bobmatnyc/claude-mpm-skills
13 skills15.7k installs689 starsGitHub
Install
npx skills add https://github.com/bobmatnyc/claude-mpm-skillsSkills in this repo
1Drizzle OrmDrizzle ORM is an agent skill that teaches solo and indie builders how to use a modern TypeScript-first SQL layer with no ORM runtime bloat, compile-time type safety, and syntax that stays close to SQL. It fits anyone shipping a SaaS API, internal tool, or CLI that must talk to PostgreSQL, MySQL, or SQLite without giving up editor autocomplete on joins and filters. The skill walks through installation, basic pgTable schema definition, client wiring with node-postgres, and when to reach for deeper topics in bundled references such as query patterns, performance tuning, advanced schemas, and a Prisma comparison. Use it while you are still choosing or implementing persistence—not after the schema is frozen—so migrations and types stay aligned. For Prism’s build-phase audience, it is practical procedural knowledge rather than a hosted integration: your agent applies the patterns inside your repo. Pair it with your stack’s auth and API layers when you need reliable, reviewable data models on edge or serverless deploy targets.4.3kinstalls2Playwright E2e TestingPlaywright E2E Testing is an agent skill that teaches solo and indie builders how to use Playwright’s modern test runner for real browser automation. It is aimed at anyone shipping a web app or SaaS who needs repeatable checks on login, checkout, settings, and other multi-step UI paths without maintaining a separate QA team. The skill emphasizes when to reach for E2E coverage—cross-browser needs, interaction-heavy flows, and evidence (screenshots/videos)—and walks through initialization, project layout, and running the suite locally or in CI. Auto-wait and locator APIs are framed as the main lever for stability, so agents generate tests that fail for real regressions rather than timing noise. Use it in Ship when you are hardening releases, wiring GitHub Actions or similar, or replacing brittle manual click-through before a launch. It complements unit and API tests by covering what users actually see in the browser.2.7kinstalls3Tailwind CssTailwind CSS is a Claude MPM hub skill that packages comprehensive utility-first styling guidance for rapid UI work inside coding agents. Solo builders reach for it when scaffolding SaaS dashboards, marketing sites, or extension popups and want consistent responsive and dark-mode patterns without maintaining a separate design doc for every screen. The skill walks configuration, plugins, JIT behavior, and integration with common JavaScript frameworks named in the hub metadata. Progressive disclosure keeps lightweight triggers cheap while allowing deep patterns when the agent is actively editing templates or component markup. Use it during implementation when classes, theme extension, or layout refactors are the task—not when you need brand strategy from scratch or backend API design. It complements framework-specific skills rather than replacing component libraries or Figma handoff processes.1.2kinstalls4Trpc Type SafetytRPC Type Safety is an agent skill from the claude-mpm-skills TypeScript toolchain that teaches how to build APIs where server and client share inferred types—no code generation step and no duplicate DTO definitions. Solo and indie builders shipping SaaS or internal tools on Next.js and React Query use it when they want RPC-style ergonomics with compile-time guarantees on every procedure call. The skill walks through router definition, procedures, Zod-backed input validation, context and middleware, error handling, client wiring, and framework-specific integration paths. It explicitly relates to Zod as a prerequisite validation layer and points to adjacent skills for state, testing, and Next.js app structure. Invoke it while scaffolding or hardening the API layer so refactors surface as TypeScript errors instead of silent runtime mismatches.1.1kinstalls5PytestThis pytest skill packages industry-standard Python testing guidance for agents helping solo builders ship reliable backends and CLIs. It centers on pytest’s fixture system, parametrization, markers, rich assertion introspection, and a plugin ecosystem for coverage and async work. Use it when writing unit or integration tests, exercising HTTP APIs, mocking dependencies, or following TDD on FastAPI, Django, or Flask codebases. The progressive-disclosure entry point maps common triggers—async code, database layers, and framework apps—to concrete patterns so agents do not default to unittest boilerplate or brittle asserts. It is a toolchain skill (not user-invocable in upstream metadata) but still belongs in Prism for builders who want consistent test scaffolding during Ship and for test-first loops during Build. Version 1.0.0 from the Claude MPM team under MIT.908installs6Sqlalchemy OrmSQLAlchemy ORM is a Claude MPM toolchain skill that gives coding agents procedural depth for SQLAlchemy 2.0: typed declarative models, the modern select() API, relationship graphs, connection pooling, Alembic migrations, and async database access. Solo builders shipping FastAPI or Flask services use it when raw SQL or ad-hoc query strings are slowing them down or risking schema drift. The skill’s when_to_use list targets relational apps that need migrations and type-safe ORM layers across common engines. It is multi-phase in practice—you model during Build, evolve schemas into Ship and Operate—but the canonical Prism shelf stays Build backend as the primary discovery point for “how do I structure my data layer.” Prerequisites assume Python 3.10+ and basic SQL; deliverables are model modules, migration steps, and integration-ready session setup rather than a finished product schema without your domain rules.882installs7Nodejs Backend TypescriptNode.js Backend TypeScript is a comprehensive agent skill from the Claude MPM toolchain for solo builders shipping REST APIs and small SaaS backends. It walks through TypeScript-first Node patterns: choosing Express or Fastify, structuring routes and middleware, handling authentication with JWT, wiring databases through ORMs, and aligning tests with production deployment habits. The skill is user-invocable false and meant as progressive disclosure—agents pull the deep reference when the task already involves nodejs-backend-typescript or adjacent data skills. Pair it with typescript-core and a data skill (Drizzle, Kysely, or Prisma) when you are moving from prototype handlers to a maintainable API boundary.748installs8DockerThe docker skill from claude-mpm-skills teaches Docker containerization for solo and team builders who need isolated, portable units with dependencies bundled. It walks from core concepts and Dockerfile basics through multi-stage builds, Compose-based local stacks, dev workflows, production hardening, framework-specific examples, light orchestration notes, debugging, and troubleshooting. Invoke it when local development should mirror CI or production, when onboarding needs one-command environments, or when you are preparing images for deploy. The skill is structured as progressive disclosure with a large full token footprint for deep sections while keeping a short entry point. It complements application code skills rather than replacing a managed PaaS checklist—your agent still needs app-specific secrets and registry choices.723installs9Daisyuidaisyui is a reference-style agent skill for the most popular Tailwind CSS component plugin, giving solo builders semantic class names instead of long utility chains. When you are in Build and need buttons, cards, modals, and forms that share one theme, the skill documents install steps, tailwind.config plugin options, and when semantic components beat raw utilities. It highlights 50+ components and 30+ themes plus dark mode defaults so you can prototype SaaS settings pages, marketing blocks, and admin tables with a coherent design system across frameworks. The skill is marked user-invocable false in upstream metadata but remains valuable as agent procedural knowledge during implementation. It pairs naturally with Tailwind-first stacks and reduces design drift when you cannot staff a dedicated UI engineer. Expect beginner-friendly patterns once Node and Tailwind are already in the project.713installs10Drizzle MigrationsDrizzle Migrations encodes a migration-first database workflow for TypeScript and JavaScript projects using Drizzle ORM. Solo builders often tempt agents to edit schema.ts first; this skill reverses that habit so every table or column change begins as an explicit SQL migration, keeping dev, staging, and production aligned and making rollbacks thinkable. It explains why migrations beat code-first for drift prevention, forced design decisions, and pipeline checks, then expects you to apply those patterns to your repo with drizzle-kit-style workflows. It is marked non-user-invocable in metadata, so it is best as background guidance while your agent implements schema work during Build. Pair it with backend feature skills when you add entities, and with Ship testing when you wire migration checks into CI.659installs11ZodZod is an agent skill that teaches solo builders how to use the Zod library for TypeScript-first runtime validation with compile-time types inferred from schemas. It walks through primitives and nested objects, composition with union/intersection/refinement, transformations, async validators, and practical error formatting for APIs and forms. The skill is aimed at indie developers shipping SaaS or API products who want one source of truth for both types and validation instead of duplicating interfaces and manual checks. Use it while modeling request bodies, query params, webhooks, and configuration objects, or when wiring validation into frameworks the integrations section covers. Progressive disclosure keeps entry tokens small while deeper sections support advanced branded types and composition. It does not replace a dedicated security audit skill; it focuses on correct schema patterns and safe parsing habits.608installs12VitestVitest is a Prism agent skill that packages a code-examples index for the Vite-native test runner aimed at solo and indie builders shipping TypeScript apps. It walks you from baseline defineConfig setups through React and Vue environment choices, then into everyday patterns: structured describes, compile-time type assertions, and module mocking with vi.fn. The skill is for developers who want their coding agent to propose correct config and test snippets during Ship phase work rather than improvising Jest-era habits on a Vitest stack. Because Vitest shares Vite’s plugin graph, the examples emphasize globals, coverage with the v8 provider, and setup files for DOM testing—decisions that otherwise slow down first green CI runs. Use it when you are adding or fixing tests in a Vite monorepo, SPA, or small CLI package and need consistent agent output aligned with modern Vitest APIs.576installs13BiomeBiome is a Claude MPM reference skill that steers agents—and solo builders—toward @biomejs/biome for JavaScript and TypeScript linting and formatting in one toolchain. The package documents when to prefer Biome over a split ESLint and Prettier setup, especially when you want fast feedback loops and minimal configuration during Build and Ship. Quick start covers dev install, init, and check --write across the repo. Because user-invocable is false, it is meant to augment the model when you are editing front-end or full-stack TS code, not as a standalone chat command. For indie SaaS and extension projects, standardizing on Biome reduces agent confusion about conflicting format rules and speeds up the review pass before you ship. Pair it with TypeScript and framework skills in the same repo for coherent toolchain guidance.551installs