convez.ai
SkillAutomaçãoAvançado
SeanVasey

00Claude — Plataforma de Skills para Agentes IA

SeanVasey/00Claude

IDE unificada para descobrir, curar, versionear e deployar capabilities de agentes

Ver no GitHub
BomPluginScore · de 0 a 100

Biblioteca centralizada e IDE para authoring, descoberta e versionamento de skills de agentes baseados em Claude. Oferece registro de personas (AGENTS.md), índice de skills (SKILLS.md) e integração nativa com a API Anthropic para código-geração assistida.

Instalação

$npx convez-ai SeanVasey/00Claude

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
1
Forks
0
Acompanhando
0
Saúde
71%
Último commit
há 2 meses
Versão
—
Issues abertas
1
Contribuidores
4

Casos de uso

  • Construir e manter biblioteca corporativa de skills reutilizáveis para agentes
  • Orquestrar múltiplas personas de subagentes com autoridades e modelos específicos
  • Colaborar no desenvolvimento de prompts e skills com preview markdown em tempo real

Veredito

Ferramenta sólida para times que já trabalham com agentes Claude e precisam de governança, versionamento e reutilização sistemática de skills. A integração com Anthropic API e o conceito de persona registry (AGENTS.md) trazem estrutura profissional, mas requer familiaridade com arquitetura de agentes e TypeScript.

README

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


What's New — v0.5.0

  • Agent + Skill platform integration — new AGENTS.md persona registry and SKILLS.md skill index, both wired into the smoke check so they cannot drift from .claude/.
  • First-class .claude/ scaffold — settings.json, six subagent personas under .claude/agents/, and three Anthropic-format skills (verify, code-review, security-review) under .claude/skills/.
  • Hardened CI — added npm run typecheck and npm audit --audit-level=high gates to .github/workflows/ci.yml.
  • Security patch — bumped next and eslint-config-next to 16.2.6, clearing the three high-severity advisories present in 16.1.6.

See CHANGELOG.md for the full history.

Features

Authoring & IDE

  • AgentVault IDE — AI-powered code generation with live markdown preview
  • Document management — Create, edit, version, and organize CLAUDE.md, AGENTS.md, and skill files
  • Anthropic API integration — Claude-based skill authoring through a server-side proxy with mock fallback

Agent & Skill platform

  • Persona registry (AGENTS.md) — cross-tool source of truth for who performs work, with what model, and with what authority
  • Skill registry (SKILLS.md) — index of Anthropic-format skills under .claude/skills/, kept in sync by the smoke check
  • Subagent definitions — .claude/agents/*.md for lead-engineer, implementer, researcher, reviewer, security-specialist, ux-specialist, and release-engineer

Design & UX

  • Glassmorphic UI — Spectrum gradient accents, ambient canvas animations, custom typography
  • PWA support — installable with a full app-icon suite (transparent source SVG → 14 raster sizes) and web manifest

Quality gates

  • Markdown linting — markdownlint-cli2 quality gates for documentation
  • Smoke checks — required files plus skill-registry drift detection
  • Typecheck — tsc --noEmit in CI
  • npm audit — fails CI on high/critical only; moderate is informational

Deployment

  • CI/CD pipeline — GitHub Actions on every PR and main push
  • Vercel-ready — production config with secure response headers
  • GitHub Pages workflow — conditional static export via GITHUB_PAGES=true

Tech Stack

LayerTechnology
FrameworkNext.js 16.2.6 (App Router)
LanguageTypeScript 5
UIReact 19.2.3, Tailwind CSS v4
Iconslucide-react
LintingESLint 9, markdownlint-cli2
CIGitHub Actions
DeploymentVercel (primary), GitHub Pages
Package Managernpm (Node.js 20+)

Getting Started

Prerequisites

  • Node.js 20+
  • npm 10+

Install

npm install

Development

npm run dev

Visit http://localhost:3000 to access the AgentVault IDE.

Build

npm run build

Run Checks

npm test            # lint:md + smoke checks
npm run lint        # ESLint
npm run lint:md     # Markdown linting
npm run typecheck   # TypeScript --noEmit
npm run test:smoke  # File + skill-registry smoke check
npm audit --audit-level=high  # Fails CI on high/critical only

Environment Variables

Copy .env.example to .env for local development:

cp .env.example .env
VariableDescription
APP_NAMEApp or repository display name
NODE_ENVEnvironment mode (development / production)
ANTHROPIC_API_KEYAnthropic API key (server-side only)

Architecture

.
├── .claude/             # Claude Code config, agents, skills, hooks
│   ├── settings.json    # Permissions + env
│   ├── agents/          # Subagent personas (lead-engineer, reviewer, ...)
│   └── skills/          # Anthropic-format skills (verify, code-review, ...)
├── .github/workflows/   # CI + GitHub Pages deploy
├── app/                 # Next.js App Router pages and API routes
│   ├── api/anthropic/   # Claude API proxy endpoint
│   ├── layout.tsx       # Root layout (fonts, metadata)
│   └── page.tsx         # Home page
├── components/          # React UI components
│   └── AgentVault.tsx   # Main IDE component (all views)
├── docs/                # Supporting project documentation
├── public/              # Static assets (icons, SVGs, manifest)
├── scripts/             # Validation helper scripts
├── tasks/               # Task tracking (todo.md, lessons.md)
├── AGENTS.md            # Agent persona registry (who does the work)
├── SKILLS.md            # Skill registry (what discrete capabilities exist)
├── CLAUDE.md            # Project-level operating contract
├── CHANGELOG.md         # Release history
├── CODE_OF_CONDUCT.md   # Contributor conduct expectations
├── LICENSE              # Apache 2.0
├── SECURITY.md          # Vulnerability disclosure policy
├── package.json         # Scripts and dependencies
└── vercel.json          # Deployment configuration

Deployment

Vercel

This repository includes a vercel.json configured with clean URLs, no trailing slashes, and secure default headers (X-Content-Type-Options, X-Frame-Options, Referrer-Policy, Permissions-Policy). Connect the repository to Vercel and deploy using the default Next.js build settings.

GitHub Pages

The CI workflow includes a build verification step. For GitHub Pages static export, configure next.config.ts with output: 'export' and enable Pages in repository settings pointing to the gh-pages branch or GitHub Actions deployment.

Contributing

  1. Fork the repo and create a feature branch
  2. Make your changes following the standards in CLAUDE.md
  3. Run npm test and npm run build to verify
  4. Open a pull request with a clear description

License

Apache 2.0

Security

See SECURITY.md for vulnerability reporting.

Ver o original no GitHub

Bom para

Relacionados