giuseppe-trisciuoglio/developer-kit

What it solves

Developer Kit provides a modular system of plugins for AI coding assistants (like Claude Code) to replace generic AI responses with domain-specific expertise. It solves the problem of AI assistants lacking deep, repeatable knowledge of specific tech stacks, frameworks, and best practices across different programming languages.

How it works

It functions as a plugin marketplace that adds specialized capabilities to AI CLIs. The system is organized into four layers:

1. Skills: Reusable, on-demand capabilities that provide patterns and templates for specific tasks (e.g., Spring Boot CRUD generation).
2. Agents: Specialized sub-agents designed for complex workflows like security audits or architecture design.
3. Rules: Path-scoped instructions that automatically activate when the AI opens files matching specific patterns (e.g., enforcing constructor injection in .java files).
4. Commands: Custom slash commands (e.g., /devkit.refactor) to trigger specific actions.

It supports multiple CLIs including Claude Code, GitHub Copilot CLI, OpenCode CLI, and Codex CLI.

Who it’s for

Software developers who use AI coding assistants and want to enforce high-quality, language-specific standards and automate repetitive development tasks across Java, TypeScript, Python, PHP, and AWS environments.

Highlights

• Extensive Library: Includes over 150 skills, 45 agents, and 45 rules.
• Modular Installation: Users can install only the plugins relevant to their stack (e.g., developer-kit-java or developer-kit-typescript).
• Auto-Activation: Rules are automatically triggered based on file globs, requiring no manual configuration during the coding session.
• Broad Compatibility: Works across several major AI terminal tools and integrates with LSP servers for various languages.