convez.ai
SkillDevOps & InfraestruturaAvançado
i-stack

AI Coding Kit

i-stack/ai-coding-kit

Orquestração de Agent Skills e sincronização MCP para workflow de IA coding

Ver no GitHub
InicialPluginScore · de 0 a 100

Kit de engenharia local para manutenção de Agent Skills, sincronização de configuração MCP entre múltiplas ferramentas de IA (Claude Code, Cursor, Codex, Xcode) e experimentação com Gateway RAG universal. Inclui skill especializada em iOS/Swift com governança e ledger de evolução.

Instalação

$npx convez-ai i-stack/ai-coding-kit

1 comando no terminal: detecta o formato do repositório (marketplace ou plugin avulso) e instala via Claude Code. Requer Node 18+ e o Claude Code instalado.

Métricas

Estrelas
3
Forks
3
Acompanhando
0
Saúde
42%
Último commit
há 1 semana
Versão
2.0.0
Issues abertas
0
Contribuidores
1

Casos de uso

  • Manutenção centralizada de Agent Skills para Cursor, Claude Code e Codex
  • Sincronização de configuração MCP entre múltiplas plataformas de IA
  • Desenvolvimento iOS/Swift com regras de engenharia auditáveis e governance
  • Construção de Gateway RAG com roteamento de providers e memória semântica

Veredito

Ferramenta robusta para equipes que precisam orquestrar múltiplas instâncias de IA coding com consistência e auditoria. Oferece arquitetura real (MCP sync, RAG gateway) além de skill templates, mas exige familiaridade com shell scripting e conceitos de agentes.

README

A tradução para o português deste README ainda está na fila.

ai-coding-kit

Agent Skills iOS Engineer Skill MCP Config Sync Universal RAG Gateway

ai-coding-kit is a local AI coding workflow kit for maintaining Agent Skills, syncing MCP configuration, and experimenting with a Universal RAG Gateway. It is built for Codex, Claude Code, Cursor, Xcode Coding Assistant, and OpenAI-compatible clients that need shared engineering rules, tool configuration, memory, retrieval, and provider routing.

中文定位:这是一个面向 AI Coding / Agentic Coding / iOS Engineering / MCP / RAG Gateway 的本地工程化工具包,用于维护 Agent Skill、同步多端 MCP 配置,并沉淀可审计的工程规则与网关架构。

Overview

This repository has three connected parts:

AreaPurposeStart Here
Agent Skill engineeringMaintain, sync, and evolve reusable AI coding skills for Codex, Claude Code, Cursor, and Xcode-related paths.skills-engineering/README.md
MCP and platform config syncRender one local configuration source into Cursor, Codex, Claude Code, and Xcode host formats.sync/README.md
Universal RAG GatewayExplore an OpenAI-compatible gateway with provider routing, transcript storage, semantic memory, declarative tools, telemetry, and budget planning.docs/universal-rag-gateway.md

Core Capabilities

  • Agent Skills: source-managed skills for engineering discipline, iOS / Swift / SwiftUI / UIKit work, problem analysis, logical reasoning, epistemic integrity, and cognitive expansion.
  • Multi-host sync: one local config template for MCP servers, shared environment values, and platform-specific settings across AI coding tools.
  • iOS engineering rules: an auditable ios-engineer skill for architecture, concurrency, networking, UI, performance, testing, review, migration, and release-risk control.
  • Gateway architecture: a TypeScript / Fastify Universal RAG Gateway with OpenAI-compatible routing, provider adapters, memory, tool execution, GraphRAG direction, and observability.

Quick Start

TaskDocumentation
Set up or sync Agent Skillsskills-engineering/README.md
Configure MCP and platform settingssync/README.md
Create local config from the templateenv/config.json.example
Inspect the iOS engineer skillskills-engineering/ios-engineer/SKILL.md
Study the Gateway design and current implementation statusdocs/universal-rag-gateway.md
Install repository-managed Git hooksinstall-hooks.sh

Documentation

TopicLink
Agent Skill engineeringskills-engineering/README.md
iOS engineer skill sourceskills-engineering/ios-engineer/SKILL.md
iOS skill rule indexskills-engineering/ios-engineer/references/rule_index.md
Cognitive expansion skillskills-engineering/cognitive-expansion/SKILL.md
Engineering discipline skillskills-engineering/engineering-discipline/SKILL.md
Logical reasoning skillskills-engineering/logical-reasoning/SKILL.md
Epistemic integrity skillskills-engineering/epistemic-integrity/SKILL.md
Problem analysis skillskills-engineering/problem-analysis/SKILL.md
MCP and platform config syncsync/README.md
Universal RAG Gatewaydocs/universal-rag-gateway.md
Token comparison notesdocs/token-comparison-results-v2.md

Repository Layout

PathRole
skills-engineering/Agent Skill sources, references, sync scripts, validation data, and skill evolution workflow.
sync/Renderers and orchestration for local MCP / platform config sync.
env/Local configuration template; the real env/config.json is intentionally gitignored.
rag-gateway/Universal RAG Gateway source, tests, providers, retrieval, memory, telemetry, and declarative tool runtime.
docs/Architecture notes, Gateway status, and token comparison reports.
.githooks/Repository-managed commit and push guards.

Who This Is For

  • Developers who want reusable Agent Skills for Codex, Claude Code, Cursor, or Xcode workflows.
  • iOS / Swift engineers who want explicit AI coding rules for Swift, SwiftUI, UIKit, Xcode, testing, review, migration, and release work.
  • AI infrastructure builders experimenting with local memory, retrieval, declarative tools, provider routing, and OpenAI-compatible gateway patterns.
  • Maintainers who want one source of truth for MCP configuration across multiple AI coding hosts.
Ver o original no GitHub

Bom para

Relacionados