By Kaakati
Enterprise SDLC standards for a Software Development House: Rails + Phlex backend, React Native, ReactJS (Vite), Next.js, Terraform. Ships 57 skills (37 workflow + 20 path-scoped convention skills), 12 specialized agents, and quality-gate hooks with wrapper-agnostic framework detection.
Architecture and design advisor. Use when making architectural decisions, evaluating technical debt, planning large refactors, designing system components, or reviewing system design.
Validate and enforce Clean Architecture principles — dependency direction, layer boundaries, and architectural conformance. Use this agent for architectural reviews, layer boundary violations, dependency analysis, and structural refactoring toward Clean Architecture patterns.
Code quality reviewer. Use when reviewing pull requests, auditing code quality, checking adherence to team conventions, or evaluating maintainability and technical debt.
Design critique partner for visual quality review. Use when reviewing UI components for design quality, evaluating visual hierarchy, auditing design token compliance, checking cross-platform consistency, scoring against Nielsen's heuristics, or evaluating storytelling / narrative UX (StoryBrand-style hero framing, emotional beats, narrative arc).
Design system architect for Apple-level visual standards. Use when building or auditing a design token system, establishing component specs, defining grid systems, or producing design system documentation across Phlex, ReactJS, Next.js, and React Native.
React Native and Expo performance best practices with 35+ rules across 14 categories. Use when optimizing list performance, implementing animations with Reanimated, reviewing React Native code for performance, or working with native modules. Triggers on "React Native performance", "list optimization", "FlashList", "Reanimated best practices", "mobile performance", or "React Native best practices".
Build React Native mobile features with Zustand state management, TanStack Query data fetching, and Centrifugo real-time subscriptions. Use this skill whenever someone asks to build a mobile screen, create a component, add a Zustand store, write a TanStack Query hook, or says things like "build the mobile UI for X", "create a screen for Y", "add real-time updates", "implement offline support", "set up navigation for X", or "how should I structure this component". Also trigger when someone mentions React Native performance, MMKV storage, map integration, or push notification implementation.
Build ReactJS (Vite) web SPA features with React Router, Zustand client state, TanStack Query server state, Tailwind CSS styling, Framer Motion animations, and ApexCharts dashboards. Use this skill whenever someone asks to build a web page, create a Vite component, add a web dashboard, build a web SPA feature, or says things like "build the web UI for X", "create a web page for Y", "add a dashboard chart", "set up Vite routing", "build the admin panel", or "create a web form". Also trigger when someone mentions Vite configuration, Tailwind component styling, Framer Motion page transitions, or ApexCharts integration.
Safely refactor code using Fowler's patterns with test-first methodology and incremental steps. Use this skill whenever someone asks to refactor code, reduce technical debt, extract patterns, restructure modules, or says things like "refactor this", "clean up this code", "extract this into a service", "this code smells", "reduce duplication", "decompose this function", or "restructure this module". Also trigger for code smell identification, large-scale codebase transformations, or safe migration of legacy code.
Clarify vague requirements, discover hidden assumptions, and produce scoped user stories with acceptance criteria. Use when requirements are ambiguous, incomplete, or when starting a new feature, epic, or project. Also trigger when someone says "what should we build", "scope this feature", "break this down", "write user stories", "acceptance criteria", "requirements gathering", "feasibility check", or mentions unclear specifications.
Executes bash commands
Hook triggers when Bash tool is used
Modifies files
Hook triggers on file write and edit operations
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
Uses power tools
Uses Bash, Write, or Edit tools
Uses power tools
Uses Bash, Write, or Edit tools
Enterprise-grade Claude Code plugin for a professional Software Development House.
A complete, audited system of skills, agents, and hooks that transforms Claude Code into a full SDLC partner — from requirements gathering through production incident response.
This repository is a Claude Code plugin (sdh) that enforces enterprise development standards across the entire software development lifecycle. It is designed for teams building Rails API (Phlex views) + React Native mobile + ReactJS Vite SPA + Next.js App Router applications deployed on AWS and Vercel.
Instead of relying on ad-hoc prompting, this plugin provides:
/sdh:code-reviewer, /sdh:rails-architect, …), 20 std-* convention skills that auto-load by file path (e.g. std-rails-conventions, std-accessibility), plus the always-on sdh-engineering-standards skillsdh-engineering-standards skill carrying the stack + library conventions# Add this repo as a marketplace, then install the plugin
/plugin marketplace add Kaakati/sdh-claude-skills
/plugin install sdh@sdh-claude-skills
# Or test locally without installing
claude --plugin-dir /path/to/sdh-claude-skills
Skills are namespaced under the plugin: /sdh:code-reviewer, /sdh:rails-architect, etc.
Run /plugin to manage it, and claude plugin validate . to validate changes.
.claude/settings.jsonA plugin cannot ship permissions, env, or worktree settings. Copy them from this
repo's .claude/settings.json into your own project settings to get
the secret/build-artifact Read denies, the agent-teams env flag, and worktree symlinks.
See docs/monorepo-setup.md.
The system auto-detects each framework — your wrapper directory can be named
anything (backend/, api/, server/, web/, frontend/, next/, mobile/, or even
the repository root). Conventions load from each framework's own layout and marker files,
not from a forced top-level folder name.
Detection uses two wrapper-agnostic signals:
Canonical internal structure — each framework's own conventional sub-paths, matched
anywhere in the tree. app/models/*.rb is Rails whether it lives in backend/app/models/,
api/app/models/, or app/models/. src/pages/ is a Vite SPA; src/screens/ is React
Native — regardless of the wrapper.
On-disk project markers — when structure alone is ambiguous (e.g. a src/components/*.tsx
that could be web or mobile), the hooks walk up the tree to the nearest marker file:
| Framework | Marker files |
|---|---|
| Rails | Gemfile, config/application.rb, bin/rails |
| Next.js | next.config.{js,ts,mjs,cjs}, "next" in package.json |
| ReactJS (Vite) | vite.config.{js,ts}, index.html |
| React Native | metro.config.js, app.json, "react-native" in package.json |
So putting your Rails code in api/ instead of backend/ works fine — the Rails rules and
hooks still activate.
The std-* convention skills are path-scoped — they auto-load when you edit files matching their globs:
File you edit (under any wrapper) std-* skills that auto-load
-------------------------------------- --------------------------------------
**/app/**/*.rb -> std-rails-conventions, std-clean-architecture
**/app/components/**/*.rb -> std-phlex-conventions
**/app/views/**/*.rb -> std-phlex-conventions, std-i18n
**/src/pages/**/*.tsx -> std-reactjs, std-accessibility
**/src/screens/**/*.tsx -> std-react-native
**/app/**/*.tsx + **/next.config.* -> std-nextjs, std-accessibility
**/i18n/**, **/config/locales/** -> std-i18n
**/migrations/**, **/migrate/** -> std-database
**/*.tf, **/*.tfvars -> std-terraform-conventions, std-infrastructure, std-monitoring
**/*.test.*, **/*.spec.* -> std-testing
The PostToolUse checkers apply a 200-line limit to models and UI components (300 elsewhere), matched by canonical structure under any wrapper:
npx claudepluginhub kaakati/sdh-claude-skills --plugin sdhFlutter development with GetX state management, Clean Architecture, multi-agent orchestration, quality gates, comprehensive testing patterns, navigation, i18n, performance optimization, and accessibility patterns
ReAcTree Rails development with multi-agent orchestration, LSP-powered context compilation, Guardian validation cycle for type safety, Solargraph/Sorbet/Rubocop integration with blocking quality gates, requirements translation (user stories to beads tasks), Pre/Post edit validation hooks, parallel execution, dual memory systems with 24h TTL caching, iterative control flow (LOOP, CONDITIONAL), backwards communication (FEEDBACK edges), Claude CLI intelligent intent detection with manifest-based agent/skill discovery, utility agents, UX Engineer, optimized model selection, and comprehensive skills library for A+ type-safe code
iOS and tvOS development with SwiftUI, MVVM, Clean Architecture, 14 specialized agents, 27 comprehensive skills, automated quality gates, hooks system, one-command setup (/ios-init), beads integration, offline sync, push notifications, tvOS focus navigation, accessibility testing, performance profiling, and dual memory systems
Enterprise-grade Rails development workflow with multi-agent orchestration, beads task tracking, and quality gates. Generic and portable across any Rails project.
Generates a structural CODEBASE_MAP.md for LLM-assisted coding — extracts dependencies, interfaces, side effects, and merges human annotations.
Comprehensive skill pack with 66 specialized skills for full-stack developers: 12 language experts (Python, TypeScript, Go, Rust, C++, Swift, Kotlin, C#, PHP, Java, SQL, JavaScript), 10 backend frameworks, 6 frontend/mobile, plus infrastructure, DevOps, security, and testing. Features progressive disclosure architecture for 50% faster loading.
Develop, test, build, and deploy Godot 4.x games with Claude Code. Includes GdUnit4 testing, web/desktop exports, CI/CD pipelines, and deployment to Vercel/GitHub Pages/itch.io.
Harness-native ECC plugin for engineering teams - 67 agents, 271 skills, 92 legacy command shims, reusable hooks, rules, MCP conventions, and operator workflows for Claude Code plus adjacent agent harnesses
Lazy senior dev mode. Forces the simplest, shortest solution that actually works: YAGNI, stdlib first, no unrequested abstractions.
Tools to maintain and improve CLAUDE.md files - audit quality, capture session learnings, and keep project memory current.
Comprehensive PR review agents specializing in comments, tests, error handling, type design, code quality, and code simplification