mrgoonie/claudekit-skills
33 skills16.1k installs72.2k starsGitHub
Install
npx skills add https://github.com/mrgoonie/claudekit-skillsSkills in this repo
1Backend DevelopmentThe backend-development skill. Build robust backend systems with modern technologies (Node.js, Python, Go, Rust), frameworks (NestJS, FastAPI, Django), databases (PostgreSQL, MongoDB, Redis), APIs (REST, GraphQL, gRPC), authentication (OAuth 2.1, JWT), testing strategies, security best practices (OWASP Top 10), performance optimization, scalability patterns (microservices, caching, sharding), DevOps practices (Docker, Kubernete. The workflow follows the source SKILL.md contract with progressive reference loading, clear trigger phrases, and practical steps developers can apply directly in agent sessions. The workflow follows the source SKILL.md contract with progressive reference loading, clear trigger phrases, and practical steps developers can apply directly in agent sessions. The workflow follows the source SKILL.md contract with progressive reference loading, clear trigger phrases, and practical steps developers can apply directly in agent sessions. The workflow follows the source SKILL.md contract with progressive reference loading, clear trigger phrases, and practical steps developers can apply directly in agent sessions.1.9kinstalls2Sequential ThinkingThe sequential thinking skill Use when complex problems require systematic step-by-step reasoning with ability to revise thoughts, branch into alternative approaches, or dynamically adjust scope. Ideal for multi-stage analysis, design planning, problem decomposition, or tasks with initially unclear scope. Documentation covers workflows, commands, and guardrails agents should follow when users invoke this capability. Key documented areas include **Iterative reasoning**: Break complex problems into sequential thought steps; **Dynamic scope**: Adjust total thought count as understanding evolves; **Revision tracking**: Reconsider and modify previous conclusions; **Branch exploration**: Explore alternative reasoning paths from any point. Reference commands include 1. Start with initial thought (thoughtNumber: 1); 2. For each step:. Use when developers or agents need structured guidance for sequential thinking tasks with evidence grounded in the bundled SKILL.md rather than generic advice. **Iterative reasoning**: Break complex problems into sequential thought steps **Dynamic scope**: Adjust total thought count as understanding evolves **Revision tracking**: Reconsider and modify previo.1.6kinstalls3ThreejsThe threejs skill build 3D web apps with Three js WebGL WebGPU Use for 3D scenes animations custom shaders PBR materials VR XR experiences games data visualizations product configurators Three js Development Build high-performance 3D web applications using Three js a cross-browser WebGL WebGPU library When to Use This Skill Use when working with 3D scenes models animations or visualizations WebGL WebGPU rendering and graphics programming Interactive 3D experiences games configurators data viz Camera controls lighting materials or shaders Loading 3D assets GLTF FBX OBJ or textures Post-processing effects bloom depth of field SSAO Physics simulations VR XR experiences or spatial audio Performance optimization instancing LOD frustum culling Progressive Learning Path Level 1 Getting Started Load references 00-fundamentals md Fundamentals Load references 01-getting-started md Scene setup basic geometries materials lights rendering loop Level 2 Common Tasks Asset Loading references 02-loaders md GLTF FBX OBJ texture loaders Textures references 03-textures md Types mapping wrapping filtering Cameras references 04-cameras md Perspective orthographic controls Lights references 05-lights md.1.1kinstalls4Devopsdevops is a Claudekit skill (version 2.0.0, MIT license) for deploying and operating cloud infrastructure from an AI coding session. The skill covers Cloudflare Workers, R2, D1, and Pages; Docker and Docker Compose containerization; GCP via gcloud for Cloud Run, GKE, and Cloud SQL; and Kubernetes cluster operations with kubectl and Helm plus GitOps flows using Argo CD and Flux. Developers reach for devops when shipping serverless APIs, containerized services, multi-region deployments, CI/CD pipelines, or when running security audits with RBAC and network policies. Platform selection guidance routes each workload to the right target instead of defaulting every app to one cloud.925installs5Chrome DevtoolsChrome DevTools is a claudekit-skills agent integration (Apache-2.0) providing browser automation through executable Puppeteer CLI scripts that output JSON for agent parsing. It supports screenshots, performance analysis, network traffic monitoring, web scraping, form automation, and JavaScript debugging. Installation may require Linux/WSL system libraries for headless Chrome. Developers reach for Chrome DevTools when agents must verify frontend behavior, capture page evidence, or profile load performance without manual browser sessions—checking working directory before script execution as documented in the skill quick start.703installs6Aestheticaesthetic is a mrgoonie/claudekit-skills generator that outputs a Design Guidelines document with sections for design style (Minimalism, Glassmorphism, Neo-brutalism, Modern, Memphis, Flat Design), primary and accent color palettes with hex values and usage notes, neutral surfaces, color psychology rationale, and layout guidance before any UI code is written. Developers reach for aesthetic at project kickoff or redesign planning when engineers and agents need a single reference for tokens, mood, and component tone instead of ad-hoc CSS choices. The template includes placeholders for primary, secondary, accent, success, warning, error, background, surface, and border colors plus generated date metadata, making the artifact ready to hand to frontend implementation skills.675installs7Ui Stylingui-styling is a Claude Code skill from mrgoonie/claudekit-skills that teaches agents how to produce Tailwind CSS and component styling guidance without inventing invalid utility classes or violating an existing design system. With 515 installs on skills.sh, it is used while implementing React, Vue, or HTML interfaces where class naming consistency matters. The skill encodes conventions for spacing, typography, responsive breakpoints, and component variants so generated markup stays production-ready. Developers reach for ui-styling when an AI assistant keeps suggesting nonexistent Tailwind classes, mixes conflicting design tokens, or drifts from a project's established UI patterns during feature work or refactors.568installs8Media ProcessingMedia Processing is an Agent Skill from mrgoonie/claudekit-skills that equips coding agents to work with images, audio, and video beyond simple file reads. The skill supports analysis, resizing, format conversion, and caption generation so agents can preprocess assets for web apps, content pipelines, or multimodal features. Developers reach for Media Processing when building upload flows, thumbnails, transcoding steps, or accessibility captions without writing ad-hoc ffmpeg or Pillow scripts from scratch. Catalog metadata shows 477 installs and rank 50 on skills.sh, indicating strong adoption for agent-driven media workflows in ClaudeKit-based projects.524installs9Problem Solvingproblem-solving is an agent skill from mrgoonie/claudekit-skills that structures how Claude tackles unclear technical problems through repeatable decomposition and solution-generation steps. Listed on skills.sh with 487 installs and rank 5 in its source repo, it is invoked when a task lacks a clear path, spans multiple components, or needs explicit trade-off analysis before implementation. The skill does not ship code generators or domain-specific APIs; instead it shapes agent behavior so debugging, architecture choices, and feature breakdowns follow a consistent method. Developers reach for problem-solving when they want Claude to slow down, enumerate constraints, compare options, and output an actionable plan rather than jumping straight to patches. It pairs well with implementation-focused skills because it front-loads clarity. Skip it when the task is a single-file fix with an obvious edit or when you already have a written spec to execute verbatim.523installs10Debuggingdebugging is a mrgoonie/claudekit-skills meta-skill with 440 catalog installs and four bundled sub-skills: systematic-debugging, root-cause-tracing, defense-in-depth, and verification-before-completion. Systematic debugging enforces a four-phase flow—Root Cause Investigation, Pattern Analysis, Hypothesis Testing, Implementation—with the rule that no fixes happen before root-cause investigation. Root-cause-tracing walks backward through call stacks to fix invalid data at the source; defense-in-depth adds validation at entry, business logic, environment guards, and instrumentation layers; verification-before-completion requires fresh command output before claiming done. The skill reports systematic debugging fixes issues in 15–30 minutes versus 2–3 hours of thrashing, with a 95% first-time fix rate versus 40% for ad-hoc guessing. Developers reach for debugging when tests fail, production errors surface, or an agent is about to declare a bug fixed without proof.500installs11Code Reviewcode-review is a mrgoonie/claudekit-skills agent workflow ranked #6 on skills.sh with 452 installs that provides structured, high-signal feedback on code changes before merging. Developers reach for code-review when they need a second pass on pull request diffs, want to catch logic errors and maintainability issues, or lack a human reviewer available in the workflow. The skill focuses on actionable review comments rather than generic praise, helping teams maintain code quality gates in CI-driven delivery. Install from mrgoonie/claudekit-skills via skills.sh and invoke it against staged or PR diffs for consistent review output.488installs12Ai Multimodalai-multimodal is a mrgoonie/claudekit-skills package with 437 installs that exposes Google Gemini multimodal capabilities to coding agents under MIT license. It transcribes audio up to 9.5 hours with timestamps, analyzes images with captioning, object detection, OCR, and visual Q&A, processes videos including YouTube URLs up to 6 hours, extracts tables and charts from PDFs, and generates images from text prompts. The skill supports Gemini 2.5 and 2.0 models with context windows up to 2M tokens and resolves GEMINI_API_KEY from environment or layered .env paths including .claude/skills/ai-multimodal/.env. Allowed tools include Bash, Read, Write, and Edit. Developers reach for ai-multimodal when implementing agent features over screenshots, recordings, long meetings, or document ingestion without bespoke SDK wiring per media type. Bundled reference docs cover audio processing, image understanding, video analysis, document extraction, and image generation scripts. Google AI Studio and Vertex AI configurations share the same GEMINI_API_KEY discovery order, reducing setup friction across ClaudeKit ai-ml-tools plugin installs.483installs13Context Engineeringcontext-engineering is a mrgoonie/claudekit-skills discipline for structuring what coding agents see and how they invoke tools. It covers prompt templates, context window budgeting, retrieval chunk boundaries, and tool schema design so agents remain accurate and bounded when working across large monorepos. Developers reach for context-engineering when agent sessions hallucinate, lose thread on long tasks, or mis-invoke tools because context is noisy or schemas are ambiguous. The skill applies during agent skill authoring, MCP tool design, and repo-specific instruction files where precision and token limits matter.458installs14Shopifyshopify is a ClaudeKit agent skill for building on the Shopify platform with @shopify/cli, GraphQL Admin API, REST legacy endpoints, Polaris design system, and Liquid templating. The skill covers OAuth app setup via shopify.app.toml scopes, checkout_ui_extension and admin_block extension types, POS UI extensions, Shopify Functions for discounts and validation, and theme development with shopify theme dev on port 9292. Three reference guides cover app development, extensions, and themes, plus a shopify_init.py scaffold script. API examples use the 2025-01 Admin API version with pagination, bulk operations, and webhook signature verification patterns. Developers reach for shopify when creating public or custom apps, customizing checkout, managing products and orders via API, or building Liquid storefront sections. Best practices emphasize minimal scopes, GraphQL field selection to control query cost, and development-store testing before production deployment.422installs15Databasesdatabases is a mrgoonie/claudekit-skills agent skill providing unified guidance for MongoDB document databases and PostgreSQL relational systems. It includes a selection guide for flexible schema versus ACID joins, quick-start commands for mongosh and psql, CRUD and indexing examples, and eight reference guides split across MongoDB CRUD, aggregation, indexing, Atlas, and PostgreSQL queries, psql CLI, performance, and administration topics. Python utilities db_migrate.py, db_backup.py, and db_performance_check.py support migrations, backups, and slow-query analysis. Best practices cover embedded versus referenced documents in MongoDB, 3NF normalization and EXPLAIN ANALYZE in PostgreSQL, and replication, sharding, and backup strategies for production. Developers reach for databases when designing schemas, writing aggregations or JOINs, optimizing indexes, or administering Atlas or self-hosted Postgres—not for frontend UI work unrelated to persistence.406installs16Web Testingweb-testing is a Claude Code skill from mrgoonie/claudekit-skills for comprehensive browser and API testing before release. It recommends a testing pyramid of roughly 70% unit (Vitest/Jest under 50ms), 20% integration (Vitest with fixtures, 100–500ms), and 10% E2E (Playwright, 5–30s). Quick-start commands include `npx vitest run`, `npx playwright test`, `k6 run load-test.js`, `npx @axe-core/cli`, and `npx lighthouse`. Reference docs cover E2E Playwright workflows, component testing, flakiness mitigation, Core Web Vitals, mobile gestures, cross-browser matrices, and CI gates where unit tests fail fast before E2E and accessibility checks. Developers reach for web-testing to automate regression checks, load validation, visual regression, and a11y scans on critical flows like login, checkout, and payment.403installs17Frontend Designfrontend-design is a mrgoonie Claudekit-skills agent skill aligned with Claudekit frontend-design principles for crafting web interfaces that are both usable and visually distinctive. The skill helps agents reason about information hierarchy, page layouts, component selection, spacing, typography, and interaction patterns before or while writing frontend code. Developers reach for frontend-design when a UI feels generic, confusing, or structurally weak and they want deliberate design direction rather than default component stacking. The skill fits early mockup conversations, refactor passes on existing pages, and agent-assisted prototyping where visual quality and clarity matter. Use frontend-design whenever interface decisions should precede or guide HTML, CSS, and component implementation.381installs18Repomixrepomix is a Claude Code skill from mrgoonie/claudekit-skills wrapping the Repomix CLI for packaging entire code repositories into single AI-friendly files. The skill supports customizable include and exclude patterns, git-aware processing that respects .gitignore, token counting for LLM context budgets, security checks for sensitive information, and four output formats: XML, Markdown, JSON, and plain text. Common workflows cover codebase analysis, third-party library evaluation, security audit preparation, documentation context generation, and bug investigation across large trees. Bundled references include configuration.md for patterns and output options and usage-patterns.md for audit and analysis workflows. Developers reach for repomix when an agent needs full repository context in one bounded artifact instead of ad-hoc file-by-file reads during review or refactor tasks.372installs19Docs Seekerdocs-seeker is a Claude Code skill (version 1.0.0) from mrgoonie/claudekit-skills for intelligent technical documentation discovery before implementing features. The skill prioritizes llms.txt and llms-full.txt standardized AI-friendly docs, falls back to GitHub repository analysis via Repomix when llms.txt is missing, and deploys parallel Explorer or Researcher agents for scattered sources. Context7 MCP integration fetches version-specific API references, migration guides, and official code examples from live documentation servers. Tool routing maps WebSearch to llms.txt discovery, WebFetch to single pages, Task Explore to parallel URLs, and Repomix to full repository packing. Developers reach for docs-seeker when implementing unfamiliar libraries—Better Auth OAuth setup, Next.js App Router handlers, or Prisma v6 migrations—instead of relying on outdated model training data.371installs20Better Authbetter-auth is a MIT-licensed agent skill (version 2.0.0) in mrgoonie/claudekit-skills for the Better Auth TypeScript authentication framework. Better Auth ships email/password, social OAuth, session management, and RBAC out of the box, with plugins for 2FA, passkeys, magic links, usernames, organizations, and rate limiting. The skill documents a feature-selection matrix, client-server architecture, and framework integrations spanning Next.js, Nuxt, SvelteKit, Remix, Astro, Hono, and Express. It references dedicated guides for email-password flows, OAuth providers, and advanced features, and points to the Better Auth CLI (`npx auth@latest generate`) for Drizzle, Prisma, or other ORM schema output. Developers reach for better-auth when adding verified login, OAuth, MFA, or multi-tenant org auth without re-reading the entire Better Auth docs during implementation.352installs21Frontend Dev Guidelinesfrontend-dev-guidelines is a mrgoonie claudekit-skills playbook for modern React and TypeScript frontend development inside Claude Code and compatible agents. It promotes Suspense-based data fetching with `useSuspenseQuery`, lazy loading via `React.lazy`, feature-based directory layouts under `features/{name}/`, and TanStack Router route files under `routes/`. Styling guidance targets MUI v7 with inline styles under 100 lines and separate files when components exceed that threshold. Import aliases such as `@/`, `~types`, `~components`, and `~features` keep modules organized, while `useMuiSnackbar` replaces ad-hoc loading spinners and early-return patterns. Topic guides cover data fetching, routing, performance, file organization, loading and error boundaries, and TypeScript best practices. Developers reach for this skill when creating components, pages, or features and want Claude-generated code to match production-grade React conventions instead of generic snippets.343installs22Web Frameworksweb-frameworks is a Claude agent skill from mrgoonie/claudekit-skills that guides framework selection and project scaffolding for modern full-stack web apps. The skill compares routing models, rendering strategies—SSR, SSG, SPA—and directory conventions so new repos start with consistent patterns instead of ad-hoc folder layouts. Developers reach for web-frameworks at greenfield project kickoff, when migrating between React, Next.js, Vue, Nuxt, SvelteKit, or similar stacks, or when an agent needs explicit conventions before generating routes and components. It outputs recommended stack choices, folder structures, config stubs, and naming patterns aligned to the chosen framework. The skill reduces rework from mismatched routing, mixed rendering modes, and missing tooling defaults that surface only after the first deploy.334installs23Mcp Managementmcp-management is a Claude Code skill from mrgoonie/claudekit-skills for end-to-end MCP server lifecycle work in agent projects. The skill guides developers through registering MCP servers, validating tool schemas, diagnosing connection failures, and keeping tool definitions aligned with what Claude agents can invoke. It targets the gap between dropping an MCP package into a project and having stable, callable tools in Claude Code, Cursor, or similar agents. Developers reach for mcp-management when agents fail tool calls, schemas drift from API changes, or new local utilities need exposure as MCP tools without manual trial-and-error.330installs24Skill Creatorskill-creator is a Claude Code meta-skill (version 2.0.0) from mrgoonie/claudekit-skills that walks developers through the full skill authoring lifecycle. The skill explains the SKILL.md anatomy—YAML frontmatter under 200 characters, a body kept under 150 lines, and optional scripts/, references/, and assets/ folders loaded via progressive disclosure. It prescribes a six-step creation process covering example gathering, internet research, reusable resource planning, init_skill.py scaffolding, SKILL.md editing, and package_skill.py validation before zip distribution. Developers reach for skill-creator when they need repeatable agent capabilities—API integrations, domain workflows, or packaged CLI scripts—that Claude should auto-trigger from metadata descriptions. The skill also documents plugin marketplace distribution and links five validation-criteria reference files plus marketplace hosting guides for teams shipping skills beyond a single repository.327installs25Mermaidjs V11mermaidjs-v11 is a Claudekit agent skill in mrgoonie/claudekit-skills for creating text-based diagrams with Mermaid.js v11 declarative syntax. It covers 24+ diagram types—including flowchart, sequenceDiagram, classDiagram, stateDiagram, erDiagram, gantt, and journey—and supports inline markdown code blocks, frontmatter theming, JavaScript browser embedding, and CLI rendering via @mermaid-js/mermaid-cli (`mmdc -i diagram.mmd -o diagram.svg`). Reference files document diagram-types, configuration, cli-usage, integration, and examples for architecture, API flows, database schemas, and user journeys. Developers reach for mermaidjs-v11 when README files, ADRs, or design docs need accurate, version-correct Mermaid v11 syntax without opening a separate drawing tool.323installs26Payment Integrationpayment-integration is a mrgoonie/claudekit-skills framework (version 2.2.0, MIT license) for production payment wiring across five providers: SePay (Vietnamese VietQR and 44+ banks), Polar (SaaS MoR subscriptions), Stripe (CheckoutSessions, Billing, Connect, Payment Element), Paddle (MoR subscriptions with tax compliance), and Creem.io (licensing and revenue splits). It bundles reference docs for Stripe server SDKs, stripe-js Payment Element, stripe-cli local testing, webhook idempotency, subscription lifecycle management, and multi-provider order orchestration. Developers reach for payment-integration when adding monetization to apps and needing provider-specific checkout patterns, verified webhooks, and regional coverage guidance rather than reading scattered provider documentation alone.319installs27Mcp Buildermcp-builder is a structured guide for building Model Context Protocol servers that let LLM agents interact with external services. From mrgoonie/claudekit-skills, it walks a four-phase workflow: research and agent-centric design, implementation, review and refinement, and evaluation creation. The skill emphasizes workflow-oriented tools over thin API wrappers, input validation with Pydantic or Zod, context-efficient responses, pagination, and corrective error messages. Implementation references cover Python with FastMCP and TypeScript with the MCP SDK, plus mcp_best_practices.md, python_mcp_server.md, node_mcp_server.md, and evaluation.md. Developers reach for mcp-builder when exposing databases, SaaS APIs, or internal workflows to Claude Code, Cursor, or other MCP clients. Deliverables include tool inventories tied to user workflows, strict schemas, and evaluation scripts confirming agent usability. It ships in the platform-tools suite alongside mcp-management for runtime discovery.282installs28Template Skilltemplate-skill is a starter scaffold from mrgoonie/claudekit-skills for authoring new Claude Code agent skills. The template provides a YAML frontmatter block with name and description fields plus a markdown body section for instructions, triggers, and usage examples. Developers reach for template-skill when creating a first custom skill, enforcing consistent metadata across a skill library, or onboarding team members to the Claude Code skill format without copying ad-hoc markdown files.266installs29BunnyGuides integration of Bunny.net CDN, object storage, and streaming services into web and ecommerce apps so teams can offload assets, reduce origin load, and configure delivery URLs correctly.135installs30Docxdocx is a Claude Code skill for ai & agent building. It helps solo builders move faster with AI-assisted coding.57installs31Pptxpptx is a Claude Code skill for ai & agent building. It helps solo builders move faster with AI-assisted coding.56installs32Xlsxxlsx is a Claude Code skill for ai & agent building. It helps solo builders move faster with AI-assisted coding.56installs33Pdfpdf is a Claude Code skill for ai & agent building. It helps solo builders move faster with AI-assisted coding.54installs