Skills)\n\n**Frameworks**:\n- Django - Full-featured web framework with ORM, admin, DRF\n- FastAPI - Modern async API framework with automatic OpenAPI\n- Flask - Lightweight WSGI framework for microservices\n\n**Testing**:\n- pytest - Fixtures, parametrization, plugins, FastAPI/Django integration\n\n**Data & ORM**:\n- SQLAlchemy - Modern ORM with 2.0 syntax, async, Alembic migrations\n\n**Async & Background Jobs**:\n- asyncio - Async/await patterns, event loops, concurrent programming\n- Celery - Distributed task queues, periodic tasks, workflows\n\n**Type Checking**:\n- mypy - Static type checker with strict mode\n- pyright - Fast type checker with VS Code integration\n\n**Validation**:\n- Pydantic - Data validation with type hints, FastAPI/Django integration\n\n### TypeScript (14 Skills)\n\n**Frameworks**:\n- React - Hooks, context, performance optimization\n- Vue 3 - Composition API, Pinia, TypeScript integration\n- Node.js Backend - Express/Fastify with Drizzle/Prisma\n- Fastify - Schema-first, high-performance backend with typed routes\n\n**Testing**:\n- Vitest - Modern testing with React/Vue\n- Jest - TypeScript testing with ts-jest\n\n**Data & ORMs**:\n- Drizzle - TypeScript-first ORM with migrations\n- Kysely - Type-safe SQL query builder\n- Prisma - Next-gen ORM with migrations and client generation\n\n**Validation**:\n- Zod - Schema validation with type inference\n\n**State Management**:\n- Zustand - Minimal React state management\n- TanStack Query - Server state, caching, optimistic updates\n\n**API**:\n- tRPC - End-to-end type safety without codegen\n\n**Build Tools**:\n- Turborepo - Monorepo with intelligent caching\n\n### JavaScript (22 Skills)\n\n**Frameworks**:\n- React Core - Component patterns, JSX, props/state, core hooks, composition\n- React Hooks Composition - Advanced custom-hook patterns (SWR, debounce, memoized contexts)\n- React State Machine - XState v5 finite-state modeling for complex UI flows\n- React Advanced - React 19 platform + rendering architecture (React Compiler, concurrent, Actions, virtualization, RSC boundary)\n- FlexLayout (React) - Docking layout manager with drag-and-drop panels\n- Vue - Progressive framework (also in TypeScript)\n- Svelte - Reactive framework with runes\n- SvelteKit - Full-stack Svelte with SSR/SSG\n- Svelte 5 Runes + adapter-static - Hydration-safe state and store bridges\n\n**Testing**:\n- Playwright - Cross-browser E2E testing with Page Object Model\n- Cypress - Browser E2E testing with network stubbing and component testing\n\n**Build Tools**:\n- Vite - Fast build tool with HMR\n\n**Tooling**:\n- Biome - Fast linter and formatter (Rust-powered)\n\n### PHP (6 Skills)\n\n**WordPress Ecosystem**:\n- wordpress-advanced-architecture - REST API, WP-CLI, performance optimization, caching strategies\n- wordpress-block-editor - Block themes, FSE architecture, theme.json, custom Gutenberg blocks\n- wordpress-testing-qa - PHPUnit integration tests, WP_Mock unit tests, PHPCS coding standards\n\n**Enterprise**:\n- espocrm-development - EspoCRM customization, entity management, API extensions\n- espocrm-advanced-features - Advanced workflows, complex business logic implementation\n- espocrm-deployment - Production deployment, security hardening, performance tuning\n\n### Golang (7 Skills)\n\n**Web & HTTP**:\n- golang-http-frameworks - net/http, Chi, Gin, Echo, Fiber patterns\n\n**gRPC**:\n- golang-grpc - Protobuf APIs, interceptors, streaming, bufconn testing\n\n**Concurrency**:\n- golang-concurrency-patterns - Context, errgroup, worker pools, bounded fan-out\n\n**Testing**:\n- golang-testing-strategies - Table-driven tests, testify, gomock, benchmarks\n\n**Data**:\n- golang-database-patterns - SQL patterns, migrations, query builders\n\n**CLI**:\n- golang-cli-cobra-viper - Cobra/Viper CLI structure and config\n\n**Observability**:\n- golang-observability-opentelemetry - Logging/metrics/traces + middleware patterns\n\n### Rust (6 Skills)\n\n**Web & Desktop**:\n- axum - Production Rust HTTP APIs with Tower middleware\n- desktop-applications - Rust desktop app architecture and integration patterns\n- tauri - Cross-platform desktop apps with Rust backend and web frontend\n\n**CLI**:\n- clap - Rust CLI parsing, subcommands, config layering, testable binaries\n\n### Next.js (3 Skills)\n\n- Next.js Core - App Router, Server Components, Server Actions\n- Next.js v16 - Turbopack, cache components, migration guide\n\n### UI & Styling (4 Skills)\n\n**CSS Frameworks**:\n- Tailwind CSS - Utility-first CSS with JIT mode\n\n**Component Libraries**:\n- shadcn/ui - Copy-paste components with Radix UI + Tailwind\n- DaisyUI - Tailwind plugin with 50+ components and themes\n- Headless UI - Unstyled accessible primitives for React/Vue\n\n### AI & LLM (10 Skills)\n\n**SDKs**:\n- Anthropic SDK - Messages API, streaming, function calling, vision\n\n**Frameworks**:\n- LangChain - LCEL, RAG, agents, chains, memory\n- DSPy - Automatic prompt optimization with MIPROv2\n- LangGraph - Stateful multi-agent orchestration\n\n**Services**:\n- OpenRouter - Unified LLM API access\n\n**Protocols**:\n- MCP - Model Context Protocol\n\n**Techniques**:\n- Session Compression - Context window compression, progressive summarization\n\n### Platforms (26 Skills)\n\n**Deployment**:\n- Vercel - Next.js deployment, Edge Functions, serverless\n- Netlify - JAMstack, Forms, Identity, Edge Functions\n\n**Database**:\n- Neon - Serverless Postgres with branching\n\n**Backend**:\n- Supabase - Postgres + Auth + Storage + Realtime + RLS\n\n### Universal (42 Skills)\n\n**Infrastructure**:\n- Docker - Containerization, multi-stage builds, compose\n- GitHub Actions - CI/CD workflows, matrix strategies, deployments\n- Kubernetes - Workloads, probes, rollouts, debugging runbook, hardening\n- Terraform - IaC workflow: state, modules, environments, CI guardrails\n\n**Observability**:\n- OpenTelemetry - Traces/metrics/logs, OTLP + Collector pipelines, sampling, troubleshooting\n\n**Security**:\n- Threat Modeling - STRIDE workshops, threat registers, mitigations → tickets + tests\n\n**Data**:\n- GraphQL - Schema-first APIs, Apollo, resolvers, subscriptions\n\n**Architecture**:\n- Software Patterns - Design patterns, anti-patterns, decision trees\n\n**Testing & Debugging**:\n- TDD - Test-driven development workflows\n- Systematic Debugging - Root cause analysis\n\n**Orchestration**:\n- MPM Orchestration Demo - Reference workflow for the Command → Agent → Skill pattern, covering preloaded skills, dynamic `Skill` invocation, and the `context: fork` pattern for forked sub-agents\n\n## Installation\n\n### Prerequisites\n\n- **Claude Code** (Anthropic's official CLI)\n- **Claude MPM** framework\n\n### Step 1: Install Claude MPM\n\n```bash\n# Install via pip (recommended)\npip install claude-mpm\n\n# Or install via Homebrew (macOS)\nbrew tap bobmatnyc/tools\nbrew install claude-mpm\n\n# Or install from source\ngit clone https://github.com/bobmatnyc/claude-mpm.git\ncd claude-mpm\npip install -e .\n```\n\n### Step 2: Initialize Claude MPM in Your Project\n\n```bash\n# Navigate to your project directory\ncd your-project\n\n# Initialize Claude MPM\n/mpm-init\n```\n\nThis creates `.claude-mpm/` directory with configuration and agent setup.\n\n### Step 3: Deploy Skills (Automatic)\n\n```bash\n# Auto-detect your project stack and deploy relevant skills\n/mpm-auto-configure\n\n# Or use the agent auto-configuration\n/mpm-agents-auto-configure\n```\n\nSkills are automatically selected based on:\n- `package.json` → TypeScript/JavaScript skills\n- `pyproject.toml` → Python skills\n- Framework configs → Next.js, React, Django, FastAPI\n- Dependencies → AI frameworks (LangChain, Anthropic)\n\n### Step 4: Verify Installation\n\n```bash\n# Check MPM status\n/mpm-status\n\n# List available skills\n/mpm-agents-list\n\n# View deployed agents\n/mpm-agents\n```\n\n### Manual Skill Installation (Alternative)\n\nClone this repository to make skills available to Claude MPM:\n\n```bash\n# Clone skills repository\ngit clone https://github.com/bobmatnyc/claude-mpm-skills.git\n\n# Link to Claude MPM skills directory\nln -s $(pwd)/claude-mpm-skills ~/.claude-mpm/skills\n```\n\n## Using these skills outside claude-mpm\n\nThese skills are built for the [claude-mpm](https://github.com/bobmatnyc/claude-mpm) **passive-injection model**. In that model, the framework injects skill content into the PM agent's context as reference material and lets the orchestrator decide when to apply it — skills deliberately do not self-trigger on keyword matches.\n\nTwo frontmatter fields encode that design and are intentional, not oversights:\n\n- `user-invocable: false` — suppresses slash-command generation, since these skills are applied by the orchestrator rather than invoked directly by a user.\n- `disable-model-invocation: true` — includes the skill as passive context rather than registering it as a callable, auto-triggering tool.\n\nThis behavior is load-bearing for multi-agent dispatch correctness within claude-mpm.\n\n### Standalone usage in a generic `.claude/skills/` setup\n\nGeneric installers that rely on auto-triggering — such as a plain `.claude/skills/` directory, `autoskills`, or `npx skills add` — expect skills to fire on keyword matches. With the two fields above present, these skills will not auto-trigger under those tools.\n\nTo use a skill standalone, copy its `SKILL.md` content into the target `.claude/skills/` location and strip the two fields so the skill can auto-trigger:\n\n```diff\n ---\n name: database-migration\n description: Safe patterns for evolving database schemas in production.\n-user-invocable: false\n-disable-model-invocation: true\n tags: [database, migration, schema, production]\n ---\n```\n\nAfter removing those lines, the skill registers as a normal auto-triggering skill in the generic setup.\n\n### Known limitation\n\nThese skills are **incompatible with auto-triggering installers as shipped**. Auto-triggering tools (`autoskills`, `npx skills add`, and similar) will not fire the skills until the `user-invocable` and `disable-model-invocation` fields are removed per the workaround above. This is a deliberate trade-off in favor of claude-mpm dispatch correctness, not a bug.\n\n## Usage\n\n### Automatic Deployment (Recommended)\n\n```bash\n# Initialize project with Claude MPM\n/mpm-init\n\n# Or use auto-configuration to detect toolchain\n/mpm-auto-configure\n\n# Deploy recommended skills based on project detection\n/mpm-agents-auto-configure\n```\n\nSkills are automatically deployed based on detected toolchain:\n- `package.json` → TypeScript/JavaScript skills\n- `pyproject.toml` or `requirements.txt` → Python skills\n- Framework configs → Next.js, React, Django, FastAPI skills\n- AI dependencies → LangChain, Anthropic, DSPy skills\n\n### Manual Skill Access\n\nSkills use progressive loading - entry points load first for quick reference:\n\n```markdown\n---\nprogressive_disclosure:\n  entry_point:\n    summary: \"Brief description (60-95 tokens)\"\n    when_to_use:\n      - \"Use case 1\"\n      - \"Use case 2\"\n    quick_start:\n      - \"Step 1\"\n      - \"Step 2\"\n---\n```\n\nFull documentation expands on-demand when needed.\n\n## Complete Development Stacks\n\n### Full-Stack TypeScript\n```\nNext.js + tRPC + TanStack Query + Zustand + Zod + Prisma +\nTailwind + shadcn/ui + Turborepo + Docker + GitHub Actions\n```\n**Coverage**: 100% - All skills available\n\n### Python Web Development\n```\nFastAPI/Django + Pydantic + SQLAlchemy + Celery +\npytest + mypy + Docker + GitHub Actions\n```\n**Coverage**: 100% - All skills available\n\n### Modern React Frontend\n```\nReact + TanStack Query + Zustand + Tailwind + shadcn/ui +\nVite + Vitest + Playwright\n```\n**Coverage**: 100% - All skills available\n\n### AI/LLM Applications\n```\nAnthropic SDK + LangChain + DSPy + LangGraph +\nSession Compression + OpenRouter + MCP\n```\n**Coverage**: 100% - All skills available\n\n## Progressive Loading Design\n\nSkills use a two-tier structure for optimal token efficiency:\n\n### Entry Point (60-200 tokens, depending on skill depth)\n- Skill name and summary\n- When to use (3-5 scenarios)\n- Quick start (3-5 steps)\n\n### Full Documentation (3,000-6,000 tokens)\n- Complete API reference\n- Real-world examples\n- Best practices\n- Framework integrations\n- Production patterns\n- Testing strategies\n- Troubleshooting\n\n**Token Savings**: ~87% during discovery (load 110 entry points vs all full docs)\n\n## Performance Benchmarks\n\n- **Discovery Phase**: 66,690 tokens (all 110 entry points) vs 512,411 tokens (all full docs)\n- **Token Efficiency**: ~87% reduction during skill browsing\n- **Coverage**: 95%+ of modern development workflows\n- **Production Adopters**: Skills based on patterns from JetBlue, Databricks, Walmart, VMware\n- **Token Reporting**: `python scripts/token_report.py --manifest manifest.json --out stats/token-summary.json` for CI/dashboard consumption\n\n## Contributing\n\nWe welcome contributions! See [CONTRIBUTING.md](docs/community/CONTRIBUTING.md) for guidelines.\n\n**Governance**: All merges to main require approval from @bobmatnyc (see [GOVERNANCE.md](docs/community/GOVERNANCE.md))\n\n### Skill Format Requirements\n\n1. **Progressive Disclosure**: YAML frontmatter with entry_point section\n2. **Token Budgets**: Entry 60-200 tokens, Full 3,000-6,000 tokens\n3. **Metadata**: Complete metadata.json with tags, related_skills, token estimates\n4. **Examples**: Real-world code examples with error handling\n5. **Versioning**: Semantic versioning (see [docs/VERSIONING.md](docs/VERSIONING.md))\n\n## Documentation\n\n### User Documentation\n\n- **[User Guide](docs/USER_GUIDE.md)** - Understanding and using Claude Code skills\n- **[Troubleshooting](docs/TROUBLESHOOTING.md)** - Common issues and solutions\n\n### Developer Documentation\n\n- **[Skill Creation Guide](docs/SKILL_CREATION_GUIDE.md)** - Building your own skills\n- **[Best Practices](docs/SKILL_SELF_CONTAINMENT_STANDARD.md)** - Self-containment standards\n- **[Contributing](docs/community/CONTRIBUTING.md)** - Contribution guidelines\n- **[Versioning Policy](docs/VERSIONING.md)** - Semantic versioning for skills\n\n### Architecture & Research\n\n- **[Architecture](docs/architecture/STRUCTURE.md)** - Repository structure\n- **[Research Documents](docs/research/)** - Pattern analysis and guides\n  - Python, TypeScript, Ruby, Rust, PHP, Java, Go advanced patterns\n  - Skills compliance analysis\n  - Coverage analysis\n\n### Reference\n\n- **[PR Checklist](docs/SKILL_CREATION_PR_CHECKLIST.md)** - Submission requirements\n- **[GitHub Setup](docs/infrastructure/GITHUB_REPOSITORY_SETUP.md)** - Repository configuration\n\n## License\n\nMIT License - See [LICENSE](LICENSE)\n\n## Links\n\n- **Claude MPM Framework**: https://github.com/bobmatnyc/claude-mpm\n- **Claude MPM Documentation**: https://github.com/bobmatnyc/claude-mpm/tree/main/docs\n- **Skills Documentation**: [docs/USER_GUIDE.md](docs/USER_GUIDE.md)\n- **Skill Creation Guide**: [docs/SKILL_CREATION_GUIDE.md](docs/SKILL_CREATION_GUIDE.md)\n- **Issues**: https://github.com/bobmatnyc/claude-mpm-skills/issues\n- **Discussions**: https://github.com/bobmatnyc/claude-mpm-skills/discussions\n\n## Acknowledgments\n\nBuilt with research from:\n- Official framework documentation (2025 versions)\n- Industry best practices (JetBlue, Databricks, Walmart, VMware, Replit)\n- Academic research (DSPy, LLMLingua, prompt optimization studies)\n- Community feedback and contributions\n\n---\n\n**Last Updated**: 2025-12-17\n**Skills Count**: 110\n**Coverage**: 95%+\n**Token Efficiency**: ~87%\n",
  "bytes": 20601,
  "sha": "3b7328e27cc5c37bb6fcbe847cb57d131dce5d5f3112562241a9d31800639449",
  "repo_slug": "bobmatnyc/claude-mpm-skills",
  "fonte": "repo",
  "truncated": false,
  "api": "https://agentalog.com/api/listings/skl_bobmatnyc_claude_mpm_skills_drizzle_migr_1588b285/readme"
}