Jettyio Skills: A Cross-Platform Unified Orchestration Framework for AI/ML Workflows

This article introduces Jettyio Skills—an open-source framework based on the MCP protocol, designed to address the fragmentation issue of AI development tools. It supports multiple environments such as Claude Code, Gemini CLI, Cursor, and VS Code Copilot, providing unified workflow definition, multi-environment execution, real-time monitoring, and an extensible skill ecosystem. It helps developers efficiently build, run, and monitor AI/ML workflows, suitable for team collaboration, MLOps construction, and rapid prototyping.

With the development of AI technology, tools like Claude Code and Gemini CLI are becoming increasingly abundant, but the problem of tool fragmentation is prominent: each tool has its own independent extension mechanism, configuration method, and runtime environment. Developers need to reconfigure and rewrite scripts when switching between tools, reducing efficiency; each link of the AI/ML workflow is managed separately, making unified monitoring and governance difficult.

Jettyio Skills is an open-source project that achieves cross-platform compatibility through the MCP (Model Context Protocol) server extension architecture. MCP is an open protocol launched by Anthropic, which standardizes the interaction between AI assistants and external tools. Therefore, Jettyio Skills can seamlessly integrate with MCP-supported clients such as Claude Code, Gemini CLI, Cursor, and VS Code Copilot, providing a unified workflow management experience.

The core features of Jettyio Skills include:

1. Unified Workflow Definition Language: A declarative syntax to describe AI/ML task flows, lowering the threshold for migration;
2. Multi-Environment Execution Engine: Supports local, container, and remote server execution, automatically handles dependencies, resource allocation, and error recovery, and supports GPU acceleration;
3. Real-Time Monitoring and Observability: Tracks execution status, resource consumption, and intermediate results, providing a unified dashboard;
4. Extensible Skill Ecosystem: Modular "skills" encapsulate common ML tasks, supporting community contributions and customization.

Jettyio Skills is compatible with mainstream AI tools:

• Claude Code: Deep integration, allowing Claude to directly call Skills to execute complex tasks;
• Gemini CLI: Ensures consistent workflow management across tools;
• Cursor and VS Code Copilot: Presents features in the form of a command palette or sidebar for easy GUI operation.

The application scenarios of Jettyio Skills include:

1. Data Science Team Collaboration: Eliminates differences between different development environments and shares unified workflow definitions;
2. MLOps Pipeline Construction: A lightweight starting point that supports complete pipelines from data processing to model deployment, improving reliability;
3. Rapid Prototyping of AI Applications: Uses the pre-built skill library to quickly combine and implement ideas, accelerating iteration.

Jettyio Skills provides a solution for the diversity of AI development tools and unified workflow governance, which is worthy of attention from developers and teams. Future directions include: enriching the pre-built skill library (establishing a skill market), adding cloud collaboration features (version management, permission control), and providing enterprise-level features (audit logs, CI/CD integration), etc.