multi-agent-dev-team_skill

Overview

This skill provides a two-agent collaborative development workflow: a PM (Project Manager) agent that plans and coordinates, and a Dev (Developer) agent that implements code. Describe the project, and the PM breaks it into tasks, spawns the Dev agent, monitors progress, and returns deliverables. It is optimized for small web apps, landing pages, prototypes, and TypeScript/React projects.

How this skill works

You act as the Director and tell the PM agent what to build. The PM creates a structured task specification, spawns the Dev agent with that spec, and monitors implementation. The Dev agent writes code, runs tests, and commits to Git; the PM reviews results and requests up to three revision cycles if needed. Configure models and workspace paths to balance speed and quality.

When to use it

• Building landing pages, marketing sites, or simple web apps
• Prototyping MVPs and front-end features quickly
• Generating TypeScript/React components or Next.js apps
• Iterating on small scopes where automated coding speeds delivery
• Learning or testing multi-agent development workflows

Best practices

• Start with an MVP scope and split complex projects into phases
• Provide specific requirements, examples, and acceptance criteria
• Use faster models for rapid iteration and higher-quality models for final polish
• Keep dev environment dependencies available (Node, Git, package managers)
• Review and test all generated code before deploying to production

Example use cases

• Create a Next.js landing page with Tailwind CSS, hero, features, and contact form
• Generate a reusable React component library with TypeScript and Storybook examples
• Implement a small API integration (e.g., GitHub profile search) and responsive UI
• Prototype a single-page app or interactive widget to validate UX quickly
• Initialize a git repo and produce deployable artifacts ready for human review