Cybrix Skills Marketplace
amrtaha12393-ai/cybrix-skills-marketplace
Marketplace de skills modulares para Claude com deploy via CLI
Ver no GitHubFramework de orquestração de plugins para Claude que permite implantar skills modulares e reutilizáveis via marketplace centralizado. Conecta definições de skills open-source com backend Cybrix, oferecendo automação de IA em produção com arquitetura composável.
Instalação
npx convez-ai amrtaha12393-ai/cybrix-skills-marketplace1 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
Casos de uso
- Instalar e gerenciar plugins de automação IA em ambiente de desenvolvimento
- Orquestrar pipelines de processamento de linguagem natural customizados
- Estender funcionalidades de código através de skills parametrizáveis
Veredito

Projeto ambicioso que propõe uma camada de orquestração de skills para Claude, mas com documentação incompleta no README fornecido e foco em arquitetura sem exemplos práticos suficientes. Útil para equipes que buscam standardizar extensões de IA, porém requer exploração aprofundada do código-fonte para compreender viabilidade real.
README
A tradução para o português deste README ainda está na fila.
Cybrix-Skills Extension Hub: Claude-Powered Plugin Deployment Framework
Version 2.4.0 | Released January 2026 | Open-Source MIT License
Overview
The Cybrix-Skills Extension Hub transforms how developers integrate Claude Code capabilities into their workflows. This repository provides a modular plugin architecture that connects open-source skill definitions with the Cybrix hosted backend, enabling rapid deployment of AI-powered automation tools. Whether you need natural language processing pipelines, data extraction agents, or custom assistant behaviors, this framework delivers production-ready skills that install via a single marketplace command.
Think of this project as a skill orchestration layer between your development environment and Claude's reasoning engine. Each skill is a self-contained unit of functionality that can be mixed, matched, and extended without modifying core infrastructure. The result is a composable AI toolkit that grows with your requirements.
Core Architecture
graph TD
A[Developer CLI] --> B{Claude Code Market}
B --> C[Skill Repository]
C --> D[Plugin Registry]
D --> E[Cybrix Backend]
E --> F[AI Processing Pipeline]
F --> G[Output Handler]
G --> H[Response Formatter]
H --> I[User Interface]
I --> J[Feedback Loop]
J --> C
The architecture follows a unidirectional data flow where each skill module acts as a self-contained transformation node. The Cybrix backend handles authentication, rate limiting, and model routing, while the skills themselves define the prompt templates and response schemas.
Installation & Setup
Prerequisites
Before deploying skills, ensure your environment meets these requirements:
Compatibility Table
| Operating System | Support Status | Tested Version |
|---|---|---|
| 🐧 Ubuntu 22.04+ | Full Support | Kernel 5.15+ |
| 🍎 macOS Sonoma | Full Support | 14.4+ |
| 🪟 Windows 11 Pro | Partial Support | Build 22621+ |
| 🐧 Debian 12 | Full Support | Bookworm |
| 🍎 macOS Ventura | Deprecated | 13.6+ |
Quick Start
Execute the following command in your terminal:
claude plugin marketplace add cybrixcc/cybrix-skills
This command pulls the latest skill definitions, validates them against your Claude environment, and registers available plugins with the Cybrix backend. The process takes approximately 15 seconds on a standard broadband connection.
Verification
Confirm successful installation by running:
claude plugin list --filter cybrix
Expected output shows a table of available skills with their version numbers and activation status.
Example Profile Configuration
Create a .cybrix-skills.yaml file in your project root to customize skill behavior:
profile:
name: developer-workflow
language: en
timezone: UTC
skill_defaults:
temperature: 0.3
max_tokens: 2048
streaming: true
plugins:
- name: code-review
config:
strictness: high
language_rules: python,javascript,typescript
- name: documentation-generator
config:
format: markdown
include_examples: true
This configuration activates two skills with specific parameters. The code-review skill applies strict linting rules across multiple languages, while the documentation generator produces markdown output with inline code examples.
Example Console Invocation
Trigger a skill from the command line:
claude execute cybrix-skills code-review --file src/main.py --config strictness=medium
The system responds with:
Analyzing src/main.py with code-review v2.4.0
▸ 12 style suggestions found
▸ 3 potential logic errors detected
▸ 2 performance optimizations recommend
▸ Estimated refactor time: 45 minutes
▸ Full report saved to reports/main_2026_01_15.md
The invocation displays real-time progress indicators followed by a structured summary. Results persist to disk for later review.
API Integration
The framework supports direct API access for programmatic control.
OpenAI API Integration
Connect to OpenAI models via the cybrix-openai bridge:
from cybrix_skills import SkillManager
manager = SkillManager(provider="openai")
response = manager.execute(
skill="data-extractor",
input={"text": "Invoice dated 2026-01-15 for $2,400"},
model="gpt-4-turbo"
)
print(response.structured_output)
Claude API Integration
Native Claude API support provides lower latency and better reasoning alignment:
from cybrix_skills import ClaudeSkillPlugin
plugin = ClaudeSkillPlugin(api_key="sk-...")
result = plugin.invoke(
skill="semantic-search",
parameters={"query": "recent server outages", "limit": 10}
)
Both APIs support asynchronous calls, streaming responses, and custom timeout configurations.
Key Features
Responsive UI Components
The skill output handler renders results differently based on terminal capabilities. On modern terminals with Unicode support, you get formatted tables and progress bars. Dumb terminals receive plain text fallback without loss of information.
Multilingual Support
Skills detect input language automatically and respond in the same language. Current support includes English, Spanish, French, German, Japanese, and Mandarin Chinese. Additional languages route through a translation layer that preserves formatting.
24/7 Uptime Backend
The Cybrix hosted infrastructure maintains 99.9% availability with automatic failover across three geographic regions. Skills process requests asynchronously even during backend maintenance windows, queuing results for delivery once the system recovers.
Plugin Hot-Reloading
Modify skill definitions in real time without restarting your Claude session. The framework watches skill directories for changes and reloads modified plugins within 500 milliseconds.
Granular Permission System
Each skill runs in a sandboxed environment with configurable permissions. Control file system access, network requests, and environment variable visibility on a per-skill basis.
Skill Development Guide
Creating a Custom Skill
Define a new skill by creating a JSON file in the skills/ directory:
{
"name": "sentiment-analyzer",
"version": "1.0.0",
"description": "Analyzes text sentiment with emotional granularity",
"prompt_template": "Analyze the sentiment of the following text. Provide a score from -1 to 1 and identify the primary emotion:\n\n{input}",
"input_schema": {
"type": "object",
"properties": {
"text": {"type": "string"}
}
},
"output_schema": {
"type": "object",
"properties": {
"score": {"type": "number"},
"emotion": {"type": "string"},
"confidence": {"type": "number"}
}
}
}
Register the skill by adding it to the registry.json manifest file:
{
"skills": ["sentiment-analyzer"],
"backend_connection": "wss://skills.cybrix.io/ws"
}
Run the validation tool to check for schema errors:
claude plugin validate --skill sentiment-analyzer
Use Cases
Automated Code Review Pipelines
Integrate the code-review skill into your CI/CD pipeline for automated pull request analysis. The skill generates inline comments on style violations, security vulnerabilities, and performance regressions.
Documentation Generation from Code
Transform source code into comprehensive documentation using the doc-generator plugin. The skill reads your codebase structure and produces formatted guides with API references and usage examples.
Data Extraction from Unstructured Sources
Convert PDFs, images, and audio files into structured data using the extraction pipeline. The skill applies OCR, speech-to-text, and entity recognition in a single invocation.
Troubleshooting
Connection Issues
If skills fail to load, verify network access to skills.cybrix.io on port 443. Corporate firewalls may require allowing WebSocket connections to this domain.
Version Mismatches
Run claude plugin update cybrix-skills to synchronize your local installation with the latest backend capabilities. Outdated skills may produce degraded results or fail entirely.
Permission Denials
Check your Claude API key has the skills:execute scope enabled. Generate a new key at your Cybrix dashboard if permissions appear incorrect.
License
This project is released under the MIT License. You are free to use, modify, and distribute the code in any project, commercial or personal.
See the full license text at: MIT License
Disclaimer
Skills interact with external AI models that may produce unexpected outputs. Always review generated code before deployment. The Cybrix backend does not log your input data, but third-party model providers may store prompts according to their own privacy policies. This framework is provided as-is without warranty of merchantability or fitness for a particular purpose. The year 2026 systems referenced throughout this document refer to projected capabilities and may differ from actual runtime behavior. Users assume all responsibility for outputs generated through these skills.
Cybrix-Skills Extension Hub – Building intelligent workflows one plugin at a time