GSD Orchestrator
dimita800/gsd-flow-cli
Orquestrador autônomo para workflows AI e gestão de estado MCP
Ver no GitHubMotor de orquestração autônoma de IA para coordenar múltiplos agentes Claude Code, gerenciar transições de estado MCP e automatizar ciclos de vida de plugins. Funciona como um supervisor de tarefas que coordena fluxos de trabalho completos, desde geração de código até revisão e atualização automática de dependências.
Instalação
npx convez-ai dimita800/gsd-flow-cli1 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
- Coordenar múltiplos agentes IA em fluxos de trabalho complexos
- Manter contexto persistente e transições de estado entre sessões
- Automatizar revisão de código com pipelines multi-etapa
- Gerenciar atualizações de plugins com rollback seguro
Veredito

MCP server robusto para equipes que precisam orquestração em escala de múltiplos agentes IA. Bem documentado mas exige experiência com MCP e CLI para explorar todo potencial; recomendado para arquitetos de sistemas de IA.
README
A tradução para o português deste README ainda está na fila.
GSD Orchestrator: Autonomous AI Workflow Engine for Claude Code and MCP Agents
What is GSD Orchestrator?
Imagine a digital foreman that never sleeps, never forgets, and never loses track of what needs to be done. GSD Orchestrator is an autonomous AI orchestration engine designed to coordinate multiple Claude Code agents, manage MCP (Model Context Protocol) state transitions, enforce review workflows, and handle plugin lifecycle automation — all from a single command-line interface.
Built as a spiritual successor to lightweight task managers, GSD Orchestrator transforms your development environment into a self-sustaining productivity ecosystem. It is not a tool you use; it is a tool that uses itself to keep your projects moving forward, 24/7.
The Big Picture: How It Works
graph TD
A[User Input / Task Queue] --> B[Orchestrator Core]
B --> C{Task Type}
C -->|Code Generation| D[Claude Code Agent]
C -->|Review| E[Review Workflow Engine]
C -->|Plugin Update| F[Auto-Update Manager]
D --> G[MCP State Manager]
E --> G
F --> G
G --> H[Context Persistence Layer]
H --> I[Output / Feedback Loop]
I --> B
The orchestration loop above illustrates how GSD Orchestrator continuously processes tasks, delegates to specialized AI agents, manages context state, and feeds results back into the system. This creates a virtuous cycle of productivity where each completed task makes the next one faster.
🌟 Key Features
🧠 Autonomous Task Delegation
- Automatically routes tasks to the most suitable AI agent based on context, complexity, and current state
- Supports parallel execution of independent tasks across multiple agents
- Handles task prioritization with built-in dependency resolution
🔄 MCP State Management
- Maintains persistent context across sessions and agent invocations
- Tracks state transitions with full rollback capabilities
- Implements context compression to prevent token overflow in long-running workflows
📋 Review Workflow Automation
- Enforces multi-stage review pipelines (draft → peer review → QA → approval)
- Integrates with GitHub Actions, GitLab CI, and custom endpoints
- Provides explainable AI reviews with confidence scores and reasoning traces
🔌 Plugin Auto-Update System
- Monitors plugin registries for updates and security patches
- Applies updates with zero-downtime rollback protection
- Generates delta change logs for every update, reducing audit overhead
🌐 Multilingual Support
- Processes natural language commands in English, Spanish, French, German, Japanese, and Mandarin
- Translates task descriptions and outputs between languages on the fly
- Supports code-switching within single task sessions
📱 Responsive CLI and API
- Terminal-native interface with adaptive layouts for small and large screens
- REST API endpoint for integration with external dashboards and bots
- WebSocket support for real-time status streaming
⚡ 24/7 Customer Support Integration
- Built-in escalation engine for unresolved user queries
- Connects to Slack, Discord, Teams, and email
- Logs all interactions for continuous improvement of the AI response model
Example Profile Configuration
Create a gsd-profile.json file in your project root to define your orchestration preferences:
{
"orchestrator": {
"name": "production-pipeline",
"log_level": "info",
"max_concurrent_agents": 3
},
"agents": {
"claude": {
"model": "claude-3-opus-20240229",
"api_key_env": "ANTHROPIC_API_KEY",
"temperature": 0.7,
"max_tokens": 4096
},
"openai": {
"model": "gpt-4-turbo-preview",
"api_key_env": "OPENAI_API_KEY",
"temperature": 0.5
}
},
"workflows": {
"review": {
"stages": ["lint", "unit_test", "peer_review", "security_scan"],
"auto_approve": false,
"required_approvals": 2
}
},
"plugins": {
"auto_update": true,
"check_interval_hours": 6,
"allowed_sources": ["github", "npm", "pypi"]
}
}
Example Console Invocation
gsd-orchestrator run \
--task "Refactor the authentication module to use OAuth 2.0" \
--agent claude \
--workflow review \
--language python \
--notify slack
This single command will:
- Parse the task and analyze its context
- Invoke Claude Code with the specified model
- Execute the code generation
- Automatically route the output through the review workflow
- Send a summary notification to your Slack channel
🖥️ Operating System Compatibility
| OS | Version | Status | Emoji |
|---|---|---|---|
| Linux | Ubuntu 20.04+ | ✅ Supported | 🐧 |
| Linux | Debian 11+ | ✅ Supported | 🐧 |
| macOS | Ventura 13+ | ✅ Supported | 🍎 |
| macOS | Sonoma 14+ | ✅ Supported | 🍎 |
| Windows | 10 (WSL2) | ✅ Supported | 🪟 |
| Windows | 11 (WSL2) | ✅ Supported | 🪟 |
| FreeBSD | 13+ | ⚠️ Beta | 🎃 |
| Alpine | 3.18+ | ⚠️ Beta | 🏔️ |
🚀 Getting Started
Prerequisites
- Python 3.9 or higher
- An OpenAI API key (optional, for GPT-4 integration)
- An Anthropic API key (required, for Claude Code integration)
- Git 2.30+ for plugin management
Installation
-
Install via pip:
pip install gsd-orchestrator -
Set your API keys:
export ANTHROPIC_API_KEY="your-key-here" export OPENAI_API_KEY="your-key-here" # optional -
Initialize a new project:
gsd-orchestrator init --name my-ai-pipeline
🛠️ Advanced Usage
2026 Roadmap Features
The development roadmap for 2026 includes:
- Federated agent orchestration across multiple machines
- Memory augmentation using vector databases (Pinecone, Weaviate)
- Autonomous bug fixing with self-healing pipelines
- Multi-modal input support (voice, image, document)
- Enterprise SSO integration (Okta, Azure AD)
Custom Plugin Development
Create your own plugin using the GSD Plugin SDK:
from gsd_orchestrator.plugin import BasePlugin
class MyCustomPlugin(BasePlugin):
def execute(self, context):
# Your custom logic here
return {"status": "success", "data": context}
🔒 Security and Privacy
- All API keys are stored in environment variables, never in configuration files
- Context data is encrypted at rest using AES-256
- Network communication uses TLS 1.3
- No telemetry data is collected without explicit user consent
- Audit logs are tamper-proof and signed with HMAC
🤝 Contributing
We welcome contributions from the community! Please see our CONTRIBUTING.md for guidelines on:
- Code style and formatting
- Pull request process
- Testing requirements
- Documentation standards
📄 License
This project is licensed under the MIT License - see the LICENSE file for details.
⚠️ Disclaimer
GSD Orchestrator is an open-source tool provided "as is" without warranty of any kind, either express or implied. The use of this software is at your own risk. The developers are not responsible for:
- AI hallucination or incorrect outputs — Always verify critical outputs before using them in production.
- API cost overruns — The tool does not limit API usage; users are responsible for monitoring their own API consumption.
- Data loss or corruption — While the tool implements rollback features, no software is immune to edge cases.
- Third-party service changes — Integration with OpenAI, Anthropic, and other services is subject to their respective terms of service and availability.
- Security vulnerabilities — Regular updates are recommended, but absolute security cannot be guaranteed.
By using this software, you acknowledge that you have read and understood this disclaimer.
📬 Support
For issues, feature requests, or questions:
- Open a GitHub Issue
- Join our Discord Community
- Check the Wiki for tutorials and FAQs
GSD Orchestrator — Because your AI agents deserve a better boss.