giuseppe-trisciuoglio__developer-kit/plugins/developer-kit-ai

Developer Kit AI Plugin

AI/ML capabilities for Claude Code — prompt engineering, Retrieval-Augmented Generation (RAG), and document chunking strategies.

Overview

The developer-kit-ai plugin provides specialized skills and agents for building AI-powered applications. It covers the full lifecycle of LLM integration: from prompt design and optimization, to RAG pipeline implementation, to document chunking strategies for vector databases.

Components

Skills

Skill	Description	Reference Files
prompt-engineering	Write, debug, and optimize prompts for LLMs. Covers few-shot learning, chain-of-thought, system prompt design, and template composition.	5 reference files
rag	Implement document chunking, embedding, vector storage, and retrieval pipelines for RAG systems. Includes LangChain4j examples.	5 reference files + 2 assets
chunking-strategy	Generate chunk size recommendations (256–1024 tokens), overlap percentages (10–20%), and semantic boundary detection methods.	8 reference files

Agents

Agent	Description
prompt-engineering-expert	Expert agent for advanced prompting techniques, LLM optimization, and AI system design. Specializes in CoT, constitutional AI, and production prompt strategies.

Commands

Command	Description
/developer-kit-ai:devkit.prompt-optimize	Optimize prompts using advanced techniques (CoT, few-shot, constitutional AI) with three levels: basic, standard, advanced.

Dependencies

• developer-kit-core (required by all Developer Kit plugins)

Usage

This plugin is typically used alongside:

• developer-kit-java — LangChain4j integration with Java examples
• developer-kit-python — Python AI/ML development
• developer-kit-typescript — TypeScript/JavaScript AI integrations

Documentation

See docs/ for detailed guides:

• docs/README.md — Plugin overview and quick start
• docs/guide-agents.md — Agent capabilities and usage
• docs/guide-commands.md — Command reference with examples