Configuration factory for Claude Code. Templates, stacks, skills, agents, audit system, and a practices pipeline — all markdown + shell scripts.
Bootstrap any project with a complete .claude/ setup in seconds. Audit it. Keep it in sync as claude-kit evolves.
# 1. Clone
git clone https://github.com/luiseiman/claude-kit.git
cd claude-kit
# 2. Install globally (symlinks skills, agents, commands into ~/.claude/)
export CLAUDE_KIT_DIR="$(pwd)"
./global/sync.sh
# 3. In any project directory:
/forge bootstrap # Initialize .claude/ with full config
/forge audit # Audit configuration and get a score (0-10)
/forge sync # Update against current templateNew here? Read the Usage Guide for a complete walkthrough with examples.
There are many Claude Code starter kits, skills collections, and CLAUDE.md generators. claude-kit is different because it's the only end-to-end configuration management system — not a one-shot bootstrap or a static collection.
| Feature | What it means | Who else does this |
|---|---|---|
| Additive stack layering | Auto-detects your tech (16 stacks) and merges matching configs on top of a base template. Multi-stack projects get all layers combined. | No one — closest is project-type scanning, but without composable layering |
| MCP server templates | Ready-to-use config, permissions, and usage rules for 5 common MCP servers (GitHub, Postgres, Supabase, Redis, Slack). Auto-detected by /forge bootstrap. |
No one |
| Template sync with markers | <!-- forge:section --> separates managed sections from your customizations. /forge sync updates the managed parts without touching yours. |
No one |
| Audit scoring (0-10) | 12-item checklist (5 obligatory scored 0-2, 7 recommended scored 0-1), normalized to 10. Security-critical items cap the score at 6.0 if missing. Project tier (simple/standard/complex) adjusts expectations. | tw93/claude-health has tiers but no numeric normalization or security cap |
| Practices pipeline | Continuous improvement lifecycle: inbox/ → evaluating/ → active/ → deprecated/. Practices arrive from capture, web watch, repo scouting, audit gaps, or post-session hooks. |
No one |
| Cross-project registry | registry/projects.yml tracks audit scores with history across all managed projects. Spot regressions, compare configurations. |
No one |
| Global config via symlinks | global/sync.sh installs skills, agents, and commands as symlinks into ~/.claude/. One source of truth, instant propagation. |
No one with this symlink-based approach |
- Skills collections (superpowers, claude-skills) give you components — claude-kit gives you the system that manages them
- Starter kits (claude-bootstrap, claude-starter-kit) bootstrap once — claude-kit bootstraps, syncs, audits, and evolves
- CLAUDE.md generators generate one file — claude-kit generates and maintains the entire
.claude/directory - Audit tools (claude-health) check once — claude-kit checks, scores, tracks over time, and feeds gaps back into the practices pipeline
/forge bootstrap detects your project's tech stack and generates:
- CLAUDE.md — project instructions with build/test commands
- .claude/settings.json — permissions (allow/deny), hooks
- .claude/rules/ — contextual rules auto-loaded by glob patterns
- .claude/hooks/ — block destructive commands, lint on save
- .claude/commands/ — audit, debug, health, review
- .claude/agents/ — 7 specialized subagents with orchestration
- CLAUDE_ERRORS.md — error log for cross-session learning
Multi-stack projects get all matching stack configs merged automatically.
claude-kit/
├── template/ # Base scaffold (CLAUDE.md.tmpl, settings, hooks, rules, commands)
├── stacks/ # Technology modules (16 stacks, additive)
├── agents/ # 7 subagents (researcher, architect, implementer, ...)
├── skills/ # 15 skills installed as ~/.claude/skills/ symlinks
├── mcp/ # MCP server templates (github, postgres, supabase, redis, slack)
├── audit/ # Checklist (12 items) + scoring normalized to 10
├── practices/ # Pipeline: inbox → evaluating → active → deprecated
├── global/ # Global ~/.claude/ management (CLAUDE.md, settings, sync.sh)
├── registry/ # Project tracking with scores and history
├── hooks/ # Global post-session change detection hook
├── integrations/ # Cross-tool bridges (OpenClaw)
├── docs/ # Guides, patterns, security checklist
└── tests/ # Hook test suite + benchmark tasks
Each stack provides contextual rules, permissions, and optional hooks. Stacks are auto-detected and additive.
| Stack | Detects | Rules |
|---|---|---|
| python-fastapi | pyproject.toml, requirements.txt |
backend.md, tests.md |
| react-vite-ts | package.json with react/vite |
frontend.md |
| swift-swiftui | Package.swift, *.xcodeproj |
ios.md |
| supabase | supabase/, @supabase/supabase-js |
database.md |
| docker-deploy | Dockerfile, docker-compose* |
infra.md |
| data-analysis | *.ipynb, *.csv, *.xlsx |
data.md |
| gcp-cloud-run | app.yaml, cloudbuild.yaml |
gcp.md |
| redis | redis in dependencies |
redis.md |
| node-express | package.json with express/fastify |
backend.md |
| java-spring | pom.xml, build.gradle, Spring |
backend.md |
| aws-deploy | cdk.json, template.yaml (SAM) |
aws.md |
| go-api | go.mod, *.go |
backend.md |
| devcontainer | .devcontainer/ |
devcontainer.md |
| hookify | Custom hook framework | hooks.md |
| trading | Custom stack | trading.md |
MCP server templates (in mcp/) complement stacks by adding connection config, tool permissions, and usage rules for external services. See mcp/README.md.
Creating a new stack: see docs/creating-stacks.md.
All skills are invoked through the /forge command:
| Command | What it does |
|---|---|
/forge init |
Quick-start: auto-detect stack, 3 questions, generate personalized config |
/forge bootstrap |
Full interactive bootstrap with preview and confirmation |
/forge sync |
Update config against current template (merge, not overwrite) |
/forge audit |
Audit configuration + calculate score (0-10) |
/forge diff |
Show what changed in claude-kit since last sync |
/forge reset |
Restore .claude/ from template with backup |
/forge capture |
Register a practice in practices/inbox/ (with args) or auto-detect from session context (no args) |
/cap |
Shorthand alias for /forge capture — same behavior, 4 chars |
/forge update |
Process practices: inbox → evaluate → incorporate |
/forge watch |
Search for upstream changes in Anthropic docs |
/forge scout |
Review curated repos for useful patterns |
/forge export |
Export config to Cursor, Codex, Windsurf, or OpenClaw format |
/forge insights |
Analyze sessions for patterns and recommendations |
/forge rule-check |
Detect inert rules by cross-referencing globs against git history |
/forge benchmark |
Compare full config vs minimal config on standardized tasks |
/forge plugin |
Generate Claude Code plugin package for marketplace submission |
/forge unregister |
Remove a project from the registry |
/forge global sync |
Auto-update claude-kit + sync global ~/.claude/ config |
/forge global status |
Show global config status |
Seven specialized subagents, deployed to every bootstrapped project:
| Agent | Role | Model | Memory |
|---|---|---|---|
| researcher | Read-only codebase exploration | haiku | transactional |
| architect | Design decisions, tradeoff analysis | opus | persistent |
| implementer | Code + tests | sonnet | persistent |
| code-reviewer | Review by severity (critical/warning/suggestion) | sonnet | persistent |
| security-auditor | Vulnerability scanning | opus | persistent |
| test-runner | Run tests + report coverage | haiku | transactional |
| session-reviewer | Post-session analysis and pattern detection | sonnet | transactional |
Model routing rules are defined in template/rules/model-routing.md — criteria for haiku/sonnet/opus selection by task type.
Orchestration follows a decision tree: researcher → architect → implementer → test-runner → code-reviewer. See agents/ for definitions.
/forge audit scores your project's Claude Code configuration on a 10-point scale:
- 5 obligatory items (scored 0-2): settings.json, block-destructive hook, CLAUDE.md, rules, deny list
- 7 recommended items (scored 0-1): lint hook, commands, error log, agents, manifest, global hook, prompt injection scan
- Project tier: simple/standard/complex adjusts scoring expectations
- Security cap: missing settings.json or block-destructive hook caps score at 6.0
Scores are tracked in registry/projects.yml with history for trending over time.
A continuous improvement system for discovering and incorporating Claude Code configuration patterns:
inbox/ → evaluating/ → active/ → deprecated/
Practices arrive from: /forge capture or /cap (manual or auto-detected from session context), /forge update (web search), /forge watch (upstream docs), /forge scout (curated repos), audit gaps, or post-session hooks.
See practices/README.md for the lifecycle and format.
- Usage Guide — Complete step-by-step guide: install, bootstrap, sync, audit, practices (Español)
- Best Practices — Claude Code configuration patterns
- Security Checklist — 62 items for pre-deploy review
- Prompting Patterns — 10 reproducible patterns
- Creating Stacks — How to add a new technology stack
- Anatomy of CLAUDE.md — Deep dive into project instructions
- Memory Strategy — 5-layer memory policy for agents
- Troubleshooting — Common problems and diagnostics
- Changelog — Version history (v0.1.0 → v2.6.0)
- Roadmap — Completed features + upcoming v2.6.0
- Claude Code CLI installed
bash(hooks are shell scripts)python3withpyyaml(for registry validation, optional)
claude-kit uses a single environment variable:
export CLAUDE_KIT_DIR="/path/to/claude-kit"This is set automatically by global/sync.sh. All skills and hooks resolve paths through this variable.
See CONTRIBUTING.md.
Fábrica de configuración para Claude Code. Plantillas, stacks, skills, agentes, sistema de auditoría y un pipeline de prácticas — todo en markdown + shell scripts.
Inicializá cualquier proyecto con una configuración .claude/ completa en segundos. Auditalo. Mantenelo sincronizado a medida que claude-kit evoluciona.
# 1. Clonar
git clone https://github.com/luiseiman/claude-kit.git
cd claude-kit
# 2. Instalar globalmente (crea symlinks de skills, agentes y comandos en ~/.claude/)
export CLAUDE_KIT_DIR="$(pwd)"
./global/sync.sh
# 3. En cualquier directorio de proyecto:
/forge bootstrap # Inicializar .claude/ con configuración completa
/forge audit # Auditar la configuración y obtener un puntaje (0-10)
/forge sync # Actualizar contra la plantilla actualPrimera vez? Leé la Guía de Uso para un walkthrough completo con ejemplos.
Hay muchos starter kits, colecciones de skills y generadores de CLAUDE.md para Claude Code. claude-kit es diferente porque es el único sistema de gestión de configuración end-to-end — no un bootstrap one-shot ni una colección estática.
| Feature | Qué significa | Quién más lo hace |
|---|---|---|
| Stack layering aditivo | Auto-detecta tu tech (16 stacks) y mergea configs coincidentes sobre una plantilla base. Proyectos multi-stack reciben todas las capas combinadas. | Nadie — lo más cercano es detección de tipo de proyecto, pero sin layering composable |
| Templates de servidores MCP | Config, permisos y reglas de uso listas para 5 servidores MCP comunes (GitHub, Postgres, Supabase, Redis, Slack). Auto-detectados por /forge bootstrap. |
Nadie |
| Template sync con markers | <!-- forge:section --> separa secciones gestionadas de tus customizaciones. /forge sync actualiza lo gestionado sin tocar lo tuyo. |
Nadie |
| Audit scoring (0-10) | Checklist de 12 ítems (5 obligatorios 0-2, 7 recomendados 0-1), normalizado a 10. Ítems críticos de seguridad capean el score a 6.0 si faltan. Tier de proyecto (simple/standard/complex) ajusta expectations. | tw93/claude-health tiene tiers pero sin normalización numérica ni security cap |
| Pipeline de prácticas | Ciclo de mejora continua: inbox/ → evaluating/ → active/ → deprecated/. Las prácticas llegan desde capture, web watch, repo scouting, audit gaps, o hooks post-sesión. |
Nadie |
| Registry cross-proyecto | registry/projects.yml trackea audit scores con historial en todos los proyectos gestionados. Detectá regresiones, compará configuraciones. |
Nadie |
| Config global via symlinks | global/sync.sh instala skills, agentes y commands como symlinks en ~/.claude/. Una sola fuente de verdad, propagación instantánea. |
Nadie con este enfoque basado en symlinks |
- Colecciones de skills (superpowers, claude-skills) te dan componentes — claude-kit te da el sistema que los gestiona
- Starter kits (claude-bootstrap, claude-starter-kit) bootstrapean una vez — claude-kit bootstrapea, sincroniza, audita y evoluciona
- Generadores de CLAUDE.md generan un archivo — claude-kit genera y mantiene todo el directorio
.claude/ - Herramientas de auditoría (claude-health) revisan una vez — claude-kit revisa, puntúa, trackea en el tiempo y alimenta los gaps al pipeline de prácticas
/forge bootstrap detecta el stack tecnológico de tu proyecto y genera:
- CLAUDE.md — instrucciones del proyecto con comandos de build/test
- .claude/settings.json — permisos (allow/deny), hooks
- .claude/rules/ — reglas contextuales cargadas automáticamente por patrones glob
- .claude/hooks/ — bloqueo de comandos destructivos, lint al guardar
- .claude/commands/ — auditoría, debug, salud, revisión
- .claude/agents/ — 7 subagentes especializados con orquestación
- CLAUDE_ERRORS.md — registro de errores para aprendizaje entre sesiones
Los proyectos multi-stack reciben todas las configuraciones de stacks coincidentes fusionadas automáticamente.
claude-kit/
├── template/ # Scaffold base (CLAUDE.md.tmpl, settings, hooks, rules, commands)
├── stacks/ # Módulos tecnológicos (16 stacks, aditivos)
├── agents/ # 7 subagentes (researcher, architect, implementer, ...)
├── skills/ # 15 skills instalados como symlinks en ~/.claude/skills/
├── mcp/ # Templates de servidores MCP (github, postgres, supabase, redis, slack)
├── audit/ # Checklist (12 ítems) + puntaje normalizado a 10
├── practices/ # Pipeline: inbox → evaluating → active → deprecated
├── global/ # Gestión global de ~/.claude/ (CLAUDE.md, settings, sync.sh)
├── registry/ # Seguimiento de proyectos con puntajes e historial
├── hooks/ # Hook global post-sesión para detección de cambios
├── integrations/ # Bridges cross-tool (OpenClaw)
├── docs/ # Guías, patrones, checklist de seguridad
└── tests/ # Suite de tests para hooks + benchmark tasks
Cada stack provee reglas contextuales, permisos y hooks opcionales. Los stacks se auto-detectan y son aditivos.
| Stack | Detecta | Reglas |
|---|---|---|
| python-fastapi | pyproject.toml, requirements.txt |
backend.md, tests.md |
| react-vite-ts | package.json con react/vite |
frontend.md |
| swift-swiftui | Package.swift, *.xcodeproj |
ios.md |
| supabase | supabase/, @supabase/supabase-js |
database.md |
| docker-deploy | Dockerfile, docker-compose* |
infra.md |
| data-analysis | *.ipynb, *.csv, *.xlsx |
data.md |
| gcp-cloud-run | app.yaml, cloudbuild.yaml |
gcp.md |
| redis | redis en dependencias |
redis.md |
| node-express | package.json con express/fastify |
backend.md |
| java-spring | pom.xml, build.gradle, Spring |
backend.md |
| aws-deploy | cdk.json, template.yaml (SAM) |
aws.md |
| go-api | go.mod, *.go |
backend.md |
| devcontainer | .devcontainer/ |
devcontainer.md |
| hookify | Framework de hooks custom | hooks.md |
| trading | Stack custom | trading.md |
Los templates de servidores MCP (en mcp/) complementan los stacks con config de conexión, permisos de tools y reglas de uso para servicios externos. Ver mcp/README.md.
Para crear un nuevo stack: ver docs/creating-stacks.md.
Todos los skills se invocan a través del comando /forge:
| Comando | Qué hace |
|---|---|
/forge init |
Setup rápido: auto-detecta stack, 3 preguntas, config personalizada |
/forge bootstrap |
Bootstrap completo con preview y confirmación |
/forge sync |
Actualizar configuración contra la plantilla actual (merge, no sobreescritura) |
/forge audit |
Auditar configuración + calcular puntaje (0-10) |
/forge diff |
Mostrar qué cambió en claude-kit desde la última sincronización |
/forge reset |
Restaurar .claude/ desde la plantilla con backup |
/forge capture |
Registrar una práctica (con args) o auto-detectar desde el contexto de sesión (sin args) |
/cap |
Alias corto para /forge capture — mismo comportamiento, 4 chars |
/forge update |
Procesar prácticas: inbox → evaluar → incorporar |
/forge watch |
Buscar cambios upstream en la documentación de Anthropic |
/forge scout |
Revisar repos curados en busca de patrones útiles |
/forge export |
Exportar config a formato Cursor, Codex, Windsurf u OpenClaw |
/forge insights |
Analizar sesiones para patrones y recomendaciones |
/forge rule-check |
Detectar reglas inertes cruzando globs contra historial de git |
/forge benchmark |
Comparar config completa vs minimal en tareas estandarizadas |
/forge plugin |
Generar paquete de plugin para el marketplace de Claude Code |
/forge unregister |
Eliminar proyecto del registro |
/forge global sync |
Auto-actualizar claude-kit + sincronizar ~/.claude/ |
/forge global status |
Mostrar estado de la configuración global |
Siete subagentes especializados, desplegados en cada proyecto inicializado:
| Agente | Rol | Modelo | Memoria |
|---|---|---|---|
| researcher | Exploración de código (solo lectura) | haiku | transaccional |
| architect | Decisiones de diseño, análisis de tradeoffs | opus | persistente |
| implementer | Código + tests | sonnet | persistente |
| code-reviewer | Revisión por severidad (crítico/advertencia/sugerencia) | sonnet | persistente |
| security-auditor | Escaneo de vulnerabilidades | opus | persistente |
| test-runner | Ejecución de tests + reporte de cobertura | haiku | transaccional |
| session-reviewer | Análisis post-sesión y detección de patrones | sonnet | transaccional |
La orquestación sigue un árbol de decisión: researcher → architect → implementer → test-runner → code-reviewer. Las reglas de routing de modelos están en template/rules/model-routing.md. Ver agents/ para las definiciones.
/forge audit puntúa la configuración de Claude Code de tu proyecto en una escala de 10 puntos:
- 5 ítems obligatorios (puntaje 0-2): settings.json, hook de bloqueo destructivo, CLAUDE.md, rules, lista de denegación
- 7 ítems recomendados (puntaje 0-1): hook de lint, commands, registro de errores, agentes, manifiesto, hook global, scan de prompt injection
- Tier de proyecto: simple/standard/complex ajusta expectations de scoring
- Tope de seguridad: si falta settings.json o el hook de bloqueo destructivo, el puntaje máximo es 6.0
Los puntajes se registran en registry/projects.yml con historial para seguimiento de tendencias.
Un sistema de mejora continua para descubrir e incorporar patrones de configuración de Claude Code:
inbox/ → evaluating/ → active/ → deprecated/
Las prácticas llegan desde: /forge capture o /cap (manual o auto-detectado del contexto de sesión), /forge update (búsqueda web), /forge watch (docs upstream), /forge scout (repos curados), brechas de auditoría, o hooks post-sesión.
Ver practices/README.md para el ciclo de vida y formato.
- Guía de Uso — Guía completa paso a paso: instalación, bootstrap, sync, auditoría, prácticas (English)
- Best Practices — Patrones de configuración de Claude Code
- Security Checklist — 62 ítems para revisión pre-deploy
- Prompting Patterns — 10 patrones reproducibles
- Creating Stacks — Cómo agregar un nuevo stack tecnológico
- Anatomy of CLAUDE.md — Análisis detallado de las instrucciones de proyecto
- Memory Strategy — Política de memoria de 5 capas para agentes
- Troubleshooting — Problemas comunes y diagnósticos
- Changelog — Historial de versiones (v0.1.0 → v2.6.0)
- Roadmap — Features completadas + próximo v2.6.0
- CLI de Claude Code instalado
bash(los hooks son shell scripts)python3conpyyaml(para validación del registro, opcional)
claude-kit usa una única variable de entorno:
export CLAUDE_KIT_DIR="/path/to/claude-kit"Se configura automáticamente con global/sync.sh. Todos los skills y hooks resuelven rutas a través de esta variable.
Ver CONTRIBUTING.md.