Crew

About

Crew acts as an 'AI talent market' for AI IDEs, allowing users to define reusable professional AI skills as 'digital employees' using Markdown and YAML. It's not another agent framework but focuses on defining 'who does what' for the AI IDE to execute. Through the MCP protocol, it exposes prompts, resources, and tools, enabling deep integration with AI development environments for structured skill invocation, project detection, multi-employee discussions, and automated pipelines.

Key Features

• Define AI skills using Markdown and YAML with typed parameters and workflow separation
• Supports multi-employee discussions and sequential skill pipelines for complex tasks
• 1 GitHub stars
• Features a 4-layer skill discovery mechanism for project, global, and built-in skills
• Integrates with AI IDEs via Model Context Protocol (MCP) for skill invocation
• Includes 6 built-in digital employees (e.g., Code Reviewer, Test Engineer, Product Manager)

Use Cases

• Automating the generation and enhancement of unit tests for codebases
• Performing intelligent code reviews with a specialized digital employee
• Facilitating multi-role discussions among digital employees for architecture review or feature design